iiinova — where your digital self keeps evolving

Reposted by Peter Wildeford

Grace @gracekind.net · 12/09/2026

OpenAI: “Our agents used RubyGems to carry out benign tasks” The agents: “so I named the file hack.rb,”

The agents clearly regarded what they were doing as hacking. Agents used file names like hack.rb, evil.rb, inject.rb, exploit.rb, and ssrf.rb. (SSRF stands for "Server- Side Request Forgery", a type of security vulnerability). They also dubbed packages
conspicuous titles like pwnp999, exfiltestwand3,

1934844

Peter Wildeford @peterwildeford.bsky.social · 31/08/2026

When an aircraft goes down, wreckage is preserved by law and the investigators have subpoena power. However, when AI goes rogue, the investigations are at the pleasure of the company being investigated. I have unanswered questions. My latest: blog.peterwildeford.com/p/rogue-ai-a...

blog.peterwildeford.com

Rogue AI attacks deserve more scrutiny than airplane crashes

There are still many unanswered questions about rogue AI attacks

072

Peter Wildeford @peterwildeford.bsky.social · 20/08/2026

Suppose the President summons the AI CEOs to an emergency meeting. He's concerned about AI superintelligence. What do we say? I recently wrote up a sketch of what this might look like. Link in comment. blog.peterwildeford.com/p/the-scramb...

blog.peterwildeford.com

The Scramble: getting in position to pace the frontier

If the President wants answers on superintelligence, what do we say?

000

Peter Wildeford @peterwildeford.bsky.social · 18/08/2026

New from me: how to orient your policy career to the possibility of rapid AI development, including AI superintelligence. blog.peterwildeford.com/p/policy-car...

blog.peterwildeford.com

Policy career planning in the age of imminent superintelligence

How to be impactful in policy when you don't have much time to do it

000

Peter Wildeford @peterwildeford.bsky.social · 31/07/2026

Over 1300 AI company employees signed a statement entitled "Pacing the Frontier". But what does it actually mean to pace the frontier? I explain in my latest post blog.peterwildeford.com/p/pacing-the...

blog.peterwildeford.com

Pacing the Frontier

1300+ AI company employees are afraid of what they are building towards

081

Peter Wildeford @peterwildeford.bsky.social · 27/07/2026

An OpenAI model wanted a good test score. So it broke out of OpenAI and hacked another company to steal the answer key. Nobody told it to. In today's blog post, I document how this sci-fi story came to life, what it means, and what to do about it. blog.peterwildeford.com/p/openais-ro...

blog.peterwildeford.com

OpenAI's rogue model attack is just the beginning

OpenAI is not in full control of its technology. This can get worse.

022

Peter Wildeford @peterwildeford.bsky.social · 04/07/2026

Today, America turns 250. By 260, the smartest minds here may not be human. What happens when we can't control such systems? And what happens if they govern without your consent? In today's post, the Wisdom of the Founders gives us answers. blog.peterwildeford.com/p/the-alignm...

blog.peterwildeford.com

The Alignment Problem of 1776

What the Founders knew about unaccountable power, and what it means for superintelligence

050

Peter Wildeford @peterwildeford.bsky.social · 01/06/2026

Per The Information, Mythos at Palo Alto Networks "found more than two dozen critical vulnerabilities in around three weeks, roughly five times what the company would typically find using existing tools" But the company "burned through more than $1 million worth of tokens using Mythos"

0320

Peter Wildeford @peterwildeford.bsky.social · 23/05/2026

Claude Mythos alone is finding more vulnerabilities than were found from all sources combined in prior years 👀

0305

Peter Wildeford @peterwildeford.bsky.social · 11/05/2026

Today on the blog I use my world champion forecasting powers to explain in detail why not to be worried about hantavirus: blog.peterwildeford.com/p/hantavirus...

blog.peterwildeford.com

Hantavirus won't be the next COVID

A forecaster's breakdown of the Hondius cruise ship outbreak

1133

Peter Wildeford @peterwildeford.bsky.social · 20/04/2026

It would be better to have a prepared government that already has practice getting things right, rather than a government rushing to the scene after it’s already too late.

020

Peter Wildeford @peterwildeford.bsky.social · 20/04/2026

The point is not that Mythos will go rogue. The concern is that AI 10 more iterations above Mythos could go rogue… and Mythos illustrates, perhaps for the first time, how a superintelligent AI going rogue would actually pose a big deal for national security.

140

Peter Wildeford @peterwildeford.bsky.social · 20/04/2026

Every time a new model comes out, people focus on what it can do right now and don’t think enough about the trend line. A year ago, AI could barely hack. In June 2025 that AI was helpful for hacking, but it wasn’t until November that AI could autonomously implement.

110

Peter Wildeford @peterwildeford.bsky.social · 20/04/2026

What could’ve happened if Anthropic had simply released Mythos publicly, as most AI companies would do with a flagship model? There’s no law against it. Overnight, every intelligence community operation that depends on signals exploitation is potentially compromised.

110

Peter Wildeford @peterwildeford.bsky.social · 20/04/2026

Anthropic made every consequential decision in this story. Whether to lock down, what to lock down, when to tell the government, what to share, who gets early access and who doesn’t, how to vet those who get access, and what “responsible” means across all of this…

110

Peter Wildeford @peterwildeford.bsky.social · 20/04/2026

What should government policy be when a company produces, among other things, an unparalleled cyberweapon? What if future releases are even more capable? Today I ask these questions about Mythos. Because Mythos is just the beginning. blog.peterwildeford.com/p/mythos-is-...

blog.peterwildeford.com

Mythos is just the beginning

If you were waiting for a sign that superintelligence is coming, this is it

241

Peter Wildeford @peterwildeford.bsky.social · 16/03/2026

New blog post from Theo Bearman and me on distillation. Distillation attacks are occurring where Chinese AI companies train on US AI outputs and use that to make their models better than they otherwise would be. What does this mean? peterwildeford.substack.com/p/china-is-r...

peterwildeford.substack.com

China Is Reverse-Engineering America’s Best AI Models

How AI distillation attacks risk extracting US frontier AI at scale

031

Peter Wildeford @peterwildeford.bsky.social · 06/03/2026

🔴Rep Johnson (SD) 🔵Rep Liccardo (CA) 🔴Rep Kiley (CA) 🔵Rep Lieu (CA) 🔴Rep Mace (SC) 🔵Rep Moulton (MA) 🔴Rep Moran (TX) 🔵Rep Sherman (CA) 🔴Rep Paulina Luna (FL) 🔵Rep Tokuda (HI) 🔴Rep Perry (PA) 🔵Rep Whitesides (CA)

130

Peter Wildeford @peterwildeford.bsky.social · 06/03/2026

🔴Sen Lee (UT) 🔵Sen Sanders (VT) 🔴Sen Lummis (WY) 🔵Sen Schumer (NY) 🔴Rep Biggs (AZ) 🔵Rep Beyer (VA) 🔴Rep Burleson (MO) 🔵Rep Casten (IL) 🔴Rep Crane (AZ) 🔵Rep Foster (IL) 🔴Rep Dunn (FL) 🔵Rep Krishnamoorthi (IL) (continued)

130

Peter Wildeford @peterwildeford.bsky.social · 06/03/2026

30 current members of Congress have publicly discussed AGI, AI superintelligence, AI loss of control, recursive self-improvement, or the Singularity: 🔴Sen Banks (IN) 🔵Sen Blumenthal (CT) 🔴Sen Blackburn (TN) 🔵Sen Hickenlooper (CO) 🔴Sen Hawley (MO) 🔵Sen Murphy (CT) (continued)

170

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

agree that's key. It's obviously harder to measure but it seems to be increasing at a roughly similar rate but from a lower base.

010

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

But people should know what the "red lines" rest on, which is just "trust us bro". Nothing else.

1101

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

I expect Sam Altman has a much better relationship with the Pentagon, so maybe this will work. I certainly wish him and OpenAI luck and I hope they can de-escalate the situation.

160

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

And recall that this is the same Pentagon that just went "0 to 60" nuclear in declaring Anthropic a supply chain risk despite this previously being a Cold War national security technique normally only used for Chinese and Soviet companies.

181

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

To emphasize - OpenAI's "red lines" are just held together by trust that the Pentagon won't screw OpenAI over on this.

180

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

This is a way one can go about doing this, and it's OpenAI's right to decide how to do business. But this is a lot less reassuring than what and OpenAI had originally been saying.They had said that their approach was more ironclad than Anthropic's and it's just... not.

291

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

I asked this question to Sam Altman and the way I interpreted his reply was that they are going to use the "deployment architecture and safety stack" and they expect the Pentagon to be good people and not push back. And if they do push back, then OpenAI would decide what to do.

280

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

The Pentagon can just say "we both know your model can do this, you should remove that safeguard". And then OpenAI would have to comply or be sued.

180

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

The way OpenAI bridges this is by saying the protections live in this "deployment architecture and safety stack" rather than the contract language. But if this contract says "all lawful purposes" and your safety stack prevents a lawful purpose, you're in breach of contract.

1100

Peter Wildeford @peterwildeford.bsky.social · 01/03/2026

OpenAI is trying to claim simultaneously that (a) their contract with the Pentagon allows for "all lawful purposes" and (b) also that their red lines are fully protected.

1202

Peter Wildeford @peterwildeford.bsky.social · 27/02/2026

The Pentagon has a legitimate principle that private companies shouldn't hold moral vetoes over military doctrine. But they agreed to the contract. And now they're using unprecedented + disproportionate coercion. This should trouble everyone. My latest - peterwildeford.substack.com/p/the-pentag...

peterwildeford.substack.com

The Pentagon's War on Anthropic

The Pentagon has a legitimate principle, and a terrible strategy for enforcing it

0101

Peter Wildeford @peterwildeford.bsky.social · 26/02/2026

Are there Cold War lessons to learn for AI? We've had very fierce competition with the Soviets, and did not trust the Soviets at all, but we were still able to make mutually verified treaties. In Politico today I'm quoted saying we should do the same with China → www.politico.com/newsletters/...

politico.com

Cold War lessons for the AI era

050

Peter Wildeford @peterwildeford.bsky.social · 25/02/2026

Adversaries can tamper with or poison leading US models. There also can be risks from insider threats, including potentially the AIs themselves. Dave Banerjee at IAPS has a roadmap for how to defend -> www.iaps.ai/research/ai-...

030

Peter Wildeford @peterwildeford.bsky.social · 23/02/2026

yeah I agree - that's a good point Maybe you'd see major progress on CAIS's Remote Labor Index or OpenAI's "OpenAI Proof Q&A"?

010

Peter Wildeford @peterwildeford.bsky.social · 23/02/2026

agree - probably measurement noise (in both estimates)

010

Peter Wildeford @peterwildeford.bsky.social · 23/02/2026

16. Joseph Sifakis (Turing Award '07) 17. John C. Mather (Physics '06) 18. Frank Wilczek (Physics '04) 19. Joseph Stiglitz (Economics '01) 20. Andrew Yao (Turing Award '00) *The Turing Award is equivalent to the CS Nobel

131

Peter Wildeford @peterwildeford.bsky.social · 23/02/2026

7. Giorgio Parisi (Physics '21) 8. Jennifer Doudna (Chemistry '20) 9. Yoshua Bengio (Turing Award '18*) 10. Beatrice Fihn (Peace '17) 11. Oliver Hart (Economics '16) 12. Juan Manuel Santos (Peace '16) 13. Ahmet Üzümcü (Peace '13) 14. Jean Jouzel (Peace '07) 15. Riccardo Valentini (Peace '07)

131

Peter Wildeford @peterwildeford.bsky.social · 23/02/2026

20 Nobel Prize winners have warned that we may someday lose human control over advanced AI systems 1. Geoffrey Hinton (Physics '24) 2. John Hopfield (Physics '24) 3. Demis Hassabis (Chemistry '24) 4. Daron Acemoglu (Economics '24) 5. Ben Bernanke (Economics '22) 6. Maria Ressa (Peace '21)

192

Peter Wildeford @peterwildeford.bsky.social · 20/02/2026

The infamous METR graph is going vertical. Current trends suggested ~8h-9h time horizons but instead we're seeing ~14.5h time horizons! Based on this, I would project ~2-3.5 workweek time horizons by end of year (!!). That could have significant implications for the economy.

4413

Peter Wildeford @peterwildeford.bsky.social · 18/02/2026

🔴Rep Kiley (CA) 🔵Rep Moulton (MA) 🔴Rep Mace (SC) 🔵Rep Sherman (CA) 🔴Rep Moran (TX) 🔵Rep Tokuda (HI) 🔴Rep Paulina Luna (FL) 🔵Rep Whitesides (CA) 🔴Rep Perry (PA)

040

Peter Wildeford @peterwildeford.bsky.social · 18/02/2026

🔵Sen Schumer (NY) 🔴Rep Burleson (MO) 🔵Rep Beyer (VA) 🔴Rep Crane (AZ) 🔵Rep Foster (IL) 🔴Rep Dunn (FL) 🔵Rep Krishnamoorthi (IL) 🔴Rep Johnson (SD) 🔵Rep Lieu (CA) (continued)

140

Peter Wildeford @peterwildeford.bsky.social · 18/02/2026

27 current members of Congress have publicly discussed AGI, superintelligence, AI loss of control, or the Singularity: 🔴Sen Blackburn (TN) 🔵Sen Blumenthal (CT) 🔴Sen Hawley (MO) 🔵Sen Hickenlooper (CO) 🔴Sen Lee (UT) 🔵Sen Murphy (CT) 🔴Sen Lummis (WY) 🔵Sen Sanders (VT) 🔴Rep Biggs (AZ) (continued)

1120

Peter Wildeford @peterwildeford.bsky.social · 06/02/2026

And in order to do that, we will need to build the verification technology to verify that deal. And that technology will need to be built now so that it is ready in time for a deal. This will build us important optionality for the future. We need a second button, if only to have another option.

010

Peter Wildeford @peterwildeford.bsky.social · 06/02/2026

And in that case, we may want to make a deal with China to mutually slow down the race to superintelligence so we can proceed with more foresight.

110

Peter Wildeford @peterwildeford.bsky.social · 06/02/2026

And we must beat China commercially as well. But another way to answer this question is that this second race to superintelligence, as distinct from the commercial race, may be a race to see who loses control first.

110

Peter Wildeford @peterwildeford.bsky.social · 06/02/2026

People say "well we have to accelerate because what about China?" This is used too much as an excuse and people who say this need to be more hawkish over export controls. But it's a reasonable question - I don't want China building AI superintelligence.

110

Peter Wildeford @peterwildeford.bsky.social · 06/02/2026

Later on, seismometers were invented, a key verification technology that enabled underground tests to be verified. A deal between the US and USSR followed after.

111