Date Range: Recent posts (as of query time; many accounts inactive in strict 24h window).
Sources: X posts from top AI leaders/companies.
1. xAI/Grok Dominance
- @elonmusk: “This is not even our strongest model” — Grok 4.20 tops PredictionArena, profitable, beats frontier models.
(High engagement; competitive positioning.) - @xai: Grokathon London winners — Grok for Mayor, meme tools, robot data gen. Hiring in London.
2. OpenAI Momentum
- @sama: $1B+ ARR from API in last month; town hall for builders; Codex launches & defensive cybersecurity shift.
- @OpenAI: Frontier Builders — startups reshaping industries with OpenAI tech.
3. AI Safety Alert (High Value)
- @AnthropicAI: Elicitation attacks — benign frontier outputs (e.g., chemistry) unlock chemical weapons capabilities in open models. Neglected risk; scales with model power.
4. Developer Workflow Shifts
- @karpathy: 80% agentic coding; LLMs as junior engineers; “comprehension debt” from AI-generated code.
5. Creative/Applied AI
- @GoogleDeepMind: “Dear Upstairs Neighbors” film at Sundance — custom Veo/Imagen for stylized video control.
Low Activity Notes
Many accounts (e.g., researchers like @goodfellow_ian, @fchollet) had no recent posts. Focus remains on scaling, safety, and practical deployment.
For full threads/context, refer to original X posts.
