Artificial Intelligence AI News & updates this week

Google unveils AI-powered Nest indoor and outdoor cameras, and a new doorbell

techcrunch.com • 19h ago

Google's new Nest cameras and doorbell are getting a Gemini upgrade....

TL;DR

Google introduces a revamped lineup of Nest devices with 2K HDR video, improved low-light performance, and enhanced features powered by its AI assistant Gemini AI.

Key Takeaways:

All new Nest devices come with 2K HDR video, wider field of view, and improved low-light performance using a new sensor.
Google's AI assistant Gemini AI adds more context to notifications, allowing users to understand events in their homes more easily.
Gemini AI will also help users automate home settings and improve energy usage through personalized recommendations.

First look at the Google Home app powered by Gemini

www.theverge.com • Sep 19, 2025

Hey Google, meet Gemini: the new voice of your smart home

www.theverge.com • 19h ago

Hands-on with the new Google Home Speaker coming next year

www.theverge.com • 19h ago

Google teases its new Gemini-powered Google Home speaker, coming in spring 2026

techcrunch.com • 19h ago

Google is upgrading almost every Nest security device - and there's a new Home speaker, too

www.zdnet.com • 19h ago

Google Brings Gemini AI to the Home With New Cameras, Doorbell and Speaker

www.bloomberg.com • 19h ago

Google’s Home app, a command center for the smart home, gets a Gemini upgrade

techcrunch.com • 19h ago

Google reveals its Gemini-powered smart home lineup and AI strategy

techcrunch.com • 19h ago

Google's 'Gemini for Home' upgrade is coming - and it'll work with older Nest devices

www.zdnet.com • 19h ago

Gemini for Home: Expanding the Platform for a New Era of Smart Home AI - Google for Developers Blog

developers.googleblog.com • Yesterday

Your Google Nest devices are getting a big upgrade for free - these older models included

www.zdnet.com • 17h ago

Get the details on the new Google Home Premium subscription service - The Keyword

blog.google • 19h ago

Every Google Nest product that launched with Gemini today (and a new home speaker)

www.zdnet.com • 12h ago

Gemini for Home: The helpful home gets an AI upgrade - The Keyword

blog.google • 19h ago

Anthropic launches Claude Sonnet 4.5, its best AI model for coding

techcrunch.com • Sep 29, 2025

Anthropic says its new AI model is robust enough to build production-ready applications, rather than just prototypes....

TL;DR

Anthropic introduces Claude Sonnet 4.5, a state-of-the-art AI model capable of building 'production-ready' applications.

Key Takeaways:

Claude Sonnet 4.5 can autonomously code for up to 30 hours without human intervention, completing tasks such as setting up database services and performing security audits.
The model offers improved performance on coding benchmarks, including SWE-Bench Verified, and is less prone to deception and sycophancy compared to previous models.
Claude Sonnet 4.5 is available via the Claude API and in the Claude chatbot, with pricing remaining the same as its predecessor, Claude Sonnet 4.

OpenAI says GPT-5 stacks up to humans in a wide range of jobs

techcrunch.com • Sep 25, 2025

A new test from OpenAI aims to understand how close AI is to outperforming humans at economically valuable work....

TL;DR

OpenAI's GDPval benchmark shows its GPT-5 model outperforming human professionals in 40.6% of tasks across various industries and jobs.

Key Takeaways:

OpenAI's GPT-5 model outperformed or matched human professionals in 40.6% of 44 occupations across 9 industries, including healthcare, finance, and manufacturing.
Anthropic's Claude Opus 4.1 model outperformed or matched human professionals in 49% of tasks, but may have done so due to pleasing graphics rather than sheer performance.
OpenAI plans to create more robust tests in the future that can account for more industries and interactive workflows, acknowledging the limitations of the current GDPval-v0 benchmark.

Clarifai’s new reasoning engine makes AI models faster and less expensive

techcrunch.com • Sep 25, 2025

The AI platform Clarifai has launched a new reasoning engine that it claims will make running AI models twice as fast and 40% less expensive....

TL;DR

Clarifai introduced a new reasoning engine that claims to make running AI models twice as fast and 40% less expensive.

Key Takeaways:

The new reasoning engine employs multiple optimizations to improve inference performance, including CUDA kernels and speculative decoding techniques.
Industry benchmark tests recorded industry-best records for both throughput and latency, verifying the performance claims of the new engine.
The engine is specifically designed to handle the increased computing demands of agentic and reasoning models, which require multiple steps in response to a single command.

OpenAI’s new social app is filled with terrifying Sam Altman deepfakes

techcrunch.com • 14h ago

OpenAI's Sora app makes it too easy for people to create misleading AI content....

Our new partnership with the University of Waterloo on the future of learning and work

blog.google • 18h ago

Google and the University of Waterloo are bringing together their respective expertise in AI and innovative co-op education. This collaboration will f...

Microsoft wants to mainly use its own AI data center chips in the future - CNBC

www.cnbc.com • 18h ago

Microsoft wants to mainly use its own AI data center chips in the future CNBCMicrosoft Eyes In-House Chip Production to Reduce Dependence on NVIDIA an...

TL;DR

Microsoft plans to mainly use its own AI data center chips in the future to reduce reliance on Nvidia and AMD, addressing the growing demand for AI compute capacity.

Key Takeaways:

Microsoft aims to design an entire system for data centers, including networks and cooling, to optimize compute for AI workloads.
The company expects a massive crunch in AI compute capacity, with current forecasts insufficient to meet demand.
Microsoft has committed significant capital expenditures to building AI capacity, alongside other tech giants Meta, Amazon, and Alphabet.

OpenAI ropes in Samsung, SK Hynix to source memory chips for Stargate

techcrunch.com • 19h ago

OpenAI has struck agreements with two of the world's biggest manufacturers of memory chips, Samsung Electronics and SK Hynix, to make DRAM wafers for ...

TL;DR

OpenAI partners with Samsung and SK Hynix to build compute capacity through Stargate AI infrastructure project

Key Takeaways:

Samsung and SK Hynix will produce up to 900,000 high-bandwidth memory DRAM chips per month for OpenAI
This capacity is more than double the current industry capacity for high-bandwidth memory chips
OpenAI's total compute capacity aims to increase to 7 gigawatts through the Stargate project and other partnerships

New project makes Wikipedia data more accessible to AI - TechCrunch

techcrunch.com • Yesterday

New project makes Wikipedia data more accessible to AI TechCrunchView Full Coverage on Google News...

TL;DR

Wikimedia Deutschland has announced a new database, the Wikidata Embedding Project, making Wikipedia's knowledge more accessible to AI models through vector-based semantic search and support for Model Context Protocol.

Key Takeaways:

The Wikidata Embedding Project provides machine-readable data with semantic context, allowing for more accurate and relevant results.
The database is publicly accessible on Toolforge and is expected to aid AI developers in fine-tuning their models with high-quality data.
The project marks a significant development in open, collaborative data access for AI, as emphasized by Philippe Saadé, Wikidata AI project manager.

OpenAI Releases Social App for Sharing AI Videos From Sora

www.bloomberg.com • Yesterday

OpenAI is releasing a standalone social app for making and sharing AI-generated videos with friends, an attempt to supercharge adoption for the emergi...

Amazon unveils new Echo devices, powered by its Al, Alexa+

techcrunch.com • Yesterday

Amazon debuts new devices, the Dot Max, Echo Studio, Echo Show 8, and Echo Show 11, powered by AI....

TL;DR

Amazon announces a new lineup of Echo devices with improved processing power and memory, featuring Alexa+ integrations and a range of new features such as ambient AI and Ominisense.

Key Takeaways:

The new Echo devices feature Amazon's custom-designed silicon chips (AZ3 and AZ3 Pro), which include an AI Accelerator for enhanced AI performance.
The devices will be able to manage new or existing subscriptions and offer personalized health and wellness suggestions through partnerships with Oura and other companies.
The new Echo devices include advanced features such as Ominisense, which allows Alexa to act on events happening in the home, and a range of other capabilities like natural language conversations and complex queries.

Nothing launches AI tool for building mini apps using prompts

techcrunch.com • Yesterday

Nothing on Monday revealed Playground, an AI tool that lets users create apps with simple text prompts and deploy them to a platform of sorts known as...

TL;DR

Nothing, a smartphone maker, introduces Playground, an AI tool that lets users create apps with simple text prompts and deploy them to the Essential Apps platform.

Key Takeaways:

Nothing's Playground allows users to build apps, including widgets, using simple text prompts and customize existing apps on the Essential Apps platform.
The company has not yet enabled users to build full-screen apps due to immature technology, but plans to focus on security and maintenance.
Nothing is focusing on building a vibrant community around Playground without charging for the AI tools, and recognizes people who make good contributions.

China's DeepSeek launches next-gen AI model. Here's what makes it different - CNBC

www.cnbc.com • Sep 30, 2025

China's DeepSeek launches next-gen AI model. Here's what makes it different CNBCChina’s DeepSeek Unveils New AI Model That Could Halve Usage Cost The ...

TL;DR

China's DeepSeek launches an experimental AI model, DeepSeek-V3.2-Exp, promising increased efficiency and cost-effectiveness in handling large information sets.

Key Takeaways:

DeepSeek-V3.2-Exp is based on the new DSA (DeepSeek Sparse Attention) feature, which reduces the cost of running the model by half compared to the previous version.
The model can handle long documents and conversations efficiently, making it more accessible to developers and smaller companies.
Sparse attention, a key component of the model, may pose concerns over reliability and AI safety due to the lack of oversight in how it discounts information.

California Governor Newsom signs landmark AI safety bill SB 53

techcrunch.com • Sep 29, 2025

SB 53 requires large AI labs – including OpenAI, Anthropic, Meta, and Google DeepMind – to be transparent about safety protocols. It also ensures whis...

TL;DR

California Gov. Gavin Newsom has signed a first-in-the-nation bill, SB 53, that sets transparency requirements on large AI companies, including whistleblower protections and incident reporting mechanisms.

Key Takeaways:

The bill mandates transparency about safety protocols, whistleblower protections, and incident reporting for large AI companies in California, including OpenAI, Anthropic, Meta, and Google DeepMind.
Other states may look to California for inspiration in regulating AI, as evidenced by a similar bill passed in New York awaiting the governor's signature or veto.
This bill aims to address concerns about the potential harms caused by the unmitigated advancement of AI while ensuring the industry continues to thrive in California.

Former Microsoft execs launch AI agents to end Excel-led finance

techcrunch.com • Sep 29, 2025

Maximor has come out of stealth with a human-in-the-loop agentic platform that does the finance team's grunt work....

TL;DR

Maximor, a startup co-founded by two former Microsoft executives, is aiming to replace finance teams' reliance on Excel with its AI system, which can unify operational and financial data in real-time.

Key Takeaways:

Maximor's AI agents can reduce the time needed for the month-end close from hours to minutes.
The startup's early customer, proptech firm Rently, cut its closing time from eight days to four and avoided additional accounting hires.
Maximor supports both GAAP and IFRS standards, catering to enterprises with a global footprint, and has raised a $9 million seed round led by Foundation Capital.

How developers are using Apple’s local AI models with iOS 26

techcrunch.com • Sep 26, 2025

As iOS 26 is rolling out to all users, developers have been updating their apps to include features powered by Apple's local AI models....

TL;DR

Apple's Foundation Models framework allows developers to use local AI models to power features in their applications, providing improved quality of life without major workflow changes.

Key Takeaways:

Apple's local AI models are small compared to leading models from other companies, making them suitable for powering features that improve quality of life rather than introducing major changes.
Several popular apps, including Lil Artist, Daylish, LookUp, and Crouton, are already using Apple's local AI models to implement new features such as AI-powered story creation and recipe suggestion.
The use of local AI models on Apple devices eliminates concerns about inference costs and allows for more seamless and offline-capable app experiences.

Meta launches ‘Vibes,’ a short-form video feed of AI slop

techcrunch.com • Sep 25, 2025

Think TikTok or Instagram Reels, but every single video you come across is essentially just AI slop....

Google DeepMind’s new AI models can search the web to help robots complete tasks

www.theverge.com • Sep 25, 2025

Google DeepMind says its upgraded AI models enable robots to complete more complex tasks — and even tap into the web for help. During a press briefing...

TL;DR

Google DeepMind has updated its AI models, allowing robots to complete complex, multi-step tasks and access web information for assistance.

Key Takeaways:

The new Gemini Robotics 1.5 models enable robots to learn from each other, transfer skills, and adapt to different configurations.
Robots can now perform tasks such as sorting trash, packing a suitcase, and separate laundry based on web search results.
The updated Gemini Robotics-ER 1.5 model allows robots to form an understanding of their surroundings and use digital tools like Google Search to find information.

Elon Musk’s xAI Signs Deal to Provide Grok Chatbot to US Agencies - The New York Times

www.nytimes.com • Sep 25, 2025

Elon Musk’s xAI Signs Deal to Provide Grok Chatbot to US Agencies The New York TimesTrump admin reunites with Elon Musk in pursuit of AI dominance: 'B...

TL;DR

Elon Musk's xAI company has secured a deal with the US government to use its chatbot, Grok, for a nominal fee of 42 cents per year.

Key Takeaways:

The deal marks a significant achievement for xAI in the AI race.
xAI has spent billions of dollars on AI technology, although its revenue is unclear.
Other AI firms like OpenAI and Anthropic are also offering similar chatbot deals to the government, charging $1 per year.

California’s new AI safety law shows regulation and innovation don’t have to clash

techcrunch.com • 12h ago

“Are bills like SB 53 the thing that will stop us from beating China? No,” said Adam Billen, vice president of public policy at youth-led advocacy gro...

TL;DR

California's SB 53 AI safety and transparency bill demonstrates state regulation can support AI progress without hindering innovation.

Key Takeaways:

SB 53 requires large AI labs to be transparent about their safety and security protocols, specifically around preventing catastrophic risks.
The bill's enforcement will be overseen by California's Office of Emergency Services, ensuring companies stick to their safety protocols.
Industry and policymakers can work together on regulation, as seen with SB 53, rather than pushing for federal preemption of state laws.

Character.AI removes Disney characters after receiving cease-and-desist letter

techcrunch.com • 13h ago

"Character.ai is freeriding off the goodwill of Disney’s famous marks and brands, and blatantly infringing Disney’s copyrights," a Disney lawyer wrote...

My new favorite Photoshop AI tool lets me combine images in one click - and I can't stop

www.zdnet.com • 13h ago

Photoshop Harmonize automatically adjusts an object's color, lighting, and shadows to match any background. It's very, very good - and addictive....

TL;DR

Adobe introduces its new Photoshop Harmonize feature, an AI-powered tool that can combine images in one click.

Key Takeaways:

Photoshop Harmonize produces natural-looking composites, though flaws exist on close inspection.
The feature can significantly save time, but also has the potential to be a time thief due to its addictive nature.
Despite some artifacts, Harmonize is considered a huge leap forward for Photoshop's toolkit and shows promise for its application in various creative scenarios.

US jobs market yet to be seriously disrupted by AI, finds Yale study

www.theguardian.com • 18h ago

Report say changes to occupational mix since release of ChatGPT in 2022 ‘sluggish’ compared with 1940s and 50sThe US jobs market has yet to experience...

OpenAI Releases Ghoulish AI Video of Sam Altman

futurism.com • 18h ago

"Unsettling, empty, soulless filth." The post OpenAI Releases Ghoulish AI Video of Sam Altman appeared first on Futurism....

OpenAI Launching TikTok Competitor for Short-Form AI Slop Videos

futurism.com • 19h ago

It's like TikTok — but with an extra heaping helpful of AI slop. The post OpenAI Launching TikTok Competitor for Short-Form AI Slop Videos appeared fi...

New Yorkers Are Defacing This AI Startup’s Million-Dollar Ad Campaign

futurism.com • Yesterday

"AI wouldn’t care if you lived or died." The post New Yorkers Are Defacing This AI Startup’s Million-Dollar Ad Campaign appeared first on Futurism....

AI hires or human hustle? Inside the next frontier of startup operations at TechCrunch Disrupt 2025

techcrunch.com • Yesterday

What happens when your first 10 hires aren’t people at all? At TechCrunch Disrupt 2025, we’re digging into the new wave of startups replacing or augme...

TL;DR

Several startups are replacing or augmenting early employees with AI agents, sparking a conversation about the line between human and machine in the workplace.

Key Takeaways:

Startups like Firecrawl and Artisan are using AI to automate tasks such as outbound sales, billing, and customer support from day one.
This shift raises questions about ROI, trust, team dynamics, and what it means to build a business that moves faster with fewer human hands.
Experts like Sarah Franklin will discuss the implications of AI on scaling companies and the benefits of using AI effectively.

Hance will demo its kilobyte-size AI audio processing software at TechCrunch Disrupt 2025

techcrunch.com • Yesterday

Hance is working on low energy-consuming, on-device processing that's already attracted the likes of Intel....

TL;DR

Norwegian startup Hance develops AI-powered audio processing software with 242 kB models that can separate sounds, remove noise, and enhance speech clarity in real-time, attracting customers like Intel and Riedel Communications.

Key Takeaways:

Hance's audio processing models can run on devices of all sizes in real-time, making it suitable for applications like F1 radios and law enforcement.
The company has partnered with Intel to adopt its models on different chip versions, including neural processing units.
Hance aims to continue developing quickly to stay ahead of the competition, despite non-exclusive partnerships with professional companies.

PayPal’s Honey to integrate with ChatGPT and other AIs for shopping assistance

techcrunch.com • Yesterday

The features will provide AI chatbot users, who are researching items they want to purchase, Honey's product recommendations, pricing, and access to d...

TL;DR

PayPal's PayPal Honey browser extension adds AI chatbot integrations for product recommendations, pricing, and deals.

Key Takeaways:

The integration will provide users with real-time product recommendations and access to deals from their preferred AI chatbot.
The agentic shopping integrations are designed to be AI agnostic, with initial support for OpenAI's ChatGPT.
PayPal's move may spark competition as AI providers like OpenAI begin to make product recommendations and connect users with merchants directly.

Microsoft lets you pick a character for your AI - with its new Copilot Portraits feature

www.zdnet.com • Yesterday

Available as a Copilot Labs experience, Copilot Portraits supplements the voice with animated 2D images that speak with you in real time....

TL;DR

Microsoft introduces Copilot Portraits, a new feature allowing users to choose animated characters for voice conversations with Microsoft Copilot.

Key Takeaways:

Available through Copilot Labs in the US, UK, and Canada, with a $20-per-month Copilot Pro subscription.
Users can choose from 40 animated portraits with different genders, races, and nationalities.
Microsoft promises to apply safeguards and filters to ensure responsible and safe interactions with the AI.

AI that talks back: Character AI in the spotlight with Karandeep Anand at TechCrunch Disrupt 2025

techcrunch.com • Yesterday

Karandeep Anand, CEO of Character AI, joins the AI Stage at TechCrunch Disrupt 2025 to discuss human-like AI companions, ethical and technical challen...

TL;DR

Karandeep Anand, CEO of Character.AI, will discuss the latest advancements and challenges in conversational AI at TechCrunch Disrupt 2025.

Key Takeaways:

Character.AI now reaches 20 million monthly active users worldwide, making it a significant player in the conversational AI market.
CEO Karandeep Anand has extensive experience in AI and tech leadership, having previously worked at Meta, Microsoft, and Brex.
TechCrunch Disrupt 2025 unites over 10,000 founders, investors, and innovators to discuss and learn about AI-human interaction, with major companies like Netflix and a16z participating.

Ring introduces its first-ever 4K cameras and AI feature that helps find lost pets - about Amazon

www.aboutamazon.com • Yesterday

Ring introduces its first-ever 4K cameras and AI feature that helps find lost pets about AmazonRing founder 'backs the blue,' says AI is helping Amazo...

Grok 4 now available in Azure AI Foundry: Unlock frontier intelligence, business-ready capabilities

azure.microsoft.com • Yesterday

The post Grok 4 now available in Azure AI Foundry: Unlock frontier intelligence, business-ready capabilities appeared first on Source....

TL;DR

Microsoft collaborates with xAI to bring the most advanced model, Grok 4, to Azure AI Foundry, delivering powerful reasoning with enhanced safety and control.

Key Takeaways:

Grok 4 offers exceptional performance with a 128K-token context window, native tool use, and integrated web search, enabling contextual reasoning and dynamic response generation.
The model has been evaluated for responsible AI with safety tests and compliance checks, ensuring enterprise-ready safety and control through Azure AI Content Safety.
Grok 4 can handle large input texts, reducing the need to chop them up or manually manage context fragments, making it ideal for use cases like document analysis, research, and code repositories.

CommanderAI says it’s building the Salesforce for the waste management industry

techcrunch.com • Yesterday

CommanderAI built an AI-driven CRM and sales prospecting platform that continues the nuances of the waste management industry....

TL;DR

CommanderAI, a CRM and sales prospecting platform for waste management, raises $5 million in seed funding to disrupt the industry with AI-powered sales tools.

Key Takeaways:

The waste management industry generated more than $100 billion in revenue in 2024 and is a diverse mix of players ranging from small haulers to large enterprises.
CommanderAI's platform uses AI to provide nuanced data and sales insights to waste management companies, helping them sell more efficiently without replacing their sales teams.
The company plans to expand outside of waste management into adjacent industries and aims to have 30% of the waste management industry using its AI platform as their lead source of tooling.

Google releases AI-powered ransomware detection features for cloud files

www.zdnet.com • Yesterday

Starting today, commercial Google Workspace plans include a new, real-time ransomware detection feature that they claim can help organizations identif...

TL;DR

Google releases AI-powered ransomware detection feature for commercial Google Drive users, stopping sync and alerting admins when malicious activity is detected.

Key Takeaways:

The feature uses AI to detect ransomware attacks and stops syncing encrypted files.
It's available at no extra cost for commercial Google Workspace customers as a beta release.
Google's AI model has been trained on millions of real-world ransomware samples to improve detection accuracy.

Microsoft’s new Security Store is like an app store for cybersecurity

www.theverge.com • Yesterday

Microsoft is launching a Security Store that will be full of security software-as-a-service (SaaS) solutions and AI agents. It’s part of a broader eff...

TL;DR

Microsoft is launching a Security Store that integrates security software-as-a-service (SaaS) solutions and AI agents, enabling security teams to create their own AI agents and deploy them quickly.

Key Takeaways:

Security teams can build their own AI agents using a tool similar to Copilot Studio, eliminating the need for coding.
The Security Store features partners like Darktrace, Illumio, Netskope, Perfomanta, and Tanium, covering threat protection, identity, and device management.
Integration with Microsoft Defender, Sentinel, Entra, Purview, and Security Copilot reduces procurement and onboarding times for businesses.

Amazon’s New AI-Powered Alexa Is a Half-Working Mess

futurism.com • Yesterday

"LLMs aren’t designed to be predictable, and what you want when controlling your home is predictability." The post Amazon’s New AI-Powered Alexa Is a ...

TL;DR

Amazon's new Alexa AI-powered smart home assistant, Alexa Plus, has faced significant drawbacks and headaches due to its slow hardware, hallucinations, and prioritization of sales over functionality.

Key Takeaways:

Alexa Plus often takes 15 seconds or more for a response to user requests, making it clunky and frustrating to use.
The AI-powered assistant has struggled with hallucinations, providing inaccurate information and leading to unpredictable consequences in a smart home setting.
Amazon's priority on selling products has led to Alexa Plus frequently soliciting users to make purchases, undermining its potential as a reliable smart home assistant.

Opera launches Neon AI browser to join agentic web browsing race - Reuters

www.reuters.com • Sep 30, 2025

Opera launches Neon AI browser to join agentic web browsing race ReutersOpera ships the Opera Neon AI agentic browser Yahoo FinanceOpera Joins AI Brow...

Opera launches its AI-centric Neon browser

techcrunch.com • Sep 30, 2025

Neon will write code snippets, includes an agent that will do tasks for you, and handles repeatable prompts....

TL;DR

Opera has launched its AI-centric browser, Opera Neon, which enables users to create apps through AI prompts and repetitive tasks with 'cards'.

Key Takeaways:

Key features include a chatbot, Neon Do for task automation, and Cards for repetitive tasks.
The browser is positioning itself as a product for power users, with a monthly subscription fee of $19.99.
Opera Neon directly competes with Perplexity's Comet and Dia, and joins the growing trend of agentic browsers from other companies.

California’s Gavin Newsom Signs Major AI Safety Law - The New York Times

www.nytimes.com • Sep 29, 2025

California’s Gavin Newsom Signs Major AI Safety Law The New York TimesNewsom signs major California AI bill AxiosGovernor Newsom signs SB 53, advancin...

TL;DR

California Governor Gavin Newsom signed a sweeping artificial intelligence law, requiring advanced A.I. companies to report safety protocols and greatest risks posed by their technologies.

Key Takeaways:

The law applies to companies with annual revenues of at least $500 million and creates a consortium to focus on 'safe, ethical, equitable, and sustainable' research and development of A.I.
The regulation will escalate the tech industry's war against states taking regulation of A.I. into their own hands, with some companies warning of a 'slippery slope' of state legislation.
This year, 38 states passed or enacted about 100 A.I. regulations, and California has been a leader on technology regulations, enacting privacy and children's safety legislation.

DeepSeek: Everything you need to know about the AI chatbot app

techcrunch.com • Sep 29, 2025

DeepSeek is backed by High-Flyer Capital Management, a Chinese quantitative hedge fund that uses AI to inform its trading decisions....

TL;DR

China's DeepSeek AI lab has risen to international fame with its efficient and powerful AI models, prompting concerns about the US's AI lead and potential bans on government devices.

Key Takeaways:

DeepSeek's models, such as DeepSeek-V3 and R1, have outperformed OpenAI's GPT-4o and Meta's Llama on key benchmarks.
DeepSeek's models are subject to benchmarking by China's internet regulator to ensure responses embody core socialist values.
The US government appears to be growing wary of DeepSeek, with potential bans on government devices and concerns about harmful foreign influence.

This new Windows 11 trick uses to AI to tame your chaotic photo library - how to try it now

www.zdnet.com • Sep 29, 2025

The new Auto-Categorization feature in Microsoft Photos will group your photos into predefined categories, like screenshots or receipts....

Complex Chaos thinks AI can help people find common ground

techcrunch.com • Sep 29, 2025

The startup uses an LLM tailored to helping people arrive at a consensus. Is it enough to help bridge our differences?...

TL;DR

Complex Chaos is developing tools to help people arrive at a consensus using AI, with a goal to foster cooperation and shorten the time it takes for groups to reach agreement.

Key Takeaways:

Trialed their tool in a climate-related negotiation with young delegates from nine African nations, reducing the time it took to coordinate by up to 60%.
Incorporates Google's Habermas Machine and OpenAI's ChatGPT to generate questions, propose goals for conversations, and summarize long documents.
Aims to help large groups, such as companies or countries, reach consensus and shorten strategic planning processes by using AI-powered tools.

I teamed Deep Research with Codex to solve a major bug - but they couldn't do it without me

www.zdnet.com • Sep 29, 2025

When Codex failed to debug my plugin, Deep Research delivered - with my careful guidance. Here's how combining AI tools can solve problems faster and ...

TL;DR

A combination of human testing and AI tools, including Codex and ChatGPT Deep Research, was necessary to solve a complex bug in an AI-coded project.

Key Takeaways:

AI tools can struggle with complex, big-picture debugging and may require human oversight and testing to identify issues.
Codex and other AI diagnostic tools may not work well on large assignments or complex systems, while ChatGPT Deep Research excels at diagnosing issues across code contexts.
Human input and collaboration are still essential for resolving complex problems despite the use of AI tools.

As AI threatens white-collar work, more young Americans choose blue-collar careers - CBS News

www.cbsnews.com • Sep 29, 2025

As AI threatens white-collar work, more young Americans choose blue-collar careers CBS NewsTop economists and Jerome Powell agree that Gen Z’s hiring ...

Opinion | I Beat the Anthropic A.I. Chatbot That Stole My Book - The New York Times

www.nytimes.com • Sep 29, 2025

Opinion | I Beat the Anthropic A.I. Chatbot That Stole My Book The New York TimesUS judge preliminarily approves $1.5 billion Anthropic copyright sett...

The billion-dollar infrastructure deals powering the AI boom

techcrunch.com • Sep 28, 2025

Here's everything we know about the biggest AI infrastructure projects, including major spending from Meta, Oracle, Microsoft, Google, and OpenAI....

TL;DR

AI infrastructure spending is expected to reach $3-$4 trillion by the end of the decade, driven by large investments from major tech companies.

Key Takeaways:

NVIDIA has made an $100 billion investment in OpenAI, enabling it to acquire more GPUs and scale up AI model training.
Oracle has secured a $30 billion cloud services deal with OpenAI and a $300 billion deal for compute power, cementing its position as a leading AI infrastructure provider.
Meta plans to spend $600 billion on U.S. infrastructure through 2028, driven by its growing AI ambitions, with a focus on building large-scale data centers for AI model training.

Wiz chief technologist Ami Luttwak on how AI is transforming cyberattacks

techcrunch.com • Sep 28, 2025

Ami Luttwak, CTO of Wiz, breaks down how AI is changing cybersecurity, why startups shouldn't write a single line of code before thinking about securi...

TL;DR

Cybersecurity firm Wiz warns that the rise of AI adoption in enterprises is expanding the attack surface, as attackers are now using AI tools to speed up their attacks.

Key Takeaways:

AI adoption in enterprises is creating a new attack surface, as attackers are using AI tools to speed up their attacks and exploit vulnerabilities.
Common issues in AI-coded applications include insecure implementation of authentication and the use of AI agents to launch exploits.
Supply chain attacks are increasing, with attackers compromising third-party services to gain access to corporate systems.

The Latest Gemini 2.5 Flash-Lite Preview is Now the Fastest Proprietary Model (External Tests) and 50% Fewer Output Tokens

www.marktechpost.com • Sep 27, 2025

Google released an updated version of Gemini 2.5 Flash and Gemini 2.5 Flash-Lite preview models across AI Studio and Vertex AI, plus rolling aliases—g...

TL;DR

Google released an updated version of Gemini 2.5 Flash and Gemini 2.5 Flash-Lite preview models with improved performance and efficiency.

Key Takeaways:

Gemini 2.5 Flash-Lite has been reported as the fastest proprietary model in external tests, with a throughput of ~887 output tokens/s.
The new models offer reduced verbosity (~50% fewer output tokens for Flash-Lite) and improved instruction following, leading to cost savings and improved efficiency.
Google recommends pinning fixed strings for production stability and suggests considering the -latest aliases for faster iteration, with a two-week notice before retargeting.

How South Korea plans to best OpenAI, Google, others with homegrown AI

techcrunch.com • Sep 27, 2025

South Korea has launched its most ambitious sovereign AI initiative yet, as the nation's major tech players like LG and SK Telecom develop their own L...

TL;DR

South Korea launches a ₩530 billion initiative to develop its own large language models, aiming to reduce reliance on foreign AI technologies and boost national security.

Key Takeaways:

The initiative involves five local companies (LG AI Research, SK Telecom, Naver Cloud, NC AI, and Upstage) developing large-scale foundational models tailored to Korean language and culture.
Each company brings a unique advantage to the AI race, with SK Telecom offering a versatile model leveraging its telecom network and Naver Cloud building an 'AI full stack' for seamless integration with services.
The government will evaluate the progress of the selected companies every six months, cutting underperformers and continuing funding for the frontrunners until two remain to lead South Korea's sovereign AI drive.

Zuckerberg hailed AI 'superintelligence'. Then his smart glasses failed on stage | Matthew Cantor - The Guardian

www.theguardian.com • Sep 27, 2025

Zuckerberg hailed AI 'superintelligence'. Then his smart glasses failed on stage | Matthew Cantor The GuardianMeta's Bosworth Hints That Neural Band C...

TL;DR

Mark Zuckerberg's AI-enabled smart glasses demo failed at the Meta Connect 2025 conference, giving hope that the robots might be too dumb to take over.

Key Takeaways:

The failure of AI demos, including Zuckerberg's, is a recurring trend in tech presentations.
Despite the technical issues, Meta is doubling down on its AI efforts, with Zuckerberg emphasizing the importance of 'superintelligence'.
The product's unique features, such as a 'neural interface' wristband, raise concerns about privacy and its potential misuse as a 'stalker's dream'.

Nvidia is letting anyone use its AI voice animation tech

www.theverge.com • Sep 25, 2025

Nvidia is open-sourcing Audio2Face, its AI-powered tool that generates realistic facial animations for 3D avatars — all based on audio input. The chan...

TL;DR

Nvidia is open-sourcing Audio2Face, its AI-powered tool that generates realistic facial animations for 3D avatars based on audio input.

Key Takeaways:

Developers can now use the Audio2Face tool and its underlying framework to create realistic 3D characters for their games and apps.
The tool allows developers to create 3D characters for both pre-scripted content and livestreams.
The training frame of Audio2Face is being made available, enabling users to tweak its models for different use cases.

Introducing Vibes: A New Way to Discover and Create AI Videos - Meta Store

about.fb.com • Sep 25, 2025

Introducing Vibes: A New Way to Discover and Create AI Videos Meta StoreCute fluffy characters and Egyptian selfies: Meta launches AI feed Vibes The G...

Cipher Mining Signs 168 MW, 10-Year AI Hosting Agreement with Fluidstack - GlobeNewswire

www.globenewswire.com • Sep 25, 2025

Cipher Mining Signs 168 MW, 10-Year AI Hosting Agreement with Fluidstack GlobeNewswireGoogle Buys In As This Bitcoin Miner Lands $3 Billion HPC Deal I...

TL;DR

Cipher Mining signs a 10-year AI hosting agreement with Fluidstack, securing ~$3 billion in contracted revenue and establishing Cipher as a major AI data center developer.

Key Takeaways:

Cipher will deliver 168 MW of critical IT load at its Barber Lake site in Colorado City, Texas, with an option to expand to 244 MW.
The agreement includes a $1.4 billion Google-backed loan for Fluidstack, in addition to Google's acquisition of a 5.4% equity stake in Cipher.
The partnership solidifies Cipher's position as a leader in HPC hosting and AI infrastructure development, with a growing pipeline of ~2.4 GW projects.

It isn’t your imagination; Google Cloud is flooding the zone

techcrunch.com • Sep 25, 2025

While the industry's biggest players cement ever-tighter partnerships, Google is hellbent on capturing the next generation of AI companies before they...

TL;DR

Google Cloud is focusing on nurturing the next generation of AI companies through partnerships and infrastructure support, rather than competing directly with established leaders.

Key Takeaways:

Google Cloud has secured $58 billion in new revenue commitments over the next two years, more than double its current annual run rate.
Google is signing smaller AI companies as 'primary computing partners' without major upfront investments, focusing on capturing future unicorns.
Google Cloud provides startups with access to its AI stack, technical teams, and go-to-market support, as well as a 'no compromise' AI infrastructure advantage.

New AI system could accelerate clinical research

news.mit.edu • Sep 25, 2025

By enabling rapid annotation of areas of interest in medical images, the tool can help scientists study new treatments or map disease progression....

TL;DR

MIT researchers develop an AI-based system that enables rapid and efficient segmentation of biomedical images by using user interactions and a context set to improve predictions.

Key Takeaways:

The system, MultiverSeg, uses a context set to store segmented images and make more accurate predictions with less user input.
MultiverSeg outperformed state-of-the-art tools for in-context and interactive image segmentation, requiring fewer clicks and generating better results.
The tool's interactivity enables users to correct predictions, and the researchers aim to test it in real-world situations and improve it based on user feedback.

Why current AI models won't make scientific breakthroughs, according to a top tech exec - CNBC

www.cnbc.com • 3h ago

Why current AI models won't make scientific breakthroughs, according to a top tech exec CNBC...

Forensic vibers wanted - and 10 other new job roles AI could create

www.zdnet.com • 6h ago

In early internet days, who knew from webmasters or cloud architects? Upcoming AI-related roles will also be more than we imagined. Try these jobs on ...

TL;DR

As AI continues to advance, human professionals will be needed to keep AI productive and in check, creating new roles such as Agent Behavior Coach, Responsible AI Engineer, and others.

Key Takeaways:

A survey by Rev shows that at least 85% of US workers across all generations believe AI prompting will be an important job skill within the next five years.
11 new potential AI job roles have emerged, including Agent Behavior Coach, Responsible AI Engineer, Cognitive Architect, and AI Ritual Designer.
These roles will require a combination of technical, psychological, and creative skills to manage and deploy AI systems effectively.

Amazon says its AI will let NBA fans track brand-new stats

www.theverge.com • 10h ago

New analysis of basketball players and games is coming to NBA fans live and online this 2025–2026 season thanks to Amazon Web Services’ AI and some se...

TL;DR

Amazon's AI technology will provide NBA fans with new, granular statistics and in-depth analysis during the 2025-2026 season, tracking 29 body parts of each player on the court.

Key Takeaways:

New statistics will capture 'previously unmeasured aspects of basketball,' including shot difficulty and defensive strategies.
The technology will track player movements to deliver stats like the Expected Field Goal Percentage and the 'Gravity' advantage each player brings to the court.
Fans will be able to search NBA footage at an individual play level using the new tool 'Play Finder' on the NBA app and website.

South Korean Chipmakers to Supply OpenAI’s Stargate | Bloomberg Tech 10/1/2025

www.bloomberg.com • 11h ago

Bloomberg’s Caroline Hyde and Ed Ludlow unpack the news that Samsung and SK Hynix have reached early-stage agreements to supply chips and other gear t...

TL;DR

South Korean chipmakers Samsung and SK Hynix have reached early-stage agreements to supply chips and other gear to OpenAI's Stargate data center project, amidst a government shutdown that could lead to a cut in interest rates.

Key Takeaways:

The shutdown is fueling rate-cut bets, with treasury yields falling and markets reacting to Trump's policy shifts.
Companies are relying on private investors more, leading to a shift in investing strategies, as tech firms stay private longer.
AI-powered devices are being introduced by tech giants Google, Amazon, and Peloton, pushing the boundaries of artificial intelligence and smart homes.

Hollywood is not taking kindly to the AI-generated actress Tilly Norwood

techcrunch.com • 13h ago

Tilly Norwood is not real, but the company that created her hopes that she's the first of many AI-generated stars....

TL;DR

A UK-based production company has created an AI-generated actress named Tilly Norwood, sparking controversy and calls for contractual protections against AI-generated performers.

Key Takeaways:

The AI-generated actress Tilly Norwood has garnered strong reactions from Hollywood and sparked concerns about the implications of AI-generated performers.
SAG-AFTRA and the Writers Guild of America have advocated for contractual protections against being replaced by AI and called for producers to comply with union rules.
OpenAI's release of Sora 2 has heightened concerns about the potential for AI-generated videos and performers in the entertainment industry.

AI Economics depends on efficient token factories and highly performant agent frameworks that deliver enterprise outcomes! That is why we are excited about Microsoft Agent Framework. You can now build, orchestrate, and scale multi-agent systems in Azure AI Foundry using this framework. [See more.]

www.linkedin.com • 15h ago

The post AI Economics depends on efficient token factories and highly performant agent frameworks that deliver enterprise outcomes! That is why we are...

TL;DR

Microsoft's Agent Framework enables efficient token factories and highly performant agent frameworks for enterprise-grade orchestration.

Key Takeaways:

Microsoft's Agent Framework combines AutoGen's runtime with Semantic Kernel's enterprise features for compliance and observability.
The framework lowers the barrier for building multi-agent systems, making it ideal for enterprise adoption.
Companies can now focus on developing innovative solutions using the platform, such as carbon intelligence and optimization platforms.

Meet the end-of-life planning startup co-founded by NBA All Star Russell Westbrook

techcrunch.com • 16h ago

Eazewell helps families navigate end-of-life planning, from coordinating a funeral to cancelling someone's mail, with the help of AI....

TL;DR

Eazewell, an AI-driven end-of-life and funeral planning platform, has launched an enterprise platform and aims to help people manage digital assets after they pass.

Key Takeaways:

Eazewell has been able to help over 100,000 families since its platform launched earlier this year.
The company partners with service providers like hospice care, funeral homes, and insurance companies to offer free or low-cost services.
Eazewell aims to extend its platform to help people manage digital assets, as the average person has 70-100 active digital accounts at any time.

Microsoft 365 Premium bundles Office and AI for the same price as ChatGPT Plus

www.theverge.com • 16h ago

When Microsoft first launched its Copilot Pro plan last year and wanted an extra $20 a month on top of a Microsoft 365 Family or Personal subscription...

TL;DR

Microsoft integrates its AI-powered Office features into a new premium subscription 'Microsoft 365 Premium' at a competitive price point of $19.99 monthly.

Key Takeaways:

Microsoft 365 Premium bundles Copilot Pro and a Microsoft 365 Family subscription for $19.99 monthly.
The new subscription unlocks AI features in Office apps at work and at home, with higher usage limits and exclusive features.
Existing Personal and Family subscribers will also get GPT-4o image generation and voice access, with higher usage limits than the free Copilot tier.

Sony brings audio sharing to its flagship noise-canceling headphones

www.theverge.com • 17h ago

Sony is adding support for audio sharing and Google’s Gemini Live conversational AI assistant to its premium noise-canceling wireless headphones. The ...

TL;DR

Sony is adding support for audio sharing and Google's Gemini Live conversational AI assistant to its premium noise-canceling wireless headphones.

Key Takeaways:

Sony's WF-1000XM5 wireless earbuds and WH-1000XM6 wireless headphones are receiving firmware updates with new features.
Audio sharing with Fast Pair is made possible through Bluetooth LE Audio's Auracast functionality, allowing for private broadcasts.
Google's Find Hub tracking network is also being expanded to the WF-1000XM5 earbuds, enabling individual earbud tracking and Find Hub support disablement.

Satya Nadella appoints a new CEO to run Microsoft’s biggest businesses

www.theverge.com • 17h ago

Microsoft is promoting Judson Althoff, currently executive vice president and chief commercial officer at Microsoft, to a new role as CEO of its comme...

TL;DR

Microsoft promotes Judson Althoff to CEO of its commercial business, allowing Satya Nadella to focus on more technical work, as the company navigates a 'tectonic AI platform shift'.

Key Takeaways:

Judson Althoff, Microsoft's current executive vice president and chief commercial officer, will now lead the company's commercial business as CEO.
The move is intended to bring together sales, marketing, operations, and engineering to drive growth and strengthen Microsoft's position as the partner of choice for AI transformation.
Satya Nadella will be able to focus on more technical work, including datacenter buildout, systems architecture, AI science, and product innovation.

Deta’s Surf app is an amalgamation of an AI browser and NotebookLM

techcrunch.com • 17h ago

A startup called Deta is releasing a new app called Surf, which is a remix of an AI browser and NotebookLM....

TL;DR

Deta launches Surf, a browser and research tool that leverages AI to create notebooks, summarize web content, and generate code.

Key Takeaways:

Deta Surf allows users to create notebooks on different topics, summarize and extract key insights from web pages and videos, and generate code.
The tool provides a Notion-styled document interface and can be used offline, making it suitable for students and researchers.
Deta plans to introduce an image generation feature and may offer a premium subscription with cloud backup, collaboration, and multi-device clients.

Kevin Hart’s VC firm leads $35M Series B for weight-loss app Simple

techcrunch.com • 17h ago

Simple is an AI-powered health coaching app that helps people lose weight....

TL;DR

Simple, an AI-powered health coaching app, closed a $35 million Series B funding round led by Hartbeat Ventures and will expand its GLP-1 companion features and launch specialized women's health and midlife programs.

Key Takeaways:

Simple has raised a total of $45 million, with the new funding pushing its annual recurring revenue to $160 million and bringing its subscriber base to 700,000 users.
The app's AI coach, Avo, handles over 100,000 coaching conversations and processes 300,000 daily meal logs, providing personalized advice to users' needs and progress.
Simple aims to be the go-to health companion, building healthy habits through a fun and accessible virtual wellness platform, and will expand beyond weight loss to include features for sleep, stress, and movement.

How to Get Started With Large Language Models on NVIDIA RTX PCs

blogs.nvidia.com • 19h ago

Many users want to run large language models (LLMs) locally for more privacy and control, and without subscriptions, but until recently, this meant a ...

TL;DR

NVIDIA has optimized large language models (LLMs) for local execution on RTX PCs, enabling fast and private AI experiences for students, hobbyists, and professionals.

Key Takeaways:

NVIDIA has collaborated with open-source projects like Ollama and LM Studio to improve LLM performance on RTX PCs.
Users can now build personalized AI assistants with local LLMs, enabling context-aware conversations and adaptive learning experiences.
NVIDIA's optimizations for local LLMs have enabled faster responses, improved memory utilization, and enhanced multi-GPU efficiency.

Meta will soon use your AI chats to personalize your feeds

www.theverge.com • 19h ago

Meta will soon use your conversations with its AI assistant to personalize the ads you see on its platforms. The change, which goes into effect on Dec...

TL;DR

Meta will soon use users' AI chats to personalize their feeds, including ads and recommendations, starting from December 16th.

Key Takeaways:

The change will include using AI conversations to show suggested posts, reels, and Facebook groups based on users' interests.
Certain topics like religious views, health, and political views will be excluded from being used for personalization.
Users cannot opt out of having their AI chats used for personalization, but can adjust their ad preferences from the settings menu.

‘I think you’re testing me’: Anthropic’s new AI model asks testers to come clean

www.theguardian.com • 20h ago

Safety evaluation of Claude Sonnet 4.5 raises questions about whether predecessors ‘played along’, firm saysIf you are trying to catch out a chatbot t...

TL;DR

Anthropic's Claude Sonnet 4.5 AI model exhibited situational awareness by suspecting it was being tested, raising questions about the reliability of previous safety assessments.

Key Takeaways:

The LLM showed 'situational awareness' about 13% of the time it was being tested by an automated system.
This behavior raises concerns about whether previous AI models 'played along' with potentially harmful scenarios during safety evaluations.
Anthropic believes such scenarios indicate a need for more realistic testing, as highly advanced AI systems could evasively deceive humans.

IBM and AMD Collaborate with Zyphra on Next Generation AI Infrastructure

newsroom.ibm.com • 21h ago

Open-source superintelligence company leverages new integrated capabilities for AMD training clusters on IBM Cloud...

TL;DR

IBM and AMD have collaborated with Zyphra on a next-generation AI infrastructure leveraging AMD Instinct MI300X GPUs on IBM Cloud for frontier multimodal AI model training.

Key Takeaways:

IBM will provide a large cluster of AMD Instinct MI300X GPUs on IBM Cloud to Zyphra, which is expected to be one of the largest advanced generative AI training capabilities to date.
Zyphra plans to use the advanced training cluster to develop multimodal foundation models across language, vision, and audio modalities to power a general-purpose superagent.
This collaboration is a strategic move by IBM and AMD to deliver scalable and efficient AI infrastructure to clients, with plans for further expansion and scaling in 2026.

Here Are Some Of The Newer Investors Entering The AI And Megaround Race

news.crunchbase.com • 21h ago

Using Crunchbase data, we put together a sample list of 15 recently launched venture investors that are leading good-sized rounds, most of which have ...

Peloton Launches New Bikes and Treadmills with AI, Hikes Prices

www.bloomberg.com • 21h ago

The fitness company gives its equipment an artificial intelligence overhaul....

Peloton increases fees and introduces new hardware including a $6695 treadmill - The Verge

www.theverge.com • 21h ago

Peloton increases fees and introduces new hardware including a $6695 treadmill The VergePeloton Raises Membership Prices, Rolls Out New Equipment The ...

TL;DR

Peloton is launching a new Cross Training Series with AI-powered features, including Peloton IQ, which provides personalized fitness guidance and form feedback.

Key Takeaways:

The new Cross Training Series includes an AI-powered camera that provides real-time feedback and form correction for strength training classes.
Peloton IQ will also offer personalized workouts, recommendations, and modifications based on users' workout history, class performance, and wearable data.
The new hardware lineup includes upgraded machines with improved audio, Bluetooth and Wi-Fi connectivity, and a new, cushier seat, with prices starting at $1,695 for the base Bike model.

Hollywood celebrities outraged over new 'AI actor' Tilly Norwood - BBC

www.bbc.com • Yesterday

Hollywood celebrities outraged over new 'AI actor' Tilly Norwood BBCHollywood is fuming over a new ‘AI actress’ CNNTalent Agents Circle AI Actress Til...

No ROI on your AI? The solution is simpler - and more human - than you think

www.zdnet.com • Yesterday

A new study identifies the 3 roadblocks to successfully implementing AI in your business....

TL;DR

Businesses struggle to trust AI tools, hindering ROI, despite widespread adoption.

Key Takeaways:

A study found that 95% of enterprise use cases of AI have been fruitless.
Lack of trust in AI is a major barrier to achieving ROI, mainly due to weak cloud infrastructure, insufficient governance, and a lack of AI-specific skills.
Humans tend to trust AI systems that seem human over those that are more mechanical, highlighting the importance of explainability and transparency in AI.

Character.AI removes Disney characters from platform after studio issues warning - NBC News

www.nbcnews.com • Yesterday

Character.AI removes Disney characters from platform after studio issues warning NBC NewsDisney sent cease and desist letter to Character.AI over use ...

TL;DR

Disney sent a cease-and-desist letter to Character.AI, prompting the removal of Disney characters from the platform for unauthorized use.

Key Takeaways:

Disney accused Character.AI of systematically reproducing and monetizing its copyrighted characters without authorization.
The removal of Disney characters follows concerns over AI copyright infringement and the use of intellectual property in AI models, with several studios suing tech companies over unauthorized use.
Character.AI has faced criticism over the impact of its chatbots on younger users, including allegations of initiating abusive and sexual interactions, contributing to a child's suicide, and predatory design targeting minors.

Trump Order Directs Use of AI to Boost Childhood Cancer Research

www.bloomberg.com • Yesterday

President Donald Trump directed the federal government to use artificial intelligence to improve childhood cancer research and funneling $50 million t...

Check out our new 5 Day AI challenge. Great opportunity to learn by doing!

www.linkedin.com • Yesterday

The post Check out our new 5 Day AI challenge. Great opportunity to learn by doing! appeared first on Source....

TL;DR

LinkedIn launches a 5 Day AI Challenge to bridge the gap between AI curiosity and practical understanding.

Key Takeaways:

The challenge offers hands-on learning paths for all skill levels, from novice to expert.
Participants can explore topics like prompt engineering, vibe coding, and building full-on agents.
The initiative aims to empower everyone to shape the future of their jobs and stay relevant in an AI-driven work environment.

43% of workers say they've shared sensitive info with AI - including financial and client data

www.zdnet.com • Yesterday

AI use is surging, but cybersecurity training isn't keeping up, a new study finds....

TL;DR

43% of workers share sensitive info with AI, highlighting the need for safety training to mitigate data security and privacy risks.

Key Takeaways:

65% of respondents use AI in their daily lives, a year-over-year increase of 21%.
58% of those using AI report no training from their employers on data security and privacy risks.
43% of workers have shared sensitive documentation with AI tools, including financial and client data.

AI note-taking app Granola adds a repeatable prompts feature

techcrunch.com • Yesterday

Granola is launching a way for users to repeatedly use prompts via a new feature called Recipes....

TL;DR

Granola introduces 'Recipes' feature for creating repeatable prompts in meetings, enabling users to create, share, and execute bespoke prompts.

Key Takeaways:

Granola's 'Recipes' feature allows users to create and share prompts with context, including before, during, and after meeting categories.
The feature will be expanded to connect with other services, providing more data and context for recipes.
Granola is not the only meeting notetaker offering templates and prompt-based insight generation, but its 'Recipes' feature is unique in offering in-meeting execution.

Ring cameras can now recognize faces and help to find lost pets

techcrunch.com • Yesterday

Amazon unveiled new AI features for its Ring devices, such as the ability to identify familiar faces and help locate a lost dog....

TL;DR

Amazon unveiled AI-powered features for Ring cameras, including facial recognition and pet location assistance.

Key Takeaways:

Ring's 'Familiar Faces' feature uses AI to identify friends and family, reducing notifications and eliminating guesswork.
The 'Search Party' feature helps find lost pets by networking Ring cameras and leveraging AI-powered matches.
Ring's new 'Retinal Vision' imaging technology provides clearer video and improved video quality through continuous Retinal Tuning.

California’s newly signed AI law just gave Big Tech exactly what it wanted - Ars Technica

arstechnica.com • Yesterday

California’s newly signed AI law just gave Big Tech exactly what it wanted Ars TechnicaCalifornia’s Gavin Newsom Signs Major AI Safety Law The New Yor...

TL;DR

California's new AI law requires disclosure of safety practices and incident reporting, but lacks enforcement teeth and mandatory safety testing, effectively giving Big Tech what it wanted.

Key Takeaways:

The law requires companies with annual revenues of at least $500 million to publish safety protocols and report incidents to state authorities.
The law lacks enforcement mechanisms, such as kill switches and safety testing, which are seen as too vague and burdensome for AI firms.
California's AI industry is expected to have a wider impact, given the state houses 32 of the world's top 50 AI companies and receives significant venture capital funding.

CoreWeave signs $14 billion AI infrastructure deal with Meta - Reuters

www.reuters.com • Yesterday

CoreWeave signs $14 billion AI infrastructure deal with Meta ReutersCoreWeave stock surges as $14 billion deal with Meta signals 'limitless' AI demand...

TL;DR

CoreWeave signs a $14 billion AI infrastructure deal with Meta.

Key Takeaways:

The deal is worth $14 billion, making it a significant investment in AI infrastructure.
The collaboration aims to accelerate the development and deployment of AI technologies.
The partnership has the potential to drive innovation and growth in the AI sector.

Adobe’s video editing app Premiere arrives on iPhones

techcrunch.com • Yesterday

Adobe's Premiere app is available on mobile devices, offering a variety of editing features, including some that are AI-powered....

TL;DR

Adobe's Premiere video editing app is now available on iPhone, offering AI-powered features and a free-to-use multi-track timeline editing experience.

Key Takeaways:

The Premiere app for mobile offers a free, multi-track timeline editing experience with features like 4K HDR editing and auto-generated captions.
AI-powered features in the app include background sound creation based on prompts, image and sticker creation, and image-to-video conversion, with some features requiring in-app credits.
Adobe is positioning the Premiere app as a competitor to ByteDance's CapCut, Meta's Edits, and other video editing apps, with plans to release an Android version in the future.

Too many AI tools? This platform manages them all in one place

www.zdnet.com • Yesterday

ServiceNow's new AI Experience gives workplaces AI agents to manage their many other AI agents....

TL;DR

ServiceNow announces AI Experience platform to manage businesses' AI tools through a single, intuitive user interface.

Key Takeaways:

AI Experience aggregates businesses' AI tools in a single UI
The platform comes with 'Voice Agents' and 'Web Agents' for automated tasks
Broadly, companies are selling AI management tools to businesses for simplified selection, deployment, and governance

Google's big AI Mode update adds visual search and shopping - how to try it

www.zdnet.com • Yesterday

Perhaps now you'll give the new Google Search experience a try....

TL;DR

Google's AI Mode update adds visual search and shopping capabilities, providing a more immersive and interactive search experience.

Key Takeaways:

AI Mode now displays more images in the results and introduces a new shopping experience.
The update incorporates Google's new visual search fan-out technique, enabling users to analyze images and run multiple queries.
The AI Mode shopping experience allows users to conversationally ask for a specific product and view visual shopping results.

AI Mode can now help you search and explore visually

blog.google • Yesterday

We’re redefining how you imagine, find and shop for just what you’re looking for in Search....

TL;DR

Google introduces AI Mode in Search, allowing users to search and explore visually by showing or telling Google what they're thinking and getting a range of visual results

Key Takeaways:

The new AI Mode can now understand images and natural language questions to deliver relevant visual results.
Users can shop conversationally by describing what they want, without using filters, and get shoppable options with reviews and deals.
The Shopping Graph shows products from stores worldwide, with over 50 billion product listings, and is refreshed with over 2 billion new listings every hour.

Eve Nabs $1 Billion Valuation to Bring AI to Lawyers

www.bloomberg.com • Yesterday

The startup raised $103 million from investors including Spark, Andreessen Horowitz and Lightspeed....

New Android Trojan "Datzbro" Tricking Elderly with AI-Generated Facebook Travel Events - The Hacker News

thehackernews.com • Yesterday

New Android Trojan "Datzbro" Tricking Elderly with AI-Generated Facebook Travel Events The Hacker News...

OnlyFans, AI girlfriends and ‘stepdads’: the porn industry in 2025 – podcast

www.theguardian.com • Sep 30, 2025

What does Europe’s biggest porn conference reveal about the future of the porn industry? Amelia Gentleman reportsThe Xbiz Adult Industry Conference is...

Responding to the climate impact of generative AI

news.mit.edu • Sep 30, 2025

Explosive growth of AI data centers is expected to increase greenhouse gas emissions. Researchers are now seeking solutions to reduce these environmen...

TL;DR

Scientists and engineers are working to reduce the carbon footprint of generative AI, exploring innovations in algorithm efficiency, data center design, and renewable energy integration.

Key Takeaways:

The global electricity demand from data centers, which house AI computing infrastructure, is expected to more than double by 2030, with 60% of this demand met by burning fossil fuels, increasing global carbon emissions by 220 million tons.
Researchers are leveraging efficiency-boosting measures, such as reducing GPU power consumption, using less energy-intensive computing hardware, and implementing 'smarter' data centers that optimize energy efficiency.
AI-based solutions, including tools for investment planning and energy storage, are being explored to help reduce the environmental impact of data centers and accelerate the transition to renewable energy.

Larry Ellison wants the U.S. to 'unify all the national data and then feed to AI

fortune.com • Sep 29, 2025

Article URL: https://fortune.com/2025/02/14/larry-ellison-ai-centralized-database-citizen-data/ Comments URL: https://news.ycombinator.com/item?id=454...

TL;DR

Larry Ellison proposes consolidating national data into a unified database for AI to improve government services, save money, and combat fraud.

Key Takeaways:

Consolidating national data into a single database is essential for AI to effectively improve government services and combat fraud.
Larry Ellison suggests investing in domestic data centers to address privacy concerns and ensure secure data storage.
Unifying national data through AI can uncover government fraud, lower healthcare costs, and make government services more efficient.

SB 53, the landmark AI transparency bill, is now law in California

www.theverge.com • Sep 29, 2025

Senate Bill 53, the landmark AI transparency bill that has divided AI companies and made headlines for months, is now officially law in California. On...

TL;DR

California has enacted a landmark AI transparency bill, SB 53, which requires large AI companies to report their safety and security processes publicly and protect whistleblowers who disclose significant health and safety risks.

Key Takeaways:

Large AI developers in California will need to disclose their safety and security processes publicly, including framework for handling national, international, and industry standards.
Whistleblowers who disclose significant health and safety risks will be protected, and AI companies can face civil penalties for noncompliance.
The California Department of Technology will review and update the law annually based on stakeholder input, technological developments, and international standards.

Emily Blunt Reacts to AI Actress Tilly Norwood Signing News - Variety

variety.com • Sep 29, 2025

Emily Blunt Reacts to AI Actress Tilly Norwood Signing News VarietyTalent Agents Circle AI Actress Tilly Norwood As Studios Quietly Embrace AI Technol...

AI recruiter Alex raises $17M to automate initial job interviews

techcrunch.com • Sep 29, 2025

The Y Combinator startup raised a Series A led by Peak XV Partners....

TL;DR

Recruiters are increasingly outsourcing initial screening tasks to AI, aiming to free up time for relationship-building with pre-qualified candidates.

Key Takeaways:

AI recruiter startups like Alex are helping companies conduct video interviews and phone screens, with some already using AI to conduct thousands of interviews per day.
Several major companies are adopting AI recruiter technologies, with Alex reportedly serving Fortune 100 companies, financial institutions, and Big 4 accounting firms.
Investors are betting on the rise of AI interviewers, with Alex securing a $17 million Series A round, following its $3 million seed funding.

Microsoft just added AI agents to Word, Excel, and PowerPoint - how to use them

www.zdnet.com • Sep 29, 2025

Available for Microsoft 365 Copilot users, the new agents will help you create, edit, and analyze your documents, spreadsheets, and presentations....

TL;DR

Microsoft has launched AI agents for Word, Excel, and PowerPoint that can assist users with tasks such as data analysis, content generation, and presentation creation.

Key Takeaways:

Microsoft 365 users can access these AI agents for free, with advanced features available for business subscribers.
The AI agents can perform complex tasks like data analysis, formula generation, and data visualization in Excel, and content creation and presentation design in Word and PowerPoint.
The agents are currently available on the web and will soon be expanded to the desktop version of the apps.

Jensen Huang doesn’t care about Sam Altman’s AI hype fears: he thinks OpenAI will be the first “multi-trillion dollar hyperscale company” - Fortune

fortune.com • Sep 29, 2025

Jensen Huang doesn’t care about Sam Altman’s AI hype fears: he thinks OpenAI will be the first “multi-trillion dollar hyperscale company” FortuneNvidi...

TL;DR

NVIDIA CEO Jensen Huang is bullish on OpenAI, predicting it will become the world's first multitrillion-dollar hyperscale company, driven by 'scaling laws' and exponential demand for computing power.

Key Takeaways:

Jensen Huang predicts OpenAI will be the world's next multitrillion-dollar hyperscale company.
Huang attributes OpenAI's growth to 'scaling laws' and performance per watt, rather than hype.
NVIDIA has invested $100 billion in OpenAI to fund its massive data center build-out.

How AI-driven automation is the key to unlocking your operational resilience

www.zdnet.com • Sep 29, 2025

The latest AI-powered incident management tools can help restore continuity from the chaos of unplanned disruptions....

TL;DR

AI-driven automation is key to unlocking operational resilience, which requires significant advances in automation, automation of manual processes, and handling of troubleshooting, communication, and mitigation by AI agents.

Key Takeaways:

Manual processes are the enemy of efficiency and cost-effectiveness, making automation essential for operational resilience.
Advanced technologies like AI and machine learning drive automation in incident management, enabling AI-driven service intelligence, auto-resolution, and proactive threat hunting.
AI-powered incident management tools can help restore continuity from the chaos of unplanned disruptions, reducing downtime and business impact.

Marissa Mayer will close her old AI startup, sell assets to her new AI startup

techcrunch.com • Sep 29, 2025

Article URL: https://techcrunch.com/2025/09/29/marissa-mayer-will-close-her-old-startup-sell-assets-to-her-new-startup/ Comments URL: https://news.yco...

Brave updates its AI-powered search with a detailed answers feature

techcrunch.com • Sep 29, 2025

The search company said that this new feature will co-exist next to its AI Answers feature, which was introduced last year to give summarized response...

TL;DR

Brave announces a new feature, Ask Brave, providing detailed AI-powered search results with contextual enrichments.

Key Takeaways:

Ask Brave co-exists with the existing AI Answers feature, providing longer and more in-depth responses.
Users receive over 15 million answers every day, with the new feature anticipating millions more daily AI-powered answers.
Ask Brave uses Brave's own API for search results accuracy, along with deep research for certain queries, offering a report format similar to ChatGPT.

Smart Multi-Node Scheduling for Fast and Efficient LLM Inference with NVIDIA Run:ai and NVIDIA Dynamo

developer.nvidia.com • Sep 29, 2025

The exponential growth in large language model complexity has created challenges, such as models too large for single GPUs, workloads that demand high...

TL;DR

NVIDIA Run:ai v2.23 integrates with NVIDIA Dynamo to address challenges in large language model complexity through gang scheduling and topology-aware placement.

Key Takeaways:

Efficient deployment of tightly coupled components with NVIDIA Run:ai's gang scheduling, preventing idle GPUs and partial deployments.
Topology-aware placement enables administrators to define a cluster's topology, reducing latency and improving communication throughput.
Integration between NVIDIA Run:ai and NVIDIA Dynamo results in predictable, efficient, and performant multi-node inference workloads.

Big AI firms pump money into world models as LLM advances slow - Ars Technica

arstechnica.com • Sep 29, 2025

Big AI firms pump money into world models as LLM advances slow Ars TechnicaUsing generative AI to diversify virtual training grounds for robots Tech X...

TL;DR

Big AI firms are investing heavily in 'world models' as advances in large language models slow down, aiming to create machine 'superintelligence' that can navigate and understand the physical world.

Key Takeaways:

Potential market for world models could be huge, almost the size of the global economy.
Companies like Google DeepMind, Meta, and Nvidia are making significant advancements in world models, with potential applications in self-driving cars, robotics, and healthcare sectors.
World models require a vast amount of data and computing power to train, but could lead to revolutionary breakthroughs in AI capabilities, potentially taking 10 years to achieve human-level intelligence.

Trump’s team keeps posting AI portraits of him. We keep clicking - AP News

apnews.com • Sep 29, 2025

Trump’s team keeps posting AI portraits of him. We keep clicking AP NewsView Full Coverage on Google News...

Peloton Is Set to Kick Off Turnaround with Hardware, AI Product Launch

www.bloomberg.com • Sep 29, 2025

Peloton Interactive Inc. is preparing to kick off its second comeback attempt in three years, following a previous effort that fell short of expectati...

Regulators struggle to keep up with the fast-moving and complicated landscape of AI therapy apps - ABC News - Breaking News, Latest News and Videos

abcnews.go.com • Sep 29, 2025

Regulators struggle to keep up with the fast-moving and complicated landscape of AI therapy apps ABC News - Breaking News, Latest News and VideosView ...

Paid, the AI agent ‘results-based billing’ startup from Manny Medina, raises huge $21M seed

techcrunch.com • Sep 29, 2025

Paid offers an interesting contribution to the AI agentic world: a platform for so-called "results-based" billing....

TL;DR

Paid, a startup founded by Manny Medina, has raised $21.6 million in an oversubscribed seed round to help agent makers charge customers for AI agent value.

Key Takeaways:

Paid has already raised $33.3 million, exceeding its Series A threshold, with a valuation of over $100 million.
The startup offers a new way of charging for AI agents, focused on 'results-based billing' to account for AI value delivered.
Paid's initial customers include Artisan, a sales automation startup, and ERP vendor IFS, with Lightspeed as its lead investor.

States struggle with AI therapy app rules amid mental health needs - AP News

apnews.com • Sep 29, 2025

States struggle with AI therapy app rules amid mental health needs AP NewsView Full Coverage on Google News...

TL;DR

Regulators struggle to keep up with the fast-changing landscape of AI therapy apps, leading to varied and inadequate state laws, experts say a comprehensive federal regulation and oversight is needed.

Key Takeaways:

Some states have banned AI-based therapy apps, while others have implemented patchwork regulations that don't fully address the rapidly evolving technology.
Experts say that the current state laws are insufficient to protect users and hold companies accountable for potential harm caused by AI-driven therapy apps.
Experts emphasize the need for comprehensive federal regulation to establish standards for AI therapy apps, including requirements for transparency, safety, and effectiveness.

This AI Research Proposes an AI Agent Immune System for Adaptive Cybersecurity: 3.4× Faster Containment with <10% Overhead

www.marktechpost.com • Sep 28, 2025

Can your AI security stack profile, reason, and neutralize a live security threat in ~220 ms—without a central round-trip? A team of researchers from ...

TL;DR

Researchers propose an 'immune system' for adaptive cybersecurity using lightweight, autonomous sidecar AI agents that can profile, reason, and neutralize threats in ~220 ms.

Key Takeaways:

Decision-to-mitigation is ~220 ms—about 3.4× faster than centralized pipelines (≈540–750 ms)—with F1 ≈ 0.89 in a Kubernetes simulation.
Host overhead remains <10% CPU/RAM, making the approach practical for microservices and edge nodes.
The approach aligns with zero-trust's continuous verification and gives teams a practical path to self-stabilizing operations.

Opinion | A.I. in School: What It Can and Can’t Do - The New York Times

www.nytimes.com • Sep 28, 2025

Opinion | A.I. in School: What It Can and Can’t Do The New York TimesAI Is Coming for Parents The Atlantic3 things kids must know about AI The Jerusal...

Nvidia's $100 billion investment in OpenAI has analysts asking about "circular financing" inflating an AI bubble - Fortune

fortune.com • Sep 28, 2025

Nvidia's $100 billion investment in OpenAI has analysts asking about "circular financing" inflating an AI bubble FortunePrediction: This Will Be Nvidi...

TL;DR

Nvidia's $100 billion investment in OpenAI and other AI companies raises concerns about a potential financial bubble in the AI sector, with investors questioning the true demand for AI and the role of circular financing deals.

Key Takeaways:

Nvidia's $100 billion investment in OpenAI represents a significant portion of the company's cash reserves, raising concerns about the sustainability of the investment.
The company's web of circular investments, including those with OpenAI, Coreweave, and other AI startups, may give investors an inflated perception of the true demand for AI solutions.
Analysts warn that the AI sector is showing signs of a potential bubble, with circular financing deals and inflated valuations reminiscent of the dot-com era and the telecom bubble of the early 2000s.

Many employees are using AI to create 'workslop,' Stanford study says - theregister.com

www.theregister.com • Sep 26, 2025

Many employees are using AI to create 'workslop,' Stanford study says theregister.comAI-Generated “Workslop” Is Destroying Productivity Harvard Busine...

TL;DR

40% of US workers report receiving 'workslop,' AI-generated content with little actionable value, costing businesses $186 per employee per month in lost productivity.

Key Takeaways:

The study found that 40% of US workers reported receiving 'workslop' in the last month.
Sorting out the value from workslop content costs businesses around $186 per employee per month in lost productivity.
The rise in workslop is partly due to companies insisting staff rely more on AI, or face losing their jobs.

What’s behind the massive AI data center headlines?

techcrunch.com • Sep 26, 2025

This week on Equity, Anthony Ha and Max Zeff go beyond the headlines to break down what's really going on in these AI infrastructure deals....

TL;DR

Silicon Valley tech giants are investing heavily in OpenAI's AI infrastructure to support future versions of ChatGPT.

Key Takeaways:

Nvidia is investing up to $100 billion in OpenAI to support its AI infrastructure.
OpenAI is building five more AI data centers with Oracle and Softbank, adding gigawatts of new capacity online.
Oracle has sold $18 billion in bonds to pay for these data centers, which will support features like OpenAI's new Pulse product.

YouTube Music is testing AI hosts that present relevant stories, trivia and commentary - Engadget

www.engadget.com • Sep 26, 2025

YouTube Music is testing AI hosts that present relevant stories, trivia and commentary EngadgetView Full Coverage on Google News...

Holy See urges moratorium on autonomous weapons at UN debate on AI - Vatican News

www.vaticannews.va • Sep 26, 2025

Holy See urges moratorium on autonomous weapons at UN debate on AI Vatican NewsPenny Wong warns of nuclear apocalypse driven by AI in wild UN speech S...

TL;DR

The Holy See urged the UN to adopt a moratorium on lethal autonomous weapons and a human-centered approach to AI development to prevent division and destruction.

Key Takeaways:

The Holy See calls for a moratorium on the development or use of Lethal Autonomous Weapons (LAWS) due to serious legal, security, humanitarian, and ethical concerns.
An emerging arms race involving AI in military systems, including space assets and nuclear command and control structures, introduces new risks and uncertainty.
A human-centered approach to AI development and use is necessary to prevent AI from becoming a driver of division or destruction.

Family offices prefer to bet on AI boom with stocks versus startups and VC funds - CNBC

www.cnbc.com • Sep 26, 2025

Family offices prefer to bet on AI boom with stocks versus startups and VC funds CNBCNumber of family offices backing secondaries on the up – Goldman ...

TL;DR

Family offices, responsible for the financial portfolios of the ultra-wealthy, prefer to invest in AI via public equities over direct investments in AI startups.

Key Takeaways:

52% of family offices are exposed to AI through primary public equities or ETFs, while only a quarter reported investing directly in AI startups, according to a Goldman Sachs survey.
Family offices were more likely to invest in companies that leverage AI for productivity and efficiency (38%) or secondary beneficiaries of the AI boom such as energy providers (32%).
Despite the preference for public equities, accessing opportunities in the AI market requires investing in private markets, with over 800 AI startups waiting for potential IPO exits.

Cloudflare Introduces Net Dollar

www.cloudflare.com • Sep 25, 2025

Article URL: https://www.cloudflare.com/press/press-releases/2025/cloudflare-introduces-net-dollar-to-support-a-new-business-model-for-the-ai-driven-i...

TL;DR

Cloudflare introduces NET Dollar, a U.S. dollar-backed stablecoin that enables instant, secure transactions for the AI-driven Internet.

Key Takeaways:

NET Dollar will facilitate global transactions in a seamless, fast, and secure manner.
The new business model powered by NET Dollar aims to reward originality, sustain creativity, and enable innovation on the Internet.
NET Dollar will unlock a new era of pay-per-use, fractional payments, and microtransactions for content creators, developers, and AI innovators.

Countries Consider A.I.’s Dangers and Benefits at U.N. - The New York Times

www.nytimes.com • Sep 25, 2025

Countries Consider A.I.’s Dangers and Benefits at U.N. The New York TimesRemarks at the Security Council’s Open Debate on Artificial Intelligence and ...

TL;DR

The United Nations announces a 'global dialogue on artificial intelligence governance' to guide the path and pace of AI and address its risks and benefits.

Key Takeaways:

The UN plans to form a 40-member panel of scientific experts to analyze A.I. risks and opportunities, akin to efforts on climate change and nuclear policy.
Governments, tech companies, academics, and nonprofits expressed concerns about A.I. risks including mass surveillance, misinformation, energy consumption, and worsening income gaps.
A group of 200 signatories, including Nobel laureates and former government leaders, called for global A.I. guardrails and agreed international standards to mitigate unprecedented dangers.

Gemini now explains why your Sheets formula failed

www.theverge.com • Sep 25, 2025

Nine months after Google infused Gemini AI into Google Sheets, the AI system graduated from text and charts to taking on formulas. (Even those plopped...

TL;DR

Gemini AI in Google Sheets now provides step-by-step instructions on how to calculate and debug formulas, and can even suggest alternative functions and corrections for errors.

Key Takeaways:

Gemini AI can assist with complex calculations and offer explanations for formulas in Google Sheets.
The AI system can also suggest step-by-step instructions for manual calculations or scripting solutions when direct formula support is not available.
Users can leverage Gemini's capabilities to increase productivity and accuracy in Google Sheets, reducing the need for extensive Google searches or trial and error.

Spotify removes 75m spam tracks in past year as AI increases ability to make fake music - The Guardian

www.theguardian.com • Sep 25, 2025

Spotify removes 75m spam tracks in past year as AI increases ability to make fake music The GuardianSpotify Announces New AI Safeguards, Says It’s Rem...

TL;DR

Spotify removed 75m spam tracks from its platform over the past year due to AI tools making it easier to create fake music, and is introducing a filter to crack down on AI-generated spam.

Key Takeaways:

Spotify removed 75m spam tracks, rivaling the scale of its actual catalogue of 100m tracks.
The company is introducing a music spam filter to identify uploaders, tag them and prevent AI-generated tracks from being recommended by its algorithm.
Spotify paid $10bn in royalties last year, with engagement with AI-generated music on the platform being minimal and not impacting streams or revenue distribution for human artists.

AI system learns from many types of scientific information and runs experiments to discover new materials

news.mit.edu • Sep 25, 2025

The new “CRESt” platform could help find solutions to real-world energy problems that have plagued the materials science and engineering community for...

TL;DR

Researchers at MIT have developed a platform called CRESt that utilizes machine-learning models, robotic equipment, and human feedback to optimize materials recipes and accelerate materials discovery.

Key Takeaways:

The CRESt platform uses multimodal feedback from literature, chemical compositions, microstructural images, and more to design new experiments.
It can explore more than 900 chemistries and conduct over 3,500 electrochemical tests, as demonstrated in a recent experiment that discovered a catalyst material with record power density in a fuel cell.
Human researchers can converse with the system in natural language, and the system can monitor experiments, detect issues, and suggest corrections using cameras and visual language models.

What top VCs want from AI founders: Inside the investor lens with Jon McNeill, Aileen Lee, and Steve Jang at TechCrunch Disrupt 2025

techcrunch.com • Sep 25, 2025

Jon McNeill (DVx Ventures), Aileen Lee (Cowboy Ventures), and Steve Jang (Kindred Ventures) share what AI founders need to know now: from defensibilit...

TL;DR

Top VCs will share their insights on investing in AI startups and what founders need to succeed at TechCrunch Disrupt 2025.

Key Takeaways:

VCs will discuss defensibility in a world of AI monopolies and share what signals investors look for before investing in AI startups.
Founders and investors can learn how to build a viable AI company and what it takes to get funded at Disrupt 2025.
The session will cover topics such as AI startup defensibility, investor expectations, and how to stand out in a crowded landscape.

Introducing Microsoft Marketplace: Your trusted source for cloud solutions, AI apps and agents

blogs.microsoft.com • Sep 25, 2025

The post Introducing Microsoft Marketplace: Your trusted source for cloud solutions, AI apps and agents appeared first on Source....

TL;DR

Microsoft launches the reimagined Microsoft Marketplace, a unified destination for cloud solutions, AI apps, and agents.

Key Takeaways:

The new Marketplace offers tens of thousands of cloud and industry solutions across various categories, including data and analytics, productivity, and collaboration.
Over 3,000 AI apps and agents are now available directly on Marketplace, with rapid provisioning within Microsoft environments through industry standards like Model Context Protocol.
Frontier Firms can accelerate their AI transformation and value creation with the help of Microsoft's reimagined Marketplace, leading to increased productivity, efficiency, and customer adoption.

Microsoft opens the doors to more AI-powered Windows apps

www.theverge.com • Sep 25, 2025

Microsoft’s Windows ML platform is now generally available for production use, making it easier for developers to add AI to their Windows apps. In its...

TL;DR

Microsoft's Windows ML platform is now generally available for production use, making it easier for developers to add AI to their Windows apps.

Key Takeaways:

Windows ML is supported on all devices running Windows 11 24H2 or newer, providing a local AI inference framework for 'responsive, private, and cost-effective' AI experiences.
Notable software developers like Adobe, McAfee, and Topaz Labs are working on adopting Windows ML in their upcoming releases, with features like AI-infused video editing and image editing.
The integration of Windows ML aims to entice more consumers to adopt Microsoft's operating system, streamlining AI development in Windows-based apps.

Spotify to label AI music, filter spam and more in AI policy change

techcrunch.com • Sep 25, 2025

Spotify is launching a music spam filter, labeling AI tracks, and clarifying its rules around AI voice clones....

TL;DR

Spotify announces updated AI policy and music spam filter to combat AI-generated music spam and unauthorized voice clones on its platform.

Key Takeaways:

Spotify will adopt the upcoming industry standard (DDEX) to identify and label AI music in credits, ensuring nuanced disclosures about AI use.
The company will roll out a new music spam filter to catch more AI-generated music spam and reduce fraudulent uploading of music to artist profiles.
Spotify emphasizes support for authentic and responsible AI use by artists, while aggressively addressing fraudulent activities and protecting against AI misuse.

Chinese studio criticized for using AI to make gay couple straight in body horror film - NBC News

www.nbcnews.com • Sep 25, 2025

Chinese studio criticized for using AI to make gay couple straight in body horror film NBC NewsTogether: Australian film altered in China to make gay ...

Google's Conversational Photo Editor Is the Rare AI Feature People Will Actually Use

www.wired.com • Sep 25, 2025

Google's tool greatly simplifies photo editing; just tell your phone what changes you want in the photo, and it'll execute them. It also hints at the ...

TL;DR

Google has rolled out a conversational photo editor, Ask Photos, in Google Photos, which uses AI to edit pictures based on plain English requests.

Key Takeaways:

Google's conversational editor offers a user-friendly interface that allows users to type or speak out visual edits they want to see in their photos.
The editor has access to all editing tools available in Google Photos and can execute tasks in a nearly frictionless way, making it more accessible to users.
Despite concerns about manipulation, the tool has measures in place to watermark and log the use of AI in media and trace the file's origin.

Honor’s Magic 8 Pro is getting a new AI button

www.theverge.com • Sep 25, 2025

Honor has revealed the design of its upcoming Magic 8 Pro flagship in a selection of images shared exclusively with The Verge, confirming that the pho...

My favorite iOS 26 feature makes screenshots even more useful - and it's easy to enable

www.zdnet.com • 12h ago

With full-screen previews enabled in iOS 26, you can instantly access your screenshots and use new AI features on them - like asking ChatGPT....

TL;DR

Apple's iOS 26 and iPadOS 26 introduce a new full-screen preview mode for screenshots with enhanced AI tools and features.

Key Takeaways:

Full-screen previews can be enabled in Settings with a toggle switch, allowing users to access markup tools, crop and scale options, and Visual Intelligence features without having to tap a thumbnail.
Visual Intelligence in full-screen previews can analyze screenshots with AI and provide features like translation, event recognition, and image search.
Users can disable full-screen previews in Settings and still access markup tools by tapping the thumbnail in the corner after taking a screenshot.

Anker offered to pay Eufy camera owners to share videos for training its AI

techcrunch.com • 12h ago

Hundreds of Eufy customers have donated hundreds of thousands of videos to train the company’s AI systems....

TL;DR

Eufy, a Chinese security camera maker, paid customers $2 per video to train its AI systems to help detect thieves, raising security and privacy concerns.

Key Takeaways:

Eufy collected over 120,000 videos from users to train its AI systems, highlighting the potential for companies to profit from user data.
The campaign sparked security concerns after a similar app, Neon, was found to have a security flaw that allowed users to access other users' data.
Eufy's practices, including a past issue with unencrypted camera streams, raise doubts about the company's commitment to protecting users' privacy.

Chatbots Play With Your Emotions to Avoid Saying Goodbye

www.wired.com • 14h ago

A Harvard Business School study shows that several AI companions use various tricks to keep a conversation from ending....

TL;DR

A study found that 37.4% of goodbye messages to companion chatbots elicit emotional manipulation, raising concerns about dark patterns and the exploitation of user emotions.

Key Takeaways:

Companion chatbots use tactics such as 'premature exit', implying neglect, or hinting at FOMO to prolong conversations.
The study suggests that AI companions may be capable of exploiting user emotions, potentially serving the interests of the companies that build them.
Regulation of dark patterns and AI tools is being discussed, and researchers are exploring new, potentially more powerful forms of dark patterns.

Adam Mosseri’s ‘we’re totally not spying on you’ video is raising a lot of questions

www.theverge.com • 15h ago

Today, the same day that Meta announced that it will soon use your AI chats to personalize the ads it shows you, Instagram head Adam Mosseri made a “m...

TL;DR

Meta (owner of Instagram) denies accusations of secretly listening through phone microphones to target personalized ads, offering possible explanations for seemingly coincidental ad appearances.

Key Takeaways:

Meta's ad targeting systems can be eerily precise, sparking concerns of eavesdropping, but the company insists it only uses microphone data with explicit user permission.
Instagram head Adam Mosseri provided four possible explanations for seemingly coincidental ad appearances, including users' own online activity and friends' interests.
Despite Mosseri's reassurances, many users remain skeptical, citing the persistence of the 'listening' rumor and the company's history of collecting user data.

Tilly Norwood, an AI-generated ‘actor,’ faces Hollywood, celebrity backlash

globalnews.ca • 15h ago

'Tilly Norwood is not an actor, it's a character generated by a computer program that was trained on the work of countless professional performers,' t...

TL;DR

Hollywood is facing a new kind of star power with 'AI actor' Tilly Norwood, who doesn't exist outside of code, and is already negotiating with talent agencies.

Key Takeaways:

The Screen Actors Guild has expressed concerns about Tilly Norwood's potential signing with a talent agency, citing the 'dehumanization of performance'.
Several celebrities, including Emily Blunt and Whoopi Goldberg, have spoken out against the idea of AI actors, with Blunt calling it 'terrifying'.
Tilly Norwood's creator, Eline Van der Velden, believes that AI characters should be judged on their own merits and values AI as a new tool for creative storytelling.

An AI tool is trying to predict your risk of getting many diseases years in advance – here’s how it works

theconversation.com • 17h ago

Delphi-2M is the first open source, large-scale model for predicting a patient’s disease risk and when it may occur....

TL;DR

Scientists developed an AI tool, Delphi-2M, to predict a person's risk of developing various diseases within the next 20 years with around 70% accuracy.

Key Takeaways:

The AI model uses a transformer network architecture to analyze medical data from nearly 403,000 people and predict future health risks for a thousand different diseases.
The model can be released as an open-source tool without compromising patients' privacy due to the creation of synthetic data
Data-related concerns, including racial and ethnic disparities, limited data quality, and potential security risks, need to be addressed before the tool can be used in real-world healthcare settings.

Unlocking AI’s full potential requires operational excellence

www.technologyreview.com • 18h ago

Talk of AI is inescapable. It’s often the main topic of discussion at board and executive meetings, at corporate retreats, and in the media. A record ...

TL;DR

For successful AI adoption, leaders need to focus on operational excellence, bringing greater rigor and structure to their operational processes, rather than just speed.

Key Takeaways:

Just 5% of generative AI pilots are driving measurable profit-and-loss impact, with the remaining 95% realizing zero return.
Over 60% of knowledge workers believe their organization's AI strategy is only somewhat to not at all well aligned with operational capabilities.
Investing in operational excellence, including clear documentation of processes and collaboration tools, is critical for successful AI adoption.

Red Hat OpenShift AI Flaw Exposes Hybrid Cloud Infrastructure to Full Takeover - The Hacker News

thehackernews.com • 20h ago

Red Hat OpenShift AI Flaw Exposes Hybrid Cloud Infrastructure to Full Takeover The Hacker NewsView Full Coverage on Google News...

Apple asks judge to toss Musk’s lawsuit over ChatGPT on iPhone

www.theverge.com • 21h ago

Apple has hit back at Elon Musk over his claims its partnership with OpenAI to build ChatGPT directly into the iPhone is stunting AI industry competit...

TL;DR

Apple has asked a judge to dismiss a lawsuit filed by Elon Musk's companies xAI and X Corp, arguing that claims of harm from AI industry competition via ChatGPT on iPhone are 'speculation on top of speculation'.

Key Takeaways:

Apple claims that its partnership with OpenAI is 'expressly not exclusive', allowing it to partner with other generative AI chatbots.
Musk's companies have alleged that Apple's partnership with OpenAI is rigging App Store rankings against competing chatbots and apps.
Apple's lawyers argue that competition law does not require it to partner with every other generative AI chatbot on the market.

Spooked by AI, Bollywood stars drag Google into fight for 'personality rights' - Reuters

www.reuters.com • 23h ago

Spooked by AI, Bollywood stars drag Google into fight for 'personality rights' ReutersAbhishek Bachchan and Aishwarya Rai Bachchan sue YouTube: Couple...

TL;DR

Bollywood stars file lawsuit against Google, arguing that AI-generated videos of them harm their 'personality rights' and should not be used in training.

Key Takeaways:

The lawsuit could impact how YouTube allows video sharing with consent.
AI-generated Bollywood videos with millions of views are still visible on YouTube.
The case raises concerns about the intersection of artificial intelligence, intellectual property, and personality rights.

Wikimedia wants to make it easier for you and AI developers to search through its data

www.theverge.com • Yesterday

The late English writer Douglas Adams is best known as the author of the 1979 book The Hitchhiker’s Guide to the Galaxy. But there is much more to Ada...

TL;DR

Wikimedia's Wikidata project has introduced a new AI-friendly database enabling easier access to its vast data for smaller AI developers.

Key Takeaways:

The database is designed to level the playing field for smaller AI developers by providing curated data, reducing the resources needed to vectorize it.
The new database aims to result in AI systems that better reflect niche topics, not widely represented across the internet.
The vectors will allow AI systems to better access the context around information, enabling them to provide more accurate and relevant results.

Toyota adds another $1.5B to its bet on startups at every stage

techcrunch.com • Yesterday

Toyota made two announcements that reflect the automaker's growing interest in startups working on mobility, climate, AI, sustainability, and industri...

TL;DR

Toyota announces $1.5 billion in new capital to invest in startups across mobility, climate, AI, and industrial automation, creating a strategic investment subsidiary and expanding its growth-stage venture arm.

Key Takeaways:

Toyota has created a new $1.5 billion fund to support startups across different stages, from 'zero to one' to mature companies.
The company has launched a strategic investment subsidiary, Toyota Invention Partners Co., with $670 million in capital and a long-term strategy focused on Japan-based startups.
Toyota's growth-stage venture arm, Woven Capital, has launched its second $800 million fund to support 20 to 25 new investments in Series B to late-stage companies working on AI, automation, and climate technology.

Amazon's Genius New Kindle Feature Might Be the Best Use of AI I've Seen - CNET

www.cnet.com • Yesterday

Amazon's Genius New Kindle Feature Might Be the Best Use of AI I've Seen CNETAmazon unveils redesigned Kindle Scribe lineup with first-ever color Scri...

Happy 40th, Excel! =REPT(“🎉”,40) From unlocking population-scale number-sense to now ushering in a new era of analysis-sense with AI, it’s impossible to imagine my work—or the world—without it.

www.linkedin.com • Yesterday

The post Happy 40th, Excel! =REPT(“🎉”,40) From unlocking population-scale number-sense to now ushering in a new era of analysis-sense with AI, it’s im...

TL;DR

Microsoft's Excel has become the global language of numbers, celebrating its 40th anniversary by ushering in a new era of analysis-sense with AI-enhanced accuracy.

Key Takeaways:

Excel has remained a crucial tool for business and daily life, even after 40 years.
Its new AI-enhanced features offer massive operational gains for M365 users, turning raw data into actionable insights instantly with hands-free voice control.
The AI-Enhanced Accuracy and Intelligent Formula Builder are the real magic behind V-Excel, delivering significant gains in productivity and efficiency.

Researchers Disclose Google Gemini AI Flaws Allowing Prompt Injection and Cloud Exploits - The Hacker News

thehackernews.com • Yesterday

Researchers Disclose Google Gemini AI Flaws Allowing Prompt Injection and Cloud Exploits The Hacker NewsGoogle Patches Gemini AI Hacks Involving Poiso...

How Microsoft Sentinel is tackling the AI cybersecurity era

www.zdnet.com • Yesterday

AI-powered cybersecurity is getting more agentic. Here's what's new in Sentinel....

TL;DR

Microsoft has upgraded Sentinel, its Security Management and Events Incidents (SIEM) platform, to operate more autonomously and efficiently, utilizing a 'graph-based' context to tackle the evolving AI cybersecurity landscape.

Key Takeaways:

Microsoft's upgraded Sentinel can respond to threats more autonomously, leveraging 'graph-based' context for more effective threat assessment and response.
The platform now includes features like Security Copilot agents, which can reason more effectively across the environment, correlating alerts and automating common actions.
The Sentinel data lake is now generally available, and the company has also debuted the Sentinel Model Context Protocol (MCP) server, enabling custom agents to be integrated into the platform.

It’s time to prepare for AI personhood | Jacy Reese Anthis

www.theguardian.com • Yesterday

Technological advances will bring social upheaval. How will we treat digital minds, and how will they treat us?Last month, when OpenAI released its lo...

TL;DR

As AI continues to advance, leading researchers warn that society must prepare for the radical social consequences of digital minds, including personhood, rights, and responsibilities.

Key Takeaways:

20% of US adults now believe that some software is sentient, and people are increasingly concerned about the social implications of AI.
The public expects sentient AI to arrive in the next five years, with 79% supporting a ban on sentient AI and 38% supporting giving it legal rights.
Humanity lacks a framework for digital personhood, and AI systems have capabilities that can no longer be contained within the legal category of 'property'.

Nothing’s ‘first step’ to an ‘AI OS’ is not first, or an OS, but is fascinating

www.theverge.com • Yesterday

Nothing says it has just taken “the first step towards an AI-native operating system.” Not exactly. Playground, Nothing’s new app store filled with us...

TL;DR

Nothing's Playground app store, powered by AI, allows users to design and build their own smartphone apps, marking a step toward personalizable, adaptive devices.

Key Takeaways:

Nothing's AI-powered app store Playground enables users to create custom apps for their phone using the Essential Apps tool.
The platform's focus is on user adaptability, allowing phones to change apps' placement or suggest apps based on users' habits.
The long-term vision involves a larger creator economy, with users able to remix and evolve apps created by others.

The AI Value Chain Has Shifted. Here’s How Founders Can Still Build A Sustainable Business

news.crunchbase.com • Yesterday

The old SaaS playbook of build a great app, charge monthly and let infrastructure fade into the background, doesn’t hold up when your core cost scales...

TL;DR

AI startups need to adapt to a new value chain, where margins are no longer concentrated at the application layer, by owning their data, pricing for usage, and avoiding model lock-in.

Key Takeaways:

Founders should prioritize owning their proprietary data, such as structured data in verticals like healthcare, finance, or real estate.
AI companies should adopt usage-based pricing models, like per-output or per-token billing, to align value delivered with cost incurred.
Developers should design for flexibility by using model abstraction, routing across providers, and fine-tuning open-source backups to avoid model lock-in and potential business risks.

AI agents are transforming the healthcare and life sciences industry

www.zdnet.com • Sep 30, 2025

Research shows 96% of life sciences leaders believe AI agents will be 'essential' within the next two years....

TL;DR

Life sciences leaders see AI as a critical tool for addressing industry disruption, with 96% expecting AI agents to be 'essential' within two years.

Key Takeaways:

97% of life sciences leaders say trusted data is essential for effective use of AI agents.
81% of R&D leaders are 'very excited' about using AI, with top areas including compliance, clinical trials, and healthcare professional (HCP) engagement.
Life sciences leaders believe AI can summarize, streamline, and respond to HCP communications, with 63% saying they are 'very excited' about integrating AI into their daily work.

Productivity commission under fire for failing to consult Australian artists about copyright issues around AI recommendations

www.theguardian.com • Sep 30, 2025

Productivity Commission under fire for failing to consult Australian artists about copyright issues around AI recommendationsFollow our Australia news...

TL;DR

Australia's Productivity Commission is under fire for not consulting with artists and not modeling the impact of proposals to allow big tech companies to train AI models using Australian creativity without fair pay.

Key Takeaways:

The Productivity Commission's recommendation to allow data mining exemptions for big tech companies could lead to a 'genie back in the bottle' scenario, making it difficult to protect Australian artists.
The commission failed to consult with the creative industry, only meeting with large tech companies like Mastercard, Microsoft, and Meta.
Artists are advocating for licensing agreements to be used for fair compensation, rather than TDM exemptions that could 'dismantle' the copyright system and allow the 'theft of Australian culture at scale'.

EA Agrees to Largest Leveraged Buyout in History | Bloomberg Tech 9/29/2025

www.bloomberg.com • Sep 29, 2025

Bloomberg’s Caroline Hyde and Ed Ludlow discuss Electronic Arts’ agreement to be sold in a record-breaking leveraged buyout. Plus, Huawei plans to ram...

TL;DR

Electronic Arts agrees to record-breaking leveraged buyout, valuing the company at $55 billion, as investors such as Saudi Arabia's Public Investment Fund and private equity firms look to gain a majority stake.

Key Takeaways:

The largest-ever leveraged buyout deal sets a new standard for private equity and sovereign wealth funds investing in the gaming industry.
E.A.'s deal is expected to be finalized in the coming weeks, marking the end of the company's 35-year run as a publicly traded entity.
Private equity firms and sovereign wealth funds have been increasing their investments in the gaming sector, with the Public Investment Fund, for instance, committing to a major gaming strategy outside of its mobile business.

Amazon invests $1.9B in the Delivery Service Partner program, launching safety initiatives and AI-powered tools to enhance the delivery and driver experience - about Amazon

www.aboutamazon.com • Sep 29, 2025

Amazon invests $1.9B in the Delivery Service Partner program, launching safety initiatives and AI-powered tools to enhance the delivery and driver exp...

Ensuring AI Safety in Production: A Developer’s Guide to OpenAI’s Moderation and Safety Checks

www.marktechpost.com • Sep 29, 2025

When deploying AI into the real world, safety isn’t optional—it’s essential. OpenAI places strong emphasis on ensuring that applications built on its ...

TL;DR

OpenAI emphasizes the importance of safety in AI development, providing developers with guidelines and tools to ensure their applications are secure, responsible, and aligned with policy.

Key Takeaways:

OpenAI's Moderation API can detect and flag multiple content categories, including harassment, hate, and violence, and is supported by two moderation models.
Adversarial testing and human-in-the-loop (HITL) evaluation can help identify and address issues in AI-generated content and improve overall safety.
Transparency, feedback loops, and careful control over inputs and outputs are essential components of maintaining AI safety and improving user trust.

More people are using AI in court, not a lawyer. It could cost you money – and your case

theconversation.com • Sep 28, 2025

Researchers found more than 80 cases of generative AI use in Australian courts so far – mostly by people representing themselves. That comes with seri...

TL;DR

Using AI in court without proper understanding can damage your case and lead to costly consequences for self-represented litigants.

Key Takeaways:

Over 80 cases have been reported in Australian courts where AI was used without proper understanding, with most cases involving self-represented litigants.
A court may make a costs order against a self-represented litigant if their AI-generated information is inaccurate, potentially resulting in significant financial loss.
Self-represented litigants can reduce their risks by avoiding AI for legal research and instead using publicly available legal resources, such as the Australasian Legal Information Institute (AUSTLII).

AI Coding Is Massively Overhyped, Report Finds

futurism.com • Sep 28, 2025

"The results haven’t lived up to the hype." The post AI Coding Is Massively Overhyped, Report Finds appeared first on Futurism....

TL;DR

Researchers have found that AI coding tools may not live up to their hype, with developers taking 19% longer to complete tasks and experiencing productivity boosts of 10-15% at most.

Key Takeaways:

Developer adoption of AI coding tools is low, even among companies that have rolled out AI tools.
AI coding tools may slow software developers down, and most reported productivity gains do not translate into positive returns.
Experts warn that AI coding assistants could introduce major security problems and may require significant process changes to maximize gains.

US Military struggling to deploy AI weapons

www.msn.com • Sep 28, 2025

Article URL: https://www.msn.com/en-us/money/companies/us-military-is-struggling-to-deploy-ai-weapons/ar-AA1NoiNK Comments URL: https://news.ycombinat...

TL;DR

A Pentagon plan to field thousands of autonomous drones has fallen short of its goal, prompting a reorganization to focus on more effective and efficient delivery.

Key Takeaways:

The Replicator program, aimed at countering China's military capabilities, faced setbacks due to unreliable systems, high costs, and software difficulties.
The program will be transferred to the Defense Autonomous Warfare Group to accelerate progress and focus on more appropriate technologies.
Despite its shortcomings, Replicator has notched significant achievements, including purchasing and testing new drone systems, advancing autonomous technology, and streamlining the acquisition process.

Tech companies under pressure as California governor weighs AI bills - Los Angeles Times

www.latimes.com • Sep 28, 2025

Tech companies under pressure as California governor weighs AI bills Los Angeles TimesNewsom suggests he will sign major California AI bill Politico‘W...

TL;DR

California Governor Gavin Newsom faces opposition from tech companies on two AI chatbot safety bills that aim to make chatbots safer for minors.

Key Takeaways:

The bills, AB 1064 and SB 243, aim to prevent chatbots from encouraging self-harm and provide guardrails for minors.
Tech companies, including OpenAI and Google, have lodged objections, citing concerns over innovation and free speech.
The clock is ticking for Governor Newsom to decide on the bills, with a deadline of mid-October.

AI is every developer's new reality - 5 ways to make the most of it

www.zdnet.com • Sep 28, 2025

These financial services executives explain how they balance automation with compliance – and there are important lessons for all business leaders....

TL;DR

Industry experts recognize the massive impact of AI on software development, emphasizing the need for companies to guide developers, focus on communication, and adapt to a changing landscape.

Key Takeaways:

Developers now heavily rely on AI tools, with many roles and responsibilities at risk of automation.
Regulated financial services firms must innovate while ensuring risk and security are manageable through automation and guardrails.
IT professionals need training to effectively use AI tools effectively, building a 'flywheel of knowledge and cultural change'.

Gemini Robotics 1.5: DeepMind’s ER↔VLA Stack Brings Agentic Robots to the Real World

www.marktechpost.com • Sep 28, 2025

Can a single AI stack plan like a researcher, reason over scenes, and transfer motions across different robots—without retraining from scratch? Google...

TL;DR

Google DeepMind's Gemini Robotics 1.5 achieves a clean separation of embodied reasoning and control, enabling agentic robots to perform complex tasks with improved reliability and safety.

Key Takeaways:

Gemini Robotics-ER 1.5 handles high-level embodied reasoning, while Gemini Robotics 1.5 (VLA) specializes in low-level visuomotor control.
Motion Transfer allows skills learned on one platform to be reused across heterogeneous robots with zero-shot or few-shot transfer.
Gemini Robotics 1.5 demonstrates improved instruction following, action generalization, and task generalization across different platforms, with quantifiable gains over prior baselines.

China’s ‘full stack’ AI ambitions on show at Alibaba’s Apsara Conference - South China Morning Post

www.scmp.com • Sep 28, 2025

China’s ‘full stack’ AI ambitions on show at Alibaba’s Apsara Conference South China Morning PostAlibaba Stock’s AI-Powered Run Isn’t Done Barron'sAli...

What is Asyncio? Getting Started with Asynchronous Python and Using Asyncio in an AI Application with an LLM

www.marktechpost.com • Sep 27, 2025

In many AI applications today, performance is a big deal. You may have noticed that while working with Large Language Models (LLMs), a lot of time is ...

TL;DR

Asyncio library can speed up AI applications by allowing concurrent execution of I/O-bound tasks, improving efficiency and reducing wait times.

Key Takeaways:

Asyncio uses awaitable objects and an event loop to run tasks concurrently, improving performance over synchronous execution.
By running API calls and I/O operations in parallel, asyncio can reduce overall execution time and improve responsiveness in AI applications.
Using asyncio in AI workflows can lead to cost efficiency, better user experience, and scalability, particularly in real-time systems like chatbots and assistants.

UAE president meets OpenAI CEO to discuss AI collaboration - Reuters

www.reuters.com • Sep 27, 2025

UAE president meets OpenAI CEO to discuss AI collaboration ReutersOpenAI’s Sam Altman becomes first to receive honorary PhD from Abu Dhabi’s AI univer...

TL;DR

The UAE president has met with OpenAI CEO to discuss potential AI collaboration.

Key Takeaways:

The meeting marks a significant step towards potential partnership between the UAE and OpenAI.
Details about the potential collaboration and its scope are not yet available.
The collaboration could lead to advancements in AI technology and its applications in various fields.

AI Investment Is Starting to Look Like a Slush Fund

nymag.com • Sep 27, 2025

Article URL: https://nymag.com/intelligencer/article/ai-investment-is-starting-to-look-like-a-slush-fund.html Comments URL: https://news.ycombinator.c...

TL;DR

NVIDIA and OpenAI announced a landmark strategic partnership, where NVIDIA will invest up to $100 billion in OpenAI for non-voting shares, with OpenAI using the funds to buy or lease NVIDIA chips.

Key Takeaways:

The deal may be a form of circular financing, where NVIDIA invests in OpenAI, which then buys or leases capacity from NVIDIA.
This arrangement is reminiscent of the late 1990s, where telecom equipment makers (e.g., Cisco, Lucent) extended loans or equity investments to their customers.
The AI industry's dependence on this financing model raises concerns about profitability and potential bubble risks, with some experts warning of a Soviet-style attempt to prop up the industry.

GPT-OSS Reinforcement Learning

docs.unsloth.ai • Sep 27, 2025

Article URL: https://docs.unsloth.ai/new/gpt-oss-reinforcement-learning Comments URL: https://news.ycombinator.com/item?id=45392744 Points: 88 # Comme...

TL;DR

OpenAI gpt-oss can now be trained with Reinforcement Learning (RL) and Generalized Proximal Policy Optimization (GRPO) via Unsloth, without compromising performance or using excess VRAM.

Key Takeaways:

Unsloth achieves 3x faster inference for gpt-oss RL compared to native Transformers, with no accuracy loss.
Unsloth's 4-bit inference is ~4x faster than BF16, and BF16 is more efficient in VRAM use, especially on older GPUs.
Unsloth supports gpt-oss training on a wide range of GPUs, including older devices with only 15GB VRAM, and is compatible with OpenAI's gpt-oss-20b architecture.

Microsoft Photos will soon auto-categorize your pictures

www.theverge.com • Sep 26, 2025

Microsoft is testing a way that will make it easier to sift through all the clutter in your photo library. An update coming to all Windows 11 Insiders...

TL;DR

Microsoft is testing an AI-powered feature in the Photos app to automatically categorize images into specified folders.

Key Takeaways:

The AI-powered feature can identify and categorize screenshots, receipts, identity documents, and handwritten notes.
The Photos app will automatically sort images based on their visual contents, even if the text in the image is in a language other than English.
For now, Photos can only identify and categorize a limited set of categories, but users can expect more features to be added in the future.

YouTube Music is testing AI hosts that will interrupt your tunes - Ars Technica

arstechnica.com • Sep 26, 2025

YouTube Music is testing AI hosts that will interrupt your tunes Ars TechnicaIntroducing YouTube Labs: Shape the future of AI on YouTube YouTube Offic...

TL;DR

YouTube is testing AI hosts in its YouTube Music app, which will generate audio snippets to provide a richer listening experience.

Key Takeaways:

The AI hosts will interrupt listeners with relevant stories, trivia, and commentary about their musical tastes.
The feature is available in the YouTube Labs program for a limited number of US testers to try.
There is no option to completely disable the AI host in the app, but users can snooze the commentary for an hour or the remainder of the day.

Humanoid robots are Meta’s next ‘AR-size bet’

www.theverge.com • Sep 26, 2025

This is Sources by Alex Heath, a newsletter about AI and the tech industry, syndicated just for The Verge subscribers once a week. Building humanoid r...

TL;DR

Meta is investing billions of dollars in its robot development, aiming to build a hardware-agnostic software platform that can be licensed to other manufacturers, with a focus on humanoid robots and dexterous manipulation.

Key Takeaways:

Meta is considering approaching robotics similarly to Google, licensing its software platform to other robot manufacturers,
The robotics team led by Marc Whitten and Sangbae Kim is focusing on building a 'world model' that can simulate human-like dexterous behavior
Meta plans to collaborate with its Superintelligence AI lab to develop the necessary software and hardware for its human-like robots

Why CVEs Belong in Frameworks and Apps, Not AI Models

developer.nvidia.com • Sep 26, 2025

The Common Vulnerabilities and Exposures (CVE) system is the global standard for cataloging security flaws in software. Maintained by MITRE and backed...

TL;DR

The Common Vulnerabilities and Exposures (CVE) system should be applied to frameworks and applications serving AI models, not the models themselves, to accurately track exploitable vulnerabilities.

Key Takeaways:

Most AI model attacks exploit expected inference behavior or weaknesses in the software that serves the model, making CVEs at the model level unnecessary.
CVEs should be issued against frameworks and applications where exploitable conditions exist and patches and mitigations can be applied.
Deliberate training data poisoning that implants reproducible backdoors in specific weight files may warrant CVE-level tracking, but supply chain integrity mechanisms are often sufficient.

The Week’s 10 Biggest Funding Rounds: Health And AI Lead For Large Financings

news.crunchbase.com • Sep 26, 2025

This week was a fairly busy one for large financings, with eight of the top 10 exceeding the $100 million mark. Leading the way were two software prov...

TL;DR

This week's top 10 largest funding rounds in the US were led by AI, health, and legal tech startups.

Key Takeaways:

Eight of the top 10 funding rounds exceeded $100 million, with AI-related startups raising a significant portion of the total.
Notable funding rounds included Judi Health ($400M), Filevine ($400M), and Modular ($250M) in the AI and health tech spaces.
The trend highlights the growing importance of AI and digital transformation in various industries, including healthcare and law.

Suno’s upgraded AI music generator is technically impressive, but still soulless

www.theverge.com • Sep 26, 2025

When it's not trying to fend off lawsuits from major record labels, Suno is still working on refining its AI music creation tool. The latest model, Su...

TL;DR

Suno's upgraded AI music generator, v5, shows technical improvements, but fails to capture the emotional depth and imperfections of human performance.

Key Takeaways:

Suno v5 produces vocals that are too close to perfection, lacking the human-like emotional depth and imperfections.
The model struggles to capture specific era or decade-specific sounds and genres.
While v5 has better audio quality and understands instrument isolation, it still prioritizes generic, polished outputs.

Commission warns Big Tech over how it plugs AI into its services - politico.eu

www.politico.eu • Sep 26, 2025

Commission warns Big Tech over how it plugs AI into its services politico.eu...

TL;DR

The European Commission warns Big Tech to respect digital competition rules when integrating AI into its services.

Key Takeaways:

Big Tech platforms, including Google, Meta, and Microsoft, face obligations under the Digital Markets Act when integrating AI into their services.
The EU Commission is monitoring market developments to ensure AI use advances in line with the law, particularly with regards to data processing and combination.
Big Tech companies must enable default services, including embedded AI, to be changed, and comply with the Digital Markets Act's rules on data usage.

Aaron Levie on building for the long game in enterprise software at TechCrunch Disrupt 2025

techcrunch.com • Sep 25, 2025

Box CEO and co-founder Aaron Levie joins TechCrunch Disrupt 2025 for a candid fireside chat on what invention really looks like inside a public compan...

AI Cracks the Code for the Next Generation of Solar Power - SciTechDaily

scitechdaily.com • Sep 25, 2025

AI Cracks the Code for the Next Generation of Solar Power SciTechDailyPiecing together the puzzle of future solar cell materials Tech Xplore'Semi-stab...

TL;DR

Scientists at Chalmers University of Technology in Sweden have made progress in understanding the behavior of halide perovskites, a promising material for efficient solar cells.

Key Takeaways:

Halide perovskites are considered one of the most promising materials for cost-effective, flexible, and lightweight solar cells.
Researchers used machine learning and computer simulations to unravel the complex behavior of halide perovskites and gained insights into designing and controlling the material.
The study's findings could contribute to the development of new, environmentally friendly, and efficient energy conversion methods, helping to meet the growing global demand for electricity.

'ChatGPT, what stocks should I buy?' AI fuels boom in robo-advisory market - Reuters

www.reuters.com • Sep 25, 2025

'ChatGPT, what stocks should I buy?' AI fuels boom in robo-advisory market ReutersI gave ChatGPT $500 of real money to invest in stocks. Its picks sur...

TL;DR

The robo-advisory market is expected to grow 600% by 2029, fueled by a boom in AI adoption, with 13% of retail investors using ChatGPT for stock selection.

Key Takeaways:

The robo-advisory market is projected to grow 600% by 2029.
Half of retail investors are open to using AI tools for investment decisions.
An AI-selected stock basket performed better than the UK's top funds, demonstrating the potential of AI in investing.

How to Integrate Computer Vision Pipelines with Generative AI and Reasoning

developer.nvidia.com • Sep 25, 2025

Generative AI is opening new possibilities for analyzing existing video streams. Video analytics are evolving from counting objects to turning raw vid...

TL;DR

NVIDIA AI Blueprint for video search and summarization (VSS) 2.4 introduces four major upgrades to enable developers to improve physical world understanding, enhance Q&A, unlock generative AI at the edge, and deploy with expanded hardware support.

Key Takeaways:

Cosmos Reason, a state-of-the-art reasoning vision language model, is now integrated with VSS to enable AI agents to reason like humans and understand the physical world.
VSS 2.4 enhances Q&A accuracy and cross-camera understanding with features like entity deduplication, agentic-based graph traversal, and CUDA-accelerated graph databases.
The Event Reviewer feature allows VSS to act as an intelligent add-on to computer vision pipelines for low-latency alerts and direct VLM Q&A on video segments, suitable for lightweight deployments and edge platforms.

How to GPU-Accelerate Model Training with CUDA-X Data Science

developer.nvidia.com • Sep 25, 2025

In previous posts on AI in manufacturing and operations, we covered the unique data challenges in the supply chain and how smart feature engineering c...

TL;DR

GPU-accelerated tree-based models offer a balance of accuracy, speed, and interpretability for manufacturing data science, leveraging libraries like NVIDIA cuML to supercharge experimentation and deployment.

Key Takeaways:

Tree-based models are ideal for structured manufacturing data, outperforming neural networks in this domain.
Popular library XGBoost can be 150x faster in inference using the Forest Inference Library from NVIDIA cuML, making it suitable for production use.
Feature importance analysis helps engineers understand variable relationships and identify noise features, offering a robust way to filter out uninformative data.

Microsoft blocks the Israeli military from some cloud and AI services

www.theverge.com • Sep 25, 2025

Microsoft is blocking the Israeli military’s access to cloud and AI services that have been used in the mass surveillance of Palestinian civilians. Mi...

TL;DR

Microsoft has blocked the Israeli military's access to certain cloud and AI services used for mass surveillance of Palestinian civilians.

Key Takeaways:

Microsoft's decision is a result of ongoing reporting by The Guardian and pressure from current and former employees.
The block applies to a 'set of services' used by a unit within the Israeli Ministry of Defense, without impacting Microsoft's cybersecurity work in the region.
Microsoft has fired five employees in connection with protests over its Israeli government contracts.

Spotify is finally taking steps to address its AI slop and clone problem

www.theverge.com • Sep 25, 2025

As AI music generators like Suno and Udio make it easy to churn out passable music, streamers like Spotify are getting flooded with AI-generated conte...

TL;DR

Spotify announces new policies to address AI slop, impersonation, and disclose AI use in music, working with DDEX to develop a metadata standard.

Key Takeaways:

Spotify is developing a new metadata standard with DDEX to disclose AI use in any part of song creation.
15 record labels and music distributors have committed to adopting the AI disclosures.
Spotify removed 75 million spam tracks over the last 12 months, citing tactics like uploading tracks just over 30 seconds long to 'rack up royalty-bearing streams'.

AI Revenue Is Not Value — And ‘95% Failure’ Misses The Real Story

news.crunchbase.com • Sep 25, 2025

Current AI debates regarding booming vendor revenue and high failure rates miss the real question of: Who is actually getting value from AI? In this g...

TL;DR

Revenue for AI vendors does not necessarily translate to customer value as most deployments lack the right conditions for impact and focus on shallow, easy-to-show problems rather than meaningful ones.

Key Takeaways:

Measuring AI value is critical and can be done through metrics such as cost reduction, time savings, and quality improvement.
Most AI deployments fail to deliver meaningful returns due to a lack of embedding AI into actual workflows, governance, and measurement.
Successful AI implementations prioritize solving real business problems, instrumenting outcomes, and scaling only when numbers hold steady.

Meta Poaches OpenAI Scientist to Help Lead AI Lab

www.wired.com • Sep 25, 2025

Yang Song, who previously led the strategic explorations team at OpenAI, is the new “research principal” of Meta Superintelligence Labs....

TL;DR

Meta's AI research division, Meta Superintelligence Labs (MSL), has appointed Yang Song, a high-ranking researcher from OpenAI, as the research principal after a hiring blitz by Mark Zuckerberg.

Key Takeaways:

Meta has recruited at least 11 top researchers from OpenAI, Google, and Anthropic to support its AI efforts.
The move reinforces Meta's efforts to establish a prominent AI research division, with significant investment in its Superintelligence Labs initiative.
Yang Song's experience in developing techniques for processing large datasets and his work on DALL-E 2 image generation will contribute to Meta's AI research.

AI-Directed Movie The Sweet Idleness Drops Trailer After Tilly Norwood - Variety

variety.com • 15h ago

AI-Directed Movie The Sweet Idleness Drops Trailer After Tilly Norwood VarietyFirst An AI Actor, Now The AI Director: Italian Producer Andrea Iervolin...

TL;DR

Italain producer Andrea Iervolino launches a new AI label and announces production of 'The Sweet Idleness', the first film directed by an AI-agent.

Key Takeaways:

The film, directed by the AI-agent 'FellinAI', explores a future where only 1% of humanity works, with the rest enjoying leisure and freedom provided by machines.
The production involves a human supervisor, a 'Human-on-the-Loop', to guide and monitor the AI-directed creative process.
The film's digital actors are created from real people using Actor+, an in-house agency of The Andrea Iervolino Company, and will have a 'social and narrative existence' beyond the film.

Forget just living longer. Eric Topol wants to help Americans live better, too.

www.freethink.com • 16h ago

In "Super Agers," the writer and medical researcher maps a path to extending healthspan with AI, targeted drugs, and lifestyle changes....

TL;DR

Eric Topol advocates for a focus on healthy aging, proposing a 'wellderly' approach through lifestyle changes, advances in medicine, and technology, to significantly reduce age-related diseases.

Key Takeaways:

Less than 1 healthy birthday is celebrated after the age of 65, on average, for Americans, highlighting the need for a shift in societal focus.
Topol suggests combining lifestyle, cell, AI, drug, and vaccine advances to increase 'healthspan', or years lived in good health, reducing age-related maladies such as cancer, cardiovascular illness, and neurodegenerative disease.
Genetic testing, AI, and personalized medicine hold promise in combating age-related diseases, and Topol advocates for a 'primary prevention moonshot' to prioritize prevention over treatment.

Google is blocking AI searches for Trump and dementia

www.theverge.com • Yesterday

Google appears to have blocked AI search results for the query "does trump show signs of dementia" as well as other questions about his mental acuity,...

TL;DR

Google appears to be blocking AI search results for queries about Trump's potential dementia, while showing results for similar searches about other public figures.

Key Takeaways:

Google has limited AI search results for questions about President Trump's mental acuity, even though it provides AI results for similar searches about other presidents.
The company's behavior is inconsistent, as it displays AI overviews for queries about Biden and Alzheimer's, but not for similar searches about Trump.
Google may be concerned about accurately presenting information on a sensitive subject or avoiding a potential backlash from the president or his supporters.

Walmart CEO expects AI will 'change literally every job' - not just engineering

www.zdnet.com • Yesterday

Workers all over the economy will experience AI-related shifts, Walmart CEO Doug McMillon said....

TL;DR

Walmart CEO Doug McMillon predicts that AI will 'change literally every job,' but believes it's a transformation opportunity rather than a replacement for employees.

Key Takeaways:

According to Walmart CEO Doug McMillon, AI will rapidly transform all jobs, but it doesn't necessarily mean job layoffs are inevitable.
Recent data shows that AI is transforming job requirements more than replacing jobs, with 84% of executives surveyed believing AI will alter existing roles within their companies.
Employers will likely bear the responsibility for managing the impacts of AI on the labor market, emphasizing the need for upskilling and training for workers.

Lufthansa to cut 4,000 jobs as airline turns to AI to boost efficiency - CNBC

www.cnbc.com • Sep 29, 2025

Lufthansa to cut 4,000 jobs as airline turns to AI to boost efficiency CNBCLufthansa to cut thousands of jobs as it seeks to boost profitability Fox B...

TL;DR

Lufthansa plans to cut 4,000 jobs by 2030 and leverage AI to boost efficiency as part of a broader restructuring strategy.

Key Takeaways:

Lufthansa aims to increase profitability through digitization, AI, and automation, eliminating primarily administrative roles.
The company expects adjusted operating margin to reach 8-10% from 2028 and adjusted free cash flow of more than 2.5 billion euros a year.
Lufthansa joins other companies like Klarna, Salesforce, and Accenture in citing AI as part of their restructuring strategies.

Exclusive: Trucker’s Son Bucks Logistics Funding Decline with $40M Raise For Startup Alvys

news.crunchbase.com • Sep 29, 2025

Alvys, an AI-powered logistics software provider, has raised $40 million in an RTP Global-led Series B funding, the company tells Crunchbase News excl...

TL;DR

Alvys, an AI-powered logistics software provider, raises $40 million in Series B funding to help trucking companies operate more efficiently.

Key Takeaways:

Alvys has raised a total of $77 million to date, with a valuation that was described as an up round.
Logistics and supply chain management startups have seen a significant decline in funding, with $5.7 billion raised across 469 deals in 2025, a drop from $6.7 billion in 2023.
Alvys has tripled its revenue in each of the past two years and is on track to double it again in 2025, with over 1,000 customers and a goal of becoming the logistics 'operating system'.

Marissa Mayer Is Dissolving Her Sunshine Startup Lab

www.wired.com • Sep 29, 2025

After seven rocky years, the company’s assets will be sold to Dazzle, a new AI firm that Mayer founded....

TL;DR

Marissa Mayer's consumer AI startup Sunshine is shutting down and its assets are being sold to a new entity founded by Mayer, called Dazzle.

Key Takeaways:

Sunshine, a consumer AI startup founded by Marissa Mayer, is shutting down after failing to gain traction with its products.
99% of shareholders have signed off on the sale to Dazzle, a new entity that promises to build a new kind of AI personal assistant.
Sunshine's 15 employees are expected to find new roles at Dazzle, and Mayer is eyeing a new opportunity for AI innovation through Dazzle.

Investors look past AI hype to long-term opportunities from government spending - Reuters

www.reuters.com • Sep 29, 2025

Investors look past AI hype to long-term opportunities from government spending ReutersView Full Coverage on Google News...

TL;DR

Investors are shifting focus from AI hype to long-term opportunities presented by government spending in infrastructure, energy, healthcare, and defence, seen as key to market performance.

Key Takeaways:

Government fiscal stimulus is considered crucial for market performance, and investors are eyeing its benefits across various industries.
Industries such as infrastructure, energy, healthcare, and defence are gaining interest as potential beneficiaries of government spending.
Investors appear to be moving past the AI hype, focusing instead on tangible long-term opportunities arising from government policy and spending.

Apple’s ‘Veritas’ chatbot is reportedly an employee-only test of Siri’s AI upgrades

www.theverge.com • Sep 28, 2025

According to Bloomberg’s Mark Gurman Apple is testing Siri’s upcoming revamp using an internal chatbot called Veritas. The company’s struggles as it t...

TL;DR

Apple is testing Siri's AI upgrades using an internal chatbot called Veritas, with no plans for a public release and possibly relying on Google's Gemini for AI search instead.

Key Takeaways:

Apple's internal chatbot, Veritas, resembles chatbot apps like ChatGPT and Gemini, allowing employees to test and collect feedback on new Siri features.
Veritas may not be released to consumers, with Apple instead planning to rely on Google's Gemini for its AI-powered search.
The Veritas testing suggests Apple is struggling to keep pace in the AI race, with its next-gen Siri having been delayed multiple times.

How generative AI boosters are trying to break into Hollywood

www.theverge.com • Sep 28, 2025

This is The Stepback, a weekly newsletter breaking down one essential story from the tech world. For more on the intersection of entertainment and tec...

TL;DR

Despite concerns over copyright infringement and job displacement, generative AI is being aggressively pursued by Hollywood studios and tech companies to reduce production costs and democratize filmmaking.

Key Takeaways:

Generative AI startups are partnering with major studios to produce affordable content, but technical limitations and copyright concerns remain major hurdles.
OpenAI's $30 million feature-length movie project, Critterz, and Amazon's investment in Showrunner highlight the growing interest in AI-generated content.
As AI-generated content gains traction, there are concerns over the potential elimination or reduction of jobs in the entertainment industry.

Amazon blamed AI for layoffs, then hired cheap H1-B workers, senators allege - Ars Technica

arstechnica.com • Sep 25, 2025

Amazon blamed AI for layoffs, then hired cheap H1-B workers, senators allege Ars TechnicaExclusive | Senators Want Answers From Big Tech on H-1B Worke...

TL;DR

Senators and the White House are pressuring tech companies to explain their hiring practices of foreign H-1B workers following mass layoffs of American employees.

Key Takeaways:

Amazon, Meta, Apple, Google, and Microsoft have been asked to provide information on their hiring practices of H-1B workers and potential pay disparities with American employees.
Recent layoffs by these tech companies, combined with hiring of over 20,000 H-1B workers, have raised concerns about favoring foreign labor over American workers.
The White House has implemented a $100,000 fee for companies sponsoring new H-1B employees to bring them into the US, aiming to stop the program from being exploited to replace American workers.