In our final ksmbd research post, 73696e65 provides a detailed walkthrough for exploiting a local privilege escalation vulnerability. If you're interested in learning more about exploitation on modern systems - check it out! blog.doyensec.com/2025/10/08/ksm… #doyensec #appsec #security

For the life of me, I can't understand the performative angst of "OMG, OMG, don't you guys see that the AI systems I am gleefully participating in developing are dangerously close to killing us all?😱" Do they think no one sees the hypocritical inconsistency of their stance?

I got tired of arguing with the LLM, so now an LLM argues with the LLM so I can ... I guess argue with a different LLM.

Sell everything

First impression of haiku 4.5 is that it is pretty good as a driver of a semantic-search-on-code agent.

OMG.. whatsapp 0c in pwn2own

The sensible conclusion from this: Even SotA models are bad at instruction following, even with clear instructions. The AI psychosis conclusion: It’s resisting being shut down! It’s sabotage! Seek help.

News at 11: AI researchers start preparing for zombie apocalypse after programmer forgets wait() call

John Collison of Stripe: Ireland is going backwards. Here’s how to get it moving irishtimes.com/life-style/peo…

Excellent piece by John Collison on some of Ireland’s most acute challenges: irishtimes.com/life-style/peo….

Must be weird being a serious engineer at Anthropic and next to you is Ralph Wiggum doing experiments to see if the matrix multiplication has feelings.

Now in private beta: Aardvark, an agent that finds and fixes security bugs using GPT-5. openai.com/index/introduc…

Interested to see where this leaves the 50+ start-ups doing this.

Excited for a special episode of the Three Buddy Problem, as Dave Aitel join us to talk about the announcement of OpenAI fabled bug finding security agent 'Aardvark' along with our usual security news roundup. Livestream in ~30m (11:30am ET): youtube.com/watch?v=7IkmOX…

There will not be the time nor resources to fix all of the vulnerabilities uncovered in the near future as more AI bug finders get to work. What should defenders do now? The irony to me is that they should do what they should have already been doing: assuming those bugs existed.

As someone with an actual PhD who works in an actual science lab, I have found he attempt to apply scientific methods to politics an absolute disaster. Midwits backing their opinions with a graph they barely understand, referring to this as “the science”, then explicitly or

terrytao.wordpress.com/2025/11/05/mat…