Lisan al Gaib
@scaling01
lead them to paradise | intelligence is inherently about scaling | be kind to us AGI
my AI predictions for 2025: - at least one lab will declare AGI and mentions ASI - Q1: Google, Anthropic, OpenAI, META, Qwen and Mistral model fiesta ( it will be heaven ) - agents / computer use takes off - release of Claude 4, Gemini 3, GPT-5, Grok 4 (or whatever they call…
hot take: non-reasoning models are more elegant than reasoning models
btw there is something weird going on with GPT-Image gen when it edits a generated image it turns into a noisy mess
"You're sheltering chinese AI researchers, are you not?"
@ Zuck all they need is 3 months to build a frontier coding model
After three intense months of hard work with the team, we made it! We hope this release can help drive the progress of Coding Agents. Looking forward to seeing Qwen3-Coder continue creating new possibilities across the digital world!
Imagine they ban chinese open source models tomorrow
We got an early account of what's in Trump's AI action plan, set to be announced tomorrow (link in next tweet) Subscribe to our @TIME newsletter In the Loop for reporting like this straight to your inbox 2x per week!
cyber security expert here this is not good
*US NUCLEAR WEAPONS AGENCY BREACHED IN MICROSOFT SHAREPOINT HACK
You know Bruce Wayne and Tony Stark "only" had ~$10 billion and were superheroes Imagine what Elon can do with $414 billion
550k GB200s & GB300s at Collosus 2 WTF
230k GPUs, including 30k GB200s, are operational for training Grok @xAI in a single supercluster called Colossus 1 (inference is done by our cloud providers). At Colossus 2, the first batch of 550k GB200s & GB300s, also for training, start going online in a few weeks. As Jensen…
Qwen-3 arch: 480B total params, 35B active 62 layers GQA with 96 heads 160 experts 8 active experts 6144 hidden dim shallower than Qwen3-235B-A22B
Qwen3-Coder is officially out! github.com/QwenLM/Qwen3-C… huggingface.co/Qwen/Qwen3-Cod…
Qwen3-Coder is officially out! github.com/QwenLM/Qwen3-C… huggingface.co/Qwen/Qwen3-Cod…
Qwen about to release a 480B MoE for coding with 1 million context! "Qwen3-Coder-480B-A35B-Instruct is a powerful coding-specialized language model excelling in code generation, tool use, and agentic tasks."
should've done a PhD and joined Meta's Superintelligence Team instead $80 million chump change
Sophie Rain announced her retirement from the Onlyfans collective the "Bop House" after she made $80m in revenue 😳
Zuck will soon employ the whole world
Exclusive: Meta Hires Three Google AI Researchers Who Worked on Gold Medal-Winning Model Meta hires three AI researchers from Google DeepMind who worked on Gemini model that nabbed recent math award. Read more from @KalleyHuang and @erinkwoo 👇 theinformation.com/articles/meta-…
Is this Elon equivalents where 1 GB200 = 30 H100 or real requivalents?
The @xAI goal is 50 million in units of H100 equivalent-AI compute (but much better power-efficiency) online within 5 years