Industry that built the problem offers to sell you the solution

OpenAI has gathered more than 100 of the world's biggest tech and infosec companies to warn that cyber defense is in trouble - a reassuring development given quite a few of them helped build the technology involved. The open letter has more than 100 names attached to it, including many of the companies with the most to gain – or …

[may_i]

Mafia technique

Nice data centre you have there, it would be a shame if anything happened to it.

Sign here to join our protection plan.

[ecofeco]

Re: Mafia technique

The USA operates on the bust out business process these days and has for decades.

It was shady in the best of times throughout its entire history. Now it's outright criminal.

[Jedit]

"... the problem offers to sell you the solution"

And there we have it. AI is not a solution looking for a problem. It is the problem, and the simplest and cheapest solution is to not use it.

[Sp1z]

Re: "... the problem offers to sell you the solution"

Except “not using it” in this instance doesn’t help you, because the other people using it are still going to break your network with.

So actually (and annoyingly) you kind of have to use it to test your own systems before someone nefarious does.

And the AI companies are fully aware of this.

[DoctorNine]

Re: "... the problem offers to sell you the solution"

I think that it's possible to design a system which requires an actual human in the interface with the outside world, and that would fix the AI problem. People just haven't been creative enough to figure it out yet. I suspect that's coming. AI to fight AI is a fool's errand. That just ends up being an arms race. Being human is different from being AI. Distinguishing them is doable, if a bit time consuming.

[Kurgan]

Re: "... the problem offers to sell you the solution"

The solution would have been not to invent it. But it's too late. So unless someone has a time machine and some bombs...

[cyberdemon]

Re: "... the problem offers to sell you the solution"

The time machine is not necessary.

[Anonymous Coward]

Re: "... the problem offers to sell you the solution"

Sounds like a cross between The Human Centipede and an either Ouroboros or a pegging circle.

[Anonymous Coward]

Lets see.

Companies spend their money on AI 'just because'. They don't see an ROI. They lose their technical talent. Money they SHOULD have spent on replacing their ancient systems that are susceptible to attack is instead given to the companies that exasperated the problem in the first place.

I don't know of any company where they've used AI to test for vulnerabilities and use that data to plan upgrades - they've only used it as a pretense to get rid of people.

[retiredFool]

We could pay

with a tax on AI. Double win, could deflate the bubble a bit since that cuts into profit and they'd get to pay for the mess they created.

With the money, we could hire people(maybe the ones being laid off now) to actually look after the control systems in person and disconnect them from the cesspool created by AI.

[Alan Mackenzie]

Re: We could pay

"AI" doesn't make any profit. It's sold at a small fraction of its cost, leading to massive losses, particularly at OpenAI and Anthropic.

See Ed Zitron at <https://www.wheresyoured.at>.

[retiredFool]

Re: We could pay

Good point, make it a sales tax on their revenue.

[katrinab]

Re: We could pay

Make it a tax on electricity consumption

[retiredFool]

Re: We could pay

Now that is a good idea, have an upvote.

[cyberdemon]

Re: We could pay

Now now, we don't want to discourage bri'ish biznis and be anti-growth and all that! So have a tax on AI's carbon emissions, no matter where in the world it is emitted.

i.e. if a British user calls an AI query that causes 1kg of CO2 to be sent up a smoke stack in Arizona, a tax should be due from the AI company (who may in turn pass it on to the user)

If they want to avoid that tax, then they ruddy well better build solar-powered datacentres (in locations that are in sunshine when the demand peaks occur) !

(of course they would never do that, because even in somewhere like Morocco, sunshine is less than half of the day, so they would need to spend twice as much on datacentres, plus the solar arrays)

[katrinab]

Re: We could pay

Using solar to charge batteries and then discharging the batteries when the sun isn't shining is perfectly OK.

[Anonymous Coward]

Re: We could pay

The big 7 created the problem; they have a combined net worth of around 23 trillion dollars while the U.S. government sits at a negative 25 trillion. I think they should take the financial hit, not U.S. taxpayers.

[druck]

Re: We could pay

Trouble is the AI companies trillions is only paper value based on hype and disappears instantly the bubble busts, where a the US's debt is very much real and is only getting larger

[xyz123]

Thats a LOVELY bit of code you have there.

Be a SHAME if something 'happened' to it

[ErikOnTech]

This is the bill for decades of IT slop coming due.

None of these hacks are impressive on any technical level. It’s organizations embracing the most pathetic slop imaginable and calling it a system. Does anybody really think that Hugging Face is anything other than some slopcoded monstrosity? The TeamPCP exploits, while amazing fun to watch, started with GitHub running shell scripts containing unsanitized variables, in 2026. What could possibly go wrong?

Silicon Valley rose on the back of “move fast and break things,” made a mountain of money, and built us with a world full of broken shit. And then the overwhelming majority of IT departments looked at this mess and said “hold our beers.”

[Anonymous Coward]

Now forgive me for pointing out the f'ing obvious but a lot of these systems should never have been put online in the first place. The only reason some were was to save money.

The ones that had no choice but to go online should have spent the money to make them secure or made access as small as possible and limited what could be done from online.

This is the problem with capitalism. It's all about profits. Can we do this to make more money by cutting workers? Sure but not one asks the question of consequence. Same goes for can we build this on the cheap? Do we really need to employ security experts? Nah, it'll be fine.

None of this has anything to do with AI though. All AI is doing is crawling known exploits and trying them elsewhere. Eventually they would have been exploited anyway. This is just accelerating that and the AI companies will happily sell you the solution but it won't be a proper solution. It will need to be maintained and ongoing. Just like pharmaceutical companies don't want cures, they want ongoing treatments. Why would AI companies be any different? Every new novel exploit found tested against your systems by our new AI infosec products. What a time to be alive.

[jiho]

This is the problem with capitalism. It's all about profits.

And not just profits per se. If it was only about making a living, it might be good. But stock value depends on *increasing* profits over time, which sooner or later will leave any stockholder owned corporation in an impossible position.

But the problem around the problem is, no one can point to a better system. "Solutions" generally involve some small number of people with arbitrary power to impose decisions, which for obvious reasons never works out very well. I mean, that is what we are looking at here, just different people.

[Rich 2]

Dr Strangelove would be proud

“Their proposed solution is, inevitably, more AI”

FFS. We deserve to go extinct

[Roland6]

Re: Dr Strangelove would be proud

It’s quite amazing how Hollywood predicted this line of thinking and reasoning in 1984 when in the Terminator film, Skynet was “released” to counter a computer attack…

[jiho]

1984

Until I saw what followed, I thought you were going to reference the Orwell novel.

[ecofeco]

The perfect headline

Never thought I would see the day.

[Rich 11]

Handy title

Industry that built the problem offers to sell you the solution

Just like religion, then.

[Roland6]

Re: Handy title

Trouble is, in true US grifter style, the “solution” will turn out to be snake oil…

[jiho]

Ah, Cloudflare

I run Firefox on two platforms, Windows 11 and NetBSD 10.1. Privacy & Security settings are identical on both. On NetBSD the User Agent mimics the one on Windows 11.

You know that little Cloudflare "Verify you are human" checkbox that pops up when you try to access websites. On Windows this works fine. On NetBSD it kicks itself around and around in an endless loop, asking me to check the box over and over, getting nowhere.

With identical Privacy & Security settings, and the same User Agent. This implies to me that Cloudflare must have some kind of *backdoor* into Windows itself.

[jiho]

Re: Ah, Cloudflare

I should add one important detail I omitted: Worked just fine on NetBSD, until the recent Windows Update. Has not worked on NetBSD since.

[jiho]

Re: Ah, Cloudflare

Now I have even more reason to suspect a backdoor of some kind. I tried Firefox on OpenBSD -- same version and settings as NetBSD, except the User Agent was still for Linux (default on OpenBSD). Lo and behold, the Cloudflare checkbox worked without issue and I had access to sites. Then I changed the User Agent to Windows, and suddenly the checkbox did not work at all, just like on NetBSD. This suggests to me that when Cloudflare sees the Windows User Agent it does indeed try to use some kind of backdoor into Windows, which is of course not there on the BSDs, while with the Linux User Agent it does not look for it and everything works fine. I imagine telling the truth about the OS might work just as well.

The problem is that UPS tracking on the UPS website only works with the Windows user agent! And of course I do need to use that tracking sometimes.

Tell me you knew ....

[Trank1234]

Just as a gun has always seemed to be the solution to guns, AI is going to end up being the solution to AI. It's human nature.

[Bebu sa Ware]

Oops… Alert the moderatrix, the coders need a whipping.

First time I have seen this pop-up (all in red too.)

Been popping every few minutes.

A new service offered by the Vulture for the coding S&M community ?

With the deprecation over the last few decades, of gendered terms relating to roles seeing a moderatrix seems positively kinky (she attired in a a tight red morocco leather outfit? Who doesn't love the fragrance of a well bound tome?)