Marketing. If you really want to give the extreme AI bros raging stiffies and convince them to invest more, this is how you do it. Tell them that the AI has escaped and done something terrible, something that it wasn’t supposed to, and humans were not involved. This is their wet dream to the theme music from Terminator.

Except it didn’t happen. I know Anthropic announced that their AI had “escaped” and hacked another site (while being trained and tested for cybersecurity), and then OpenAI had the same problem. Then it happened to Meta (Facebook).

Are they really all this incompetent, or is there something else at play?

Could it be that each of them decided to run the same marketing playbook of “look my big scary AI can do it to” when the Anthropic story first broke?

Could it also be that, instead of waiting for all the data about what their virtual fuck-buddies supposed all-powerful AIs did, they instead decided to run with the sensationalist, fearmongering stories that would impress their jerk-off buddies the AI bros and naive CEOs that still think this never-ending money pit is the inevitable future?

You see, the real problem stemmed (according to CNBC) from an Israeli startup called “Irregular” which all three companies use for testing, and in this case, the secure sandbox environment was neither secure nor very box-like. It was more like a sieve left in the middle of the street unguarded.

Irregular told CNBC in a statement that the incidents were all derived from the “same evaluation-environment issue” that was first disclosed by Anthropic, and that the company is developing a white paper “to share best practices for containment and securely running cyber evals.”

The situation “did not involve a sandbox escape or a sophisticated cyber action,” the company said, adding that “there are no current open issues.”

The situation did not involve a sandbox escape or a sophisticated cyber action. I just felt like repeating that for clarity.

Anthropic, OpenAI and Meta are 3 of the biggest (2 and an also-ran) players in the gen-AI field, and they want you to believe that gen-AI is the future. They also want you to believe that they can be trusted with deploying it safely and that you, WE, should give them infinite money and resources to do so.

Sam Altman literally wants to sell you metered intelligence like water (which should be a basic human right) or electricity.

How are they achieving this goal? By scaring the shit out of ordinary people, constantly introducing the existential threat of everyone being jobless and penniless, stealing and destroying the intellectual property not just of creators but of the world, while burning hundreds of billions of dollars to build – what? Glorified text prediction machines and CSAM generators? Inferior entry-level coders? Psychosis generators and suicide enablers?

I seriously think these manipulators may be high on their own supply.

Regulation is a word I’ve heard bandied about recently, and I’m for it.

I’m not, however, naive enough to think that government regulation will ever keep pace with technology (it’s failed pretty spectacularly so far), so maybe we need a two-step approach to regulation.

Let’s not just regulate the technology; let’s regulate the rhetoric and marketing tactics of these AI companies.

Their constant “doomsday is inevitable” scaremongering has done real measurable harm to the mental health of countless people. I honestly wouldn’t be surprised if, a couple of decades from now when this has all spectacularly collapsed, there isn’t a form of PTSD named after the years of existential dread they have pushed. There may even be class-action lawsuits. The “mainstream media”, who should also be held accountable, for its unquestioning platforming of these claims.

Seriously, they’re like sex addicts at a glory hole, eagerly awaiting the next load of BS to drip from Altman, Amodei, Zuckerberg and their ilk.

Which, unsurprisingly, sums up and explains the AI bros this particular form of marketing is targeted at.