Weaponized curiosity. CTO @ https://xitter.cf/t.co/6gglPaqDol
Riga, Latvia
Joined March 2010
- Tweets13.9K
- Following402
- Followers1.2K
- Likes2K
I am not chief security officer of Anthropic but I think we need just hold for around 3-5 years with may be peak at around 18 month. Unless as other Anthropic employees suggested we are killed earlier.
What a time to be alive!
皆様に、今夜に対応して頂きたい大至急のお願いです。
多くのウェブサイトで使用されているPHPに関し、PHPのセキュリティチームが、重大な２件の脆弱性に関する注意喚起を行っています。
公開しているウェブサイトでPHPを使用しているか確認し、今夜、10月21日午後11時（日本時間）に公開される修正パッチについて、可及的速やかに適用できるよう準備をお願いします。
修正パッチを速やかに適用することが困難な場合は、例えばウェブサイト等を一時的に停止させることも、リスクを低減させるための有効策の一つと考えられます。
国家サイバー統括室のXアカウントでも関係する情報を発信していますので、そちらも参考にして下さい。
国家サイバー統括室(注意・警戒情報)（@cyber_forecast） / X
I think the concept of doom trolling emerged in a same way how horror stories emerged. There are people who like to be entertained by the possibility of a catastrophic/scary things to happen while understanding that things themselves are unlikely to happen.
If it would be crypto, CT would say: this is why we are not a serious industry
found the perfect use case for @typesafeai Jev:
instant compaction
in 2026, why is compaction still a summarization prompt?
Jev can make it instant by scoring every tool call and dropping what’s irrelevant
I do review every line of the code that LLM generates, but I caught myself that I am not really reviewing, I am vibe reviewing it. Basically, I am not trying to understand what this code does, just that it looks ugly or suspicious. May be tools should now be built around that.
Vyacheslav retweeted
this post on the chinese internet by a deepseek kernel engineer explains software crafter's daily dread so perfectly
Thinking that my every day should start with 5-10 chess games and depending on the result I should either lock in and work 10 hours straight or just go touch grass cause nothing good will come out today.
Agree
Astra is really, really cool but I cannot current trust it for my present day engineering. At least until I have adjusted. Some random thoughts including code samples of what I recovered from my traces. lucumr.pocoo.org/2026/9/7/as…
“If you feel like you’re not shipping enough, you just need more agents”. No the problem is that current LLM coding flows are either: you are producing shit that doesn’t work, or you literally go line by line reviewing AI code and making a lot of iterations.
I just published a manifesto for all the developers out there who use an AI coding tool, but feel like they're not shipping much faster
I wrote 10 principles for changing how you build software with AI, based on what I've seen work for teams across Amazon
kiro.dev/topics/frontier-eng…
Vyacheslav retweeted
Terence Tao just posted this really insightful thread which seems to have been written in reaction to the Navier Stokes announcement. Please give it a read. mathstodon.xyz/@tao/11723732…
As models get more powerful sometimes they feel even less useful, I'm struggling to use Astra because any kind of small idea is built like enterprise project which takes 50% week limit and 8 hours.
So I guess you can brute force math problems with generating Lean code this way, where the types matter not the code/proof (especially if you got tips from your user sessions, lol), but I don't feel any progress for coding for a while. At times it feels it becomes worse.