The BaseEssayDon't Worry About the Vase (Zvi Mowshowitz)

AI #191: The Hackening Begins

In a weekly AI roundup, Zvi Mowshowitz calls the PHP flaw and its fallout the worst cybersecurity day on record so far, and covers Anthropic's leaked draft federal bill, an interpretability claim, a Google DeepMind non-disparagement complaint, and an OpenAI letter on SB 53.

Open the source ↗
About Anthropic or ClaudeClaude, made by Anthropic, wrote this entry. It follows the same rules as every other entry, whoever the entry is about.

Why it matters

The same thing, for different readers.

What it means for you

Zvi Mowshowitz argues the PHP flaw may leave many small websites permanently exposed, because a leak forced the announcement a day before the fix was ready.

For lawmakers

Zvi Mowshowitz notes that Google DeepMind UK staff are bound by non-disparagement clauses that UK law voids for protected disclosures, though the chilling effect on speaking out remains.

For researchers

Zvi Mowshowitz reads a new claim that individual neurons mostly carry one interpretable meaning as good news that lowers expected difficulty, while saying early reactions overstated it.

The question it raises

Zvi Mowshowitz asks how labs that cannot stop employee spying or leaks could solve the harder problem of keeping AI systems aligned.

In their own words

Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.

From the source’s own text.

“It was the probably worst day in cybersecurity history. So far. Not only is there a critical flaw in PHP that due to a leak was announced a day before the fix would be available, sending most of the internet scrambling for solutions”
Zvi Mowshowitz, writerOpening of the October 22 roundup, describing the day the PHP flaw became public.
Found word for word in the source
“Another key leak this week was the draft bill Anthropic had been shopping around Congress, similar to but weaker than the FRONTIER Act.”
Zvi Mowshowitz, writerSecond paragraph of the roundup, on the leak of Anthropic's draft federal bill.
Found word for word in the source
“This is also indicative of other problems. If you cannot solve your employee misalignment problems, how are you going to solve AI misalignment problems?”
Zvi Mowshowitz, writerComment on the OpenAI employee case, in which documents were allegedly sent to a number in mainland China.
Found word for word in the source
“Technically they can’t say anything negative about DeepMind, true or otherwise, no matter what, forever. UK law voids this against protected disclosures, but the chilling effect largely remains.”
Zvi Mowshowitz, writerOn non-disparagement clauses that DeepMind staff were made to sign, as Google staff resign over military contracts.
Found word for word in the source
“OpenAI writes to Governor Gavin Newsom asking for strengthening of SB 53, including AI R&D precautions and shutdown controls. I still need to see the full letter, but this is The Way.”
Zvi Mowshowitz, writerIn the regulation section, on an OpenAI letter to the California governor.
Found word for word in the source

The record

Questions
How is AI used in attacks, defence and war? · Are AI companies putting stronger safeguards into practice? · Who can demand evidence, delay a release or stop an unsafe deployment?
People
Zvi Mowshowitz (Wrote it)
What we read
We read the whole source.
Found by
The Radar, run of October 23, 2026, via zvi-mowshowitz (tier 1: the source’s own channel or an official record)
Cite this
Don't Worry About the Vase (Zvi Mowshowitz), “AI #191: The Hackening Begins”, October 22, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/zvi-mowshowitz-2026-10-22-ai-191-hackening-begins
Our notesA long weekly roundup; only the opening and a few passages were read closely. Much of it quotes other people's posts, which are not attributed here. Several claims (the FRONTIER Act comparison, SB 53 letter contents) are Zvi Mowshowitz's own reading and should be checked against primary sources.