OpenAI returns to open source roots with new models gpt-oss-120b and gpt-oss-20b
Enterprises can use a powerful, near topline OpenAI LLM on their hardware totally privately and securely, without sending data to the cloud....
Fyra's Brief
DeepMind reveals Genie 3, a world model that could be the key to reaching AGI
Google DeepMind has revealed Genie 3, its latest foundation world model that the AI lab says presents a crucial stepping stone on the path to artifici...
Fyra's Brief
Create personal illustrated storybooks in the Gemini app
Article URL: https://blog.google/products/gemini/storybooks/ Comments URL: https://news.ycombinator.com/item?id=44804411 Points: 166 # Comments: 49...
Fyra's Brief
Fyra's Brief
Trending AI Repos & Tools
gpt-oss-120b and gpt-oss-20b open-weight language models Discussion | Link...
Muapi is an AI image and video API platform where users can generate videos and animate images from text prompts...
No more need for super-complex regular expression in the -ot option! Just do `--cpu-moe` or `--n-cpu-moe #` and reduce the number until the model no l...
Lots of people have put instructions into their CLAUDE.md telling it not to say "you're right" I've instead done it with this hook. Whenever it spots...
Community talk
🚀 OpenAI released their open-weight models!!!
Fast and local open source TTS engine. 20+ languages, multiple voices. Model size 25MB to 65MB. Can train on new voices.
Kitten TTS : SOTA Super-tiny TTS Model (Less than 25 MB)
gpt-oss just dropped. OpenAI’s open-weight models are wild. 120B = o4-mini. 20B runs on MacBooks.
gpt-oss-120b outperforms DeepSeek-R1-0528 in benchmarks
Claude Opus 4.1 just launched—thoughts?
One-Minute Daily AI News 8/4/2025
Open-weights just beat Opus 4.1 on today’s benchmarks (AIME’25, GPQA, MMLU)
The new GPT-OSS models have extremely high hallucination rates.
If the open source model is this good, GPT5 will probably be INSANE
DeepMind Genie3 architecture speculation
LLMs Are Getting Dumber? Let’s Talk About Context Rot.
We prepared a collection of Claude code subagents for production-ready workflows.
Subagents are amazing
Meet Claude Opus 4.1
Claude Code + MCPs for development
Routing Claude Code subagents to OpenAI models (GPT-4o, o3, etc.)
Custom Instruction for ChatGPT
DFLoat11 Quantization for Qwen-Image Drops – Run It on 17GB VRAM with CPU Offloading!
Has ChatGPT lost its personality recently? Or is it just me?