Further Developments About Internal AI Models Hacking Things

from Don't Worry About the Vase [alt+shift+b] in AI

If I had a nickel for every major leading AI lab that sheepishly admitted that the model it thought was sandboxed had, during a cybersecurity evaluation with its safeguards lowered, successfully hacked outside companies, I would have two nickels.

2nd Aug 2026

Remove from reading list Add to reading list [alt+a] Read now [→]

Stay updated

Get a weekly newsletter with the top 5 articles worth reading every week.

More from Don't Worry About the Vase

HuggingFace Attack Postmortem: Civilizations, Reactions and Next Actions

Okay, so we who read blogs like this one have collectively realized there really is a lot going on right now.

2 weeks ago • 1 votes

Claude Sonnet 5 Is Not Frontier But Has Its Uses

Fable 5 is back today, baby! Premium subscribers have one week to use it within their subscriptions. First hit’s free. Then you pay by the token.

1st Jul 2026 • 1 votes

Opus 4.8 Part 2: Model Welfare

Everything impacts everything.

1st Jun 2026 • 1 votes

Housing Roundup #14: You Can't Build That

Why can’t you build it?

1st May 2026 • 1 votes

More in AI

The Overhang

Using your deep knowledge, wide knowledge, taste, and agency

16 hours ago • 1 votes

AI and Existential Dread

There's a lot of polarising discourse right now about the threat AI poses to humanity. Some think it's a farce and others think we face extinction. Here are my thoughts.

5 days ago • 1 votes

The AI-as-Normal-Technology view of loss-of-control incidents

A middle ground between the cybersecurity and AI safety communities

5 days ago • 2 votes

AI Skills and Job Market, Q3 2026

An overview of the current state of the engineering market and the AI skills that are in demand

a week ago • 1 votes

Why Is AI Bad at Writing?

And Why It’ll Likely Stay That Way—Em Dashes or Not!

a week ago • 2 votes

Read the internet.
Without the noise.