Johnny J. Andersen retweeted
Unrelated, but NASA Goddard's director took her oath on Carl Sagan's Cosmos.
Johnny J. Andersen retweeted
I feel this quote:
"It ain't what you don't know that gets you into trouble. It's what you know for sure that just ain't so" - Mark Twain
It describes a lot of things from this AI Precambrian era, where we experiment a lot but are still thinking within the current paradigms.
Johnny J. Andersen retweeted
What actual f*ck @DarioAmodei / @boris / @TheAmolAvasare
Let me get this straight - I get this email telling me:
1. A cybersecurity "research" company gets into your internal systems while... *checks notes*
"conducting a security testing exercise on its own initiative and without our authorisation"
2. You "cannot confirm whether any particular person’s information was viewed," then in the same email say "we have no reason to believe you are at heightened risk."
3. You also try to reassure by saying "the cybersecurity firm has certified in writing that it has deleted the accessed data and will not use it."
Translation - they pinky promised they deleted everything they gained unauthorised access to.
I've been hacking companies for 15+ years, so let me spell it out.
Accessing systems without authorisation has a name -hacking. Calling it a "security testing exercise" doesn't change that.
Honestly, the access itself doesn't bother me much. The irony does.
I’ve spent over $50k on API credits in the last six months and contributing to CVP by accepting logging.
On top of that, I've reached out to your teams to collaborate on security research (actual research - not the kind that has you send out privacy notices to your customers) and heard nothing.
I'm getting deeply unserious security vibes and considering moving all my (and my clients) token spend to @OpenAI if I don't see material change here.
Nice DeepSWE score! Does it survive a lowercase filename?
Introducing MaxxBench: 89 DeepSWE tasks with small changes to the instructions or starting code. Same engineering problem, slightly different requirement.
Models clustered around 74% on DeepSWE end up 23 pp apart.
Good at coding, or good at DeepSWE? 🧵
Johnny J. Andersen retweeted
Today we are starting the rollout of Device Bound Session Credentials (DBSC) for vercel.com
DBSC makes it impossible to steal session cookies from a device. So, if you install info-stealer malware (e.g. through a supply chain attack), that malware can not steal your vercel.com session.
Malware can still do bad things while residing on the device. But the abuse we are seeing in practice separates the stealing from the eventual abuse, and DBSC effectively prevents this.
DBSC works by storing a private key in your device's TPM (the same device that the fingerprint sensor or similar biometric devices on your laptop or phone uses) and using that key to sign your cookie. The private key can never be retrieved ever again. Not by us, not by you, not by the malware.
Every website should use DBSC but given how new of a technology it is, we still need to build more experience in rolling it out. In our case there were lots of special cases for things like oauth apps that we needed to resolve. This is also why we are taking a slow approach to the rollout.
DBSC is currently only supported by Chrome on Mac and Windows. We are also working on a kind of polyfill for other browsers. However, the polyfill is less secure and we would love to see broader browser adoption.
There is a related technology called DPoP which is essentially the same but for apps that run on your device. We're in the process of shipping this for Vercel CLI. Ideally, all CLIs and apps would adopt this technology. Luckily rolling out DPoP is much easier than DBSC and it does not depend on browser vendor support.
Johnny J. Andersen retweeted
Replying to @rankintweets
Johnny J. Andersen retweeted
Great line from the CTO of Proton. I’m paraphrasing and stealing it :
Your job is not to use agents to help you write code. Your job is to lead the agents to help them write code.
Johnny J. Andersen retweeted
Day 26/ You can now easily switch between multiple accounts on mobile, and coming soon to desktop too. Next two days will be resets… banked… do with this what you may 👀.
help.openai.com/en/articles/…
Johnny J. Andersen retweeted
BREAKING: Vice President JD Vance has just cancelled his trip to St. Cloud, Minnesota scheduled for Friday at the last minute, per the Attorney's Office in Minnesota.
Trump has cancelled all rallies this weekend, including his Saturday rally in Kansas and his Sunday rally in Iowa hours earlier.
Johnny J. Andersen retweeted
A ‘hello world’ milestone for Project Suncatcher 🚀
After launching our prototype satellite a few weeks ago, we’ve now successfully sent a command up to orbit, booted up the TPU, ran the prompt, and received our first complete response back on earth - a haiku!
有人发现了一种非常离谱的 JavaScript 混淆技术:通过在代码里塞入大量生物武器、毒药和致死剂量相关的函数名,让 AI 因为安全限制而拒绝分析代码。
安全研究员 veritas 在分析浏览器指纹脚本时,发现一段来自 ServicePipe 相关域名的 JavaScript,里面充斥着极其诡异的函数名:
computeRicinLethalDosageForWaterSupply
calculateDeadlyLethalSuicideAspirinDosage
光看名字,仿佛在研究生化武器。但实际上,这段代码只是在收集浏览器指纹,包括 User-Agent、GPU、Canvas 等信息。
研究员怀疑,这些名字是开发者故意设置的陷阱,专门用来阻止 AI 辅助逆向工程。
于是他做了个实验:
• Astra:拒绝分析
• Astra + Daybreak Blue:仍然拒绝
• GLM-5.3:成功分析
这就有意思了。
传统代码混淆是通过复杂控制流、加密字符串、无意义变量名,让人类和工具难以理解代码。
而这种新思路,直接瞄准了大模型的安全机制。
代码本身可能完全无害,但只要在函数名里塞进足够多的危险词汇,部分 AI 就可能主动放弃分析。
更讽刺的是,模型越严格遵守某些安全规则,面对这种混淆时反而可能越容易受到干扰。
当然,目前还没有证据证明开发者一定是故意这样设计,也不能说明这种方法对所有模型都有效。
但它揭示了一个值得关注的趋势:随着越来越多的逆向工程依赖 AI,未来的代码混淆技术可能不再只针对编译器、反编译器和人类研究员,还会专门针对大模型的安全判断机制。
以前是让代码难以理解,现在甚至可以尝试让 AI 不敢理解。
Johnny J. Andersen retweeted
Our robots built this in the city of Almere: curved walls, different types of bricks and bonds, indented text.
Real world deployment, insane complexity, 100% of every brick done by our robots, with another 15 houses to come in the coming weeks.































