🚨GPT6.2 Astra will drop today! > Tibo hinted it may be a big parameter model > Maybe 10T > costs will be like opus 5.5 and absolutely cheaper than GPT6 Astra https://t.co/nTEoPBwbaf

new X update apparently, the s = parameter in a tweet link can reveal which device was used to post it. → s=19 - android → s=20 - browser → s=46 - iphone but that's not all. the part after the ? in an X link contain tracking parameters that may reveal how the link was shared or accessed. something worth knowing before sharing links publicly. stay aware. stay safe.

zed pro free for 1 year students, this one's for you. get access to powerful ai models like gemini 4 argon, gpt 6.1 sol, and upcoming models. → $10 monthly ai credits ($120/year) → unlimited ai edit predictions → real-time multiplayer collaboration → access to zed pro features and hosted ai models eligibility: 18+ students enrolled at accredited universities, verified university email, and a github account older than 30 days. apply here: https://t.co/sbB5iM6MZh available to eligible students worldwide. set it up, choose your model, and start coding. 1 year of zed pro, completely free.

🚨 DeepSeek V4.1 Flash-1017 is running 250 tokens/s It's faster faster than GPT-6.1 Sol in Ultrafast mode, GPT-6 Astra. Claude Opus 5.5 And the Quality that feels close to GPT-6 Astra and Claude Opus 5.5 Do we even need to pay premium prices anymore? https://t.co/vhIZY4I7ES

600 OpenAI employees cashed out $6.6B in a single tender last October. About 75 sold the full $30M cap. The other 525 averaged roughly $8.3M each. The sale priced OpenAI at $500B. The company is now valued at $852B and has confidentially filed for an IPO. No shares have traded publicly. Google's 2004 IPO made about 1,000 paper millionaires. OpenAI made 600 cash ones while still private. Staff had to hold stock for two years before selling, so anyone who joined after ChatGPT launched was locked out of this window. Paper wealth waits for a bell. Cash doesn't. The bell hasn't rung. The money already left the building.

testnet tokens got reset for everyone, and people are still grinding like crazy 😭 bro, it's a testnet token with no value right now. > you literally can't do anything with it outside the testnet. i get that everyone wants to stay early and hopes their efforts pay off someday > but why are we treating these tokens like real money? 💀 are we actually here to test the network and help it grow, or are we just farming everything because we want an airdrop? just asking lol. @Ismail5656th @ericargent31113

🚨 ANTHROPIC'S IPO COULD MAKE 4,257 PEOPLE MILLIONAIRES OVERNIGHT. Not just because of the valuation. Because of how many people own a piece of it. → 7 co-founders become billionaires → 50 early employees reach $100M+ → 1,200 employees reach $10M to $99M → Roughly 3,000 more reach millions Total: 4,257 new millionaires. For context, SpaceX's IPO created about 4,000 millionaires, with roughly 400 crossing the $100M mark. Anthropic is doing it with fewer people and a wider spread. CEO Dario Amodei is worried. He has said the financial incentive could pull focus away from the mission. The company even asks candidates in interviews: if Anthropic gave up AI development for safety and the stock went to zero, how would you feel? When a company creates this much wealth before it even goes public, you are watching one of the largest wealth creation events in history.

The rumors of OpenAI achieving recursive self improvement appear to be correct. I'm seeing massive speed gains out of Sol 6.1 this morning. I'm going to test out Astra 6 next. It appears that OpenAI is finally looking to compete with Anthropic. I bet we hear from Tibo about this today. What's your opinion on this

🚨 GPT-6.1 SOL IN CODEX IS BECOMING UNUSABLE 💀 “Reconnecting… waiting for network” AGAIN?! OpenAI, fix this mess. I’m tired of fighting Codex instead of getting work done. 😭 https://t.co/5UuwUJAIns

🚨 I'm Cancelling My Chatgpt Subscription I'm seriously considering switching to Claude Opus 5.5. I love OpenAI's models, but the experience lately has been frustrating. GPT-6 Astra has become almost unusable for me. The limits are frustrating, and even with usage remaining, you can hit “model at capacity” and lose hours of productivity. GPT-6.2 Sol is also painfully slow. You ask it to handle actual work, and it can take hours to finish. Then suddenly, you get “Your network is having issues. Try again” even while using 5G. I genuinely love OpenAI's models, and they have some incredible technology. But great models mean nothing if you can't rely on them when you need them most. OpenAI needs to improve speed, usage limits, and reliability. Otherwise, more users will keep moving to Anthropic. Anyone else feel like Codex just isn't worth paying for anymore?