Claude's Fable 5 announcement remains the dominant topic (28134 likes): 'Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits.' Tomorrow is the day. The follow-up about giving users more certainty got 3108 likes.

Sottiaux's reset streak continues to dominate OpenAI engagement: 'Oops... I did it again' (9072 likes) + 'GPT-5.6 Sol confirmed to be an extremely good model' (10876 likes) + transitive rate limit resets (1159 likes). Combined: ~21K likes across three tweets.

Steipete's philosophical question gained traction over the weekend (2502 likes): 'Are we still talking loops or did we shift to graphs yet?' — a genuine architectural question about whether agent orchestration has moved from sequential loops to DAGs. He also built a codexbar icon editor (306 likes) and shared the Codex browser-use pain (317 likes).

Swyx's automation tip (1019 likes): 'btw if you havent set your {codex | claude | gemini | devin} automations to autoresearch how to improve your seo/aeo every week you are really truly missing out.' Thariq's token-saving wisdom (634 likes): 'building prototypes of mockups, schemas, data models, proof of concepts, etc. is the best way to avoid spending tons of tokens before realizing you don't want [what you asked for].'

Levie on cost economics (331 likes): 'The cheaper AI gets, the more opportunity there is for the entire ecosystem — especially including end-customers — to benefit.' Madhu Guru on the enterprise talent gap (264 likes): 'The reason enterprises struggle to go beyond basic chat bots is the talent gap to build harnesses and evals.'

Zara's two insights: building in public tip (206 likes) — 'show the work already happening inside your product' — and meeting recording culture shift (141 likes): 'Just a few years ago, many people were uncomfortable with recording meetings. Now it's just assumed that all business meetings are recorded (not for humans, but for AI).' Rauchg: 'Ship like @shadcn' (193 likes). Peter Yang previewed his Thariq episode dropping today (33 likes).

Sachin Katti (OpenAI Head of Industrial Compute, ex-Intel CTO, ex-Stanford professor) on MAD with Matt Turck, recorded at RACE conference in Paris. The definitive deep-dive on AI compute infrastructure. Key topics: (1) 'Demand far outstrips compute supply — anything we bring online, we consume immediately.' (2) Data centers as giant factories turning electrons into tokens, liquid-cooled football-field-scale. (3) OpenAI spending ~$50B on compute this year; industry ~$700B. (4) Jalapeno custom silicon optimizing tokens-per-watt. (5) Inference is now the majority of compute. (6) 'Our biggest worry is we can't build fast enough — the physical world doesn't move that fast.' (7) Grid investment: 'We are not taking power away from the grid. We invest in new generation.' (8) 'The world of recursion is not far — AI will design the systems it needs to train the next generation of AI.' If you only watch one infrastructure podcast this year, this is it.