2026-08-25

RT Tim Ferriss Has anyone used Matic (robot vacuum)? If so, from 0-10 (no 7 allowed), how strongly would you recommend and why? Thanks! 🙏

RT CatFly Claude Code 源码泄露事件后续越来越精彩了。 有人拿泄露的源码丢给 OpenAI 的 Codex 分析，竟然找到了 Claude Code 疯狂消耗 token 的元凶——autoCompact（自动上下文压缩）机制在失败后会无限重试，完全没有上限。据源码注释记录，曾有会话连续失败高达 3272 次。 修复方法简单到离谱：加一个 MAX_CONSECUTIVE_AUTOCOMPACT_FAILURES = 3 的限制，连续失败 3 次就停止重试。三行代码，搞定。 打完补丁后，这位老哥表示使用额度恢复正常了——之前被吐槽的"用两下就触发限速"，很可能有一部分就是这个 bug 在背后偷偷烧 token。 仓库地址放下面了。 Lydia Hallie ✨: We're aware people are hitting usage limits in Claude Code way faster than expected. Actively investigating, will share more when we have an update!

RT Cailyn Y. been using @garrytan's gstack office-hours for quite a bit now. it now gives me relevant yc youtube vids and PG's essays based on my situation. it also ruthlessly cuts off any bs if i'm about to spiral into my thoughts. pretty g now

RT clem 🤗 That’s why I usually say that comparing open models with closed-source APIs or products is like comparing apples and oranges. Or comparing an engine with a full car. Or comparing an ingredient with a Michelin dinner (missing ingredients, prep and chef). There’s a lot of scaffolding and tricks that is done behind APIs that make it unfair to compare to a raw open-source model! However that also means that if you put the work to really try to make open-source models work, they can give you much better results than what benchmark suggests (and that’s without even post-training/fine-tuning) Sebastian Raschka: http://x.com/i/article/2038978163389112321

RT Thariq /buddy

GStack does more to try to remember failures and learn from them. Open source lets us create a memory system to self-improve in the open. We're still on a path to a software factory, but we're going to build it in the open.

RT Boris Cherny Re Mistakes happen. As a team, the important thing is to recognize it’s never an individuals’s fault — it’s the process, the culture, or the infra. In this case, there was a manual deploy step that should have been better automated. Our team has made a few improvements to the automation for next time, a couple more on the way.

RT Boris Cherny Re Mistakes happen. As a team, the important thing is to recognize it’s never an individuals’s fault — it’s the process, the culture, or the infra. In this case, there was a manual deploy step that should have been better automated. Our team has made a few improvements to the automation for next time, a couple more on the way.

Proud to be a part of this insane lineup Reva Jariwala: how is this a class? absolutely insane line-up

Have to admit some parts of GStack's prompts do feel like this David J Phillips: "Make no mistakes DO NOT HALLUCINATE. YOU ARE AN EXPERT SOFTWARE ENGINEER"

btw this is not the way to do it, but it's quite funny Josh Cohenzadeh: I will out-accelerate you all My newest repo ~codemaxxed~ is already at 68M+ LOC & over 6,800 commits in just a day Bring it on

Look sometimes you have to be stern Miguel Serna: Best part of going through Claude Code's leaked source is seeing how $380B Anthropic's core engineering strategy is treating their model like a misbehaving child

Philip Johnston: @garrytan @marvinliao The average age of the three @Starcloud_ founders during the batch 18 months ago was 41! 😱