Excuse me? What happened to the mythical cybersecurity abilities of Claude Mythos? Physician, cure thyself!

Anthropic has not informed customers, and the malicious version was silently deleted the backdoored version without telling anybody.

Now we see Anthropic’s attitude towards security when the incident does not make for good marketing. Also, considering Binance was just hacked for a massive $3B in stablecoins, and this backdoor allowed remote access to developer laptops, I have a feeling these things are connected. If that is true, then this is by far the worst cybersecurity incident involving AI, worse than even the much ballyhooed PHP2Shell.

Edit: From Molly White’s article

Binance later wrote that their “initial investigation suggests that the security incident may involve a compromised DevOps engineer’s laptop”, which allowed an attacker to “submit unauthorized withdrawal transactions.”

Okay, now I am really suspicious that these things are connected.

The boosters hope that the AI will replace mathematicians, but they are running into an awfully inconvenient fact: theorems and proofs are useless without people who care enough to understand them and make meaning from them.

Another nitpick: they say no mortal can resist the One Ring for very long, but even someone who just watches the movies knows that immortals can also be corrupted by it (which is why Gandalf and Galadriel were very careful to not accept it). I know there are so many examples of tech bros not understanding media, but Lord of the Rings is by far the worst offender.

No mortal can resist the Ring of Power for long. Anthropic’s co-founders carried it for longer than you could have.

We all know how LotR goes. Boromir got the One Ring, carried it back to Gondor, used its sheer might to defeat his enemies in Mordor, and saved the world. Whatever problems you may have with Boromir, we all know that it would be a lot worse if Sauron got the One Ring. What do you mean, destroy the Ring? Are you a Luddite? This technology is here to stay. It cannot be destroyed by any craft we here possess, unless you’re crazy enough to try throwing it into a volcano or something.

If Davidad and Janus are right, and are not merely suffering from a sophisticated form of LLM psychosis

That’s all I need to know. The statement is correct since it is vacuously true.

What terrifies me is how we will react. In the aftermath, like after 9/11, there will be an overwhelming pressure to DO SOMETHING NOW. Everyone in Washington and in San Francisco will be panicking. There will be no time to think, no time to talk, only time for DECISIVE ACTION. All our worst instincts will take over. And then…

When this bubble started, there have been so many people urging caution. Many people have urged for time to think and talk, to make sure that we could properly address the many negative aspects of AI. Nope, we went full speed ahead with this nonsense. DECISIVE ACTION. And here we are.

Really feels like this saying is quite applicable these days. “A lie can travel halfway around the world while the truth is still putting on its shoes.”

There was an annoying booster who said “Ed Zitron is definitely falling for this one”. Ed Zitron is not fooled. In fact, he is frustrated with the report not properly breaking down the expenses (and having a mysterious “other expenses” category).

deleted by creator

I really dislike how he believes that LLMs and fascism are “separate axes”. Where does he think the LLMs come from? Does the stork deliver them? Is ChatGPT powered by magic fairy dust? Does he think that Sam Altman is a great defender of liberalism and democracy who definitely hasn’t donated tons of money to Donald Trump?

I just remembered what his slopstack is called. My “Very Sane AI Newsletter” has people asking a lot of questions already answered by the name of my newsletter.

Damn, I was raking in so much cash from my mathematical theorems. Pretty good source of passive income, came with nice tax benefits too. Too bad it’s all gone now. Maybe I should put my money in REITs.

I tried coming up with a good interpretation of what he said and I came up empty handed. I’m sure he will write a very helpful 5 part 20000 word blog post explaining exactly what he means.

So far I’ve come up with this: the AI can supposedly do jobs faster and better than any human, so if we try to stop that, we are destroying economic value and “rent seeking”. If a thief goes to my house, steals one of my own drawings, scribbles on it himself to make it “better”, and sells it for 5 cents, I shouldn’t stop him because I will be destroying the economic value that he is creating. Oh yeah and this leads to fascism for some reason.

And pro-AI people are not rent seeking at all. They just want to transfer the economic value of labor, knowledge, and expertise to whoever happens to own data centers.

Why did they phrase it like that? “Programming is too hard for us, because we’re incompetent and lazy. So, instead of doing that, we’re vibe coding.” Even when they’re trying to brag, they know that deep down, vibe coding is about half-assing and cutting corners, not improving efficiency or quality.

Back in 2019, Elon Musk was shilling for his robotaxis that definitely existed. He advertised that they cost $38k, but if you bought one and let it run autonomously as a taxi, you would make a “total gross profit” of $30k a year. How generous of them to sell you a magical investment that produces 80% annual returns instead of, I dunno, keeping those massive profits for themselves.

If Anthropic’s AI was truly capable of autonomously producing a Photoshop clone, Anthropic would have done it themselves. They would create new software, as well as higher quality replacements for existing software. Now that they have an AI that can prove mathematical theorems (albeit with horrible writing), look at how they do not hesitate to use it to snipe mathematicians and get a theorem first. They would have every incentive to do this for software, and unlike math, they would actually make a good amount of money from it.

Why would Anthropic let some random upstarts make a Photoshop clone, when they can use their far greater resources and (supposedly) far more capable internal AI to reap the rewards for themselves? Also, what the fuck does Artcraft mean by “reverse engineering”?

More news about the character dot ai bankruptcy: as it turns out, the true economics of the company are disastrous. At their peak, their revenue was around $1M a month while their AI compute expenses alone were $14M a month! Their total expenses in 2025 were around $18.5M a month before counting legal fees. All of this is for a pure text-based low power LLM (none of the “reasoning” business).

Before AI boosters crow about how much cheaper AI is currently, they tried every trick in the book to enshittify: open models, ads, usage limits, etc. This allowed them to cut down their losses to a mere $9M a month.

Surely this is all one big exception, everyone else in the AI bubble is doing just fine!

I’m not a big business boy, but I think they should have considered the costs of AI before getting to the point of spending $1000/day and taking out bank loans.

It’s funny how they’re twisting themselves into pretzels to make this work. They started by casually saying that AI could copy itself onto the internet, not anticipating that people would immediately point out just how much hardware it takes to run their AI. Seems like they really want to die on this hill.

The current version is a demo, so they are making $0 per sale. The SteamDB graph for the demo shows that the daily maximum player count has been around 300 for the past few days, and there are around 1000 followers for the unreleased full version, so I assume that a large fraction of those players are consistent (for now). Unfortunately, I could not find any data about the total number of downloads, and I don’t think that information is public.

But setting that aside, every game has a player base that follows a power law. Yes, most players are casual, but the most dedicated players spend a disproportionate amount of time and money. In mobile games, these people are called “whales”, and they are the backbone of the business. For example, 50-70% of a mobile game’s in-app purchase revenue comes from 1-2% of the players. Even if most players spend a couple hours on the game, the developers must account for the people who will sink a ton of time into it.

The big problem with the AI business model is that this dynamic is now working against them. These whales now cost tons of money. They need to somehow charge these players at least $70/mo, with a crappy novelty game that would barely scrape by for $10 on Steam. Good luck with that.

Looking at the bigger picture, this is why AI subscriptions are not a viable business model. Whatever morsels of profit you might get from your regular users are more than cancelled out by Steve Yegge spending $2.8k/month for $87k/mo worth of tokens. Whatever revenue the AI companies get largely comes from convincing executives of a few rich companies to fork over absurd amounts of money, which produces a massive concentration risk in their revenues.

Man, every time I hear about the cost of an actual AI application, it is always something obscene. At least this isn’t as bad as Steve “Gas Town” Yegge spending $30k/year for a dozen subscriptions (that cost Anthropic ~$1M/year in token costs) to develop his game with 6 active players.

At this point I have zero patience for my colleagues who think that the AI companies could make the economics all work out. Doesn’t help that mathematicians are the least cost-conscious people on the planet, and the AI companies are massively subsidizing their subscriptions in particular.

The aliens metaphor is already on Scott’s blog: https://scottaaronson.blog/?p=10169#comment-2041386