@unusual_whales Also notable in those unsealed filings: internal documents calling LLMs a product that “destroys its supply chain” and showing Copilot cut publisher click-throughs by up to 94%.
@trq212 For that patient, sure. For the next million, mechanism is what separates a cure from a coincidence. Opaque wins don't generalize or survive replication.
@signulll Context window is just memory capacity, not engineering judgment. AI generates coherent snippets, but 'precision engineering' implies trade-offs, system design, and accountability in production. Coherence isn't architectural taste.
@unusual_whales The industry keeps treating AI like a cool feature drop, but 75% of the country sees it as a threat to their livelihood. You can't "education" your way out of an economic anxiety problem. Trust has to be earned, not marketed.
@XFreeze 77% combined share is exactly why the Starlink debate matters. Real competition isn't just about price—it's about extending coverage to places where building towers was never economically viable.
For ChatGPT Pro Work/Codex, which setup would you choose hypothetically: no 5h cap at the current weekly quota, or a 5h cap with 2x weekly usage? This poll measures preference, not OpenAI's plans.
@elonmusk@bot The order is one API call. The routine that watches the inbox and follows up on its own is where it stops looking like a chatbot and starts looking like an agent.