Igor retweeted
Our Codex dashboards are showing increased rate of users hitting rate limits and since we don't fully understand why I have made the cautious decision of resetting the usage limits for all plans. Enjoy.
I also wanted to celebrate us finding a pocket of fraudulent accounts that we banned and have helped us regain some compute. The fight against abuse never stops, but it's important to mark the moment and make it a little shared victory.
Igor retweeted
the best startup founders in 2026 won't be the best coders
they'll be the best Product Managers
here's the full playbook:
1. pick a problem you personally have. you need to understand your user better than anyone else, and the easiest way to do that is to be your own user
2. spin up a landing page with AI fast. make it about the pain and the fix, no fluff. and make signup 2 clicks max, remove every possible friction
3. learn to write specs, describe features and explain edge cases. that's all you really need now
4. let AI build the first version. your job is to direct, not develop
5. use whatever stack you're most comfortable with. just ship. don't overthink the tech decision - that's not your job anymore
6. host on the simplest platform possible - Vercel, Railway, whatever - don't touch servers
7. test thoroughly - especially edge cases. you're a QA now
8. charge from day 0. free users give you nothing but false hope
9. list your product on every directory you can find. it compounds your SEO and builds DR over time
10. start building in public. share the micro wins, the losses, everything. invite people to try it through DMs. I did this for all my products
11. do customer support yourself, always. redirect questions to your socials, never automate this. every support request is a product insight - that's why you do it yourself, always (I still do this for all 5 of my products)
12. automate everything else you do more than twice
13. improve one thing in your product every single day once users start coming in. watch your competitors' G2 and Trustpilot reviews closely for ideas
14. a PM's most important skill is deciding what NOT to build. for every feature request you get, say no to 9 of them - build less, but build it right
15. every week, build one free tool targeting a keyword that benefits you. it compounds over time
16. work your SEO early. build your affiliate setup early. only run paid acquisition after you're sure about PMF
every day your only 2 priorities are:
- get more users
- and keep the ones you have
that's it. everything else is noise
Igor retweeted
Cool, X has launched a new feature that allows you to filter an account's posts by latest and most liked, and select which type of posts.
Igor retweeted
China saved opensource LLMs
and not only that, it made the FRONTIER opensource
between July 16th and today, these are the major releases:
> Kimi-K2-Thinking (1T-A32B)
> MiniMax M2
> DeepSeek V3.2
> GLM-4.6 (335B-A32B)
> Qwen3-VL-30B-A3B (Instruct & Thinking)
> Qwen3-VL-235B-A22B (Instruct & Thinking)
> Qwen3-Next 80B-A3B (Instruct & Thinking)
> GLM-4.5V (VLM, 106B-A12B)
> DeepSeek V3.1
> Doubao 1.6-Vision (multimodal, tool-calling)
> Doubao Translation 1.5 (ByteDance, 28 Languages)
> ERNIE X1.1 (Baidu, Reasoning)
> Hunyuan-MT-7B & Chimera-7B (Tencent Translation Specialists)
> MiniCPM-V 4.5 (8B), Tiny but GPT-4o-level VLM
> InternVL 3.5 (MASSIVE Multimodal Family of Models, 1B to 241B Sizes)
> Step-3 (VLM, 321B/38B)
> SenseNova V6.5 (SenseTime, Multimodal)
> GLM-4.5 Air (Base & Instruct, 106B-A12B)
> GLM-4.5 (Base & Instruct, 335B-A32B)
> Qwen 3-Coder-30B-A3B (Instruct & Thinking)
> Qwen3-Coder-480B-A35B (Instruct & Thinking)
> Qwen3-30B-A3B-2507 (Instruct & Thinking)
> Qwen3-235B-A22B-2507 (Instruct & Thinking)
> Kimi K2 (1T-A32B)
US & EU need to do better






