Chris 🇨🇦
@llm_wizard
Working on cool open-source AI stuff @ NVIDIA Views my own.
قد يعجبك
People who conflate AI safety with doomers are ngmi
Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,…
these are the takes i see on twitter. in the real world, i hear that engineers feel like they're getting addicted to something that is making them worse at their job. truth is somewhere in-between i'm sure.
We are so honored. NVIDIA DGX Spark was just named one of @TIME's Best Inventions of 2025 🏆 Our new desktop AI supercomputer will deliver an unprecedented 1 petaflop of AI compute, right to your workspace, making state-of-the-art AI more accessible than ever. 🙌 We can’t…
this tho
nvidia is the western qwen in terms of open releases but yall are not ready for this conversation
> GPT 2 > Alpaca (LoRA) > Llama 2 > DSR1 > Kimi K2 No question.
the top 5 most influential LLM releases that defined opensource AI > LLaMA 2 > Mistral 7B > LLaMA 3 > Qwen 2.5 > DeepSeek R1
All this does is expose that Nikita has never been part of deep internet culture, only Twitter.
X has always been the origin of internet culture. And as part of that, we are now exploring new ways to recognize and reward creators who were the original catalyst of a global trend or discussion. This will be forward-looking on anything new happening on the app.
Progress toward the neural kernel.
New APPLE paper says a small base model plus fetched memories can act like a bigger one. With about 10% extra fetched parameters, a 160M model matches models over 2x its size. Packing all facts into fixed weights wastes memory and compute because each query needs very little.…
arxiv.org/abs/2309.08632 - but every part of the training stack
- pretrain on math and reasoning dialog - mid train on math and reasoning traces - release as base model LLM RL Researchers: WE HAVE DISCOVERED ALIEN INTELLIGENCE BEYOND OUR COMPREHENSION WITH RLVR!
Watching Cursor with Sonnet 4.5 saying: "This is tedious" and then creating a script for itself is the most black mirror shit in the world.
Theo's video provides hilariously apt justification for this tier list, but I just love Claude Code's vibes so I'm still mad.
I made a tier list of every vibe coding app and I’m sure everyone will agree with me
Honestly, I go to conferences these days to see what people have *stopped* talking about in AI and what has been absorbed into the collective intelligence.
Ye ye ye ye. This.
RE: the agent/workflow debate Agents and workflows are a spectrum. A system can be more or less 'agentic'. A pure 'agent' is too volatile to be sent to production - you need a bit of determinism to rein it in.
United States الاتجاهات
- 1. Bills 125K posts
- 2. Falcons 40.4K posts
- 3. Josh Allen 20.1K posts
- 4. Snell 10.4K posts
- 5. Bears 55.8K posts
- 6. Bijan 24.4K posts
- 7. phil 139K posts
- 8. AFC East 4,605 posts
- 9. Joe Brady 3,918 posts
- 10. Caleb 35.2K posts
- 11. McDermott 5,674 posts
- 12. #RiseUp 1,688 posts
- 13. Drake London 6,291 posts
- 14. #RaiseHail 6,457 posts
- 15. #NLCS 7,978 posts
- 16. Commanders 35.7K posts
- 17. Freddie 15K posts
- 18. James Cook 4,156 posts
- 19. Penix 5,881 posts
- 20. Chris Moore 2,045 posts
Something went wrong.
Something went wrong.