That's what
happens after
US companies
force a take over.

Now that we have cheap AI code completion, maybe we should flood the project with patches to steer the code base back towards Unix?

If only he had lived in the US, he could have appealed to the Supreme Court!

You're hallucinating

I disagree. It's not up to a college to get you a degree. If the college deems you to have failed to pass the standard, you should get NO degree.

Why would they have a patch? Sounds like a systemic problem. Maybe think out of the box and add human drivers to supplement the algos? Passengers could take over driving and get a discount on the fare.

Why did you not list their lack of profits? Nobody cares about revenue growth, it looks impressive but is made up nonsense full of circle jerk deals among AI companies. Investors care about ROI and don't like capex unless there's a big payday coming very soon.

Here's a better number: Anthropic announced their first ever operating profit just ten days ago for Q2. Very suspicious timing given the IPO. And interestingly the profit claim happens to coincide with a compute discount for said Q2 from Elon. No discount for Q3+Q4. Elon has to make his profit claims.

Like I said: Circle jerk all around.

I stand by what I said. The *hype* IS dying. Investors are getting impatient.

double rot13 is the way to go!

Google and Anthropic are perfectly able to build their own compute-centers, but it may not be economically a good idea at this point. Remember, the AI hype is dying and the investors are starting to smell a rat. Leasing GPUs from Elon means Google and Anthropic are not stuck with tons of hardware they can't monetize in the future. If things go bad, they just stop paying for access and Elon is left holding the bag.

It's important to distinguish the model from the generated sequence dynamics. The "model" (in the traditional scientific sense) is a fixed function of the input data. It cannot improve itself, except when the scientist replaces it. But that's not improvement, that's revision.

Most regular people don't understand the word model, and think it refers to their interaction with the tool. That can indeed evolve, in the same way that if you repeatedly run an input output loop with a given transformation, and add some random perturbations (aka user control), you will get a trajectory that evolves. It will probably converge to some fixed point, but that could take a long time to detect by users, who think it's evolving before their eyes.

Ask yourself what AI models actually do? They *mimick* what humans have been doing themselves. Things like writing sentences, drawing pictures, collaging photographs, collaging movies. But without ever having built-in attribution mechanisms. So it looks novel but never is. It's just a compression model of real life data artefacts, generating short random samples.

And now ask yourself what the obvious killer app is for a human output mimicking system? A system that allows the owner to pretend some human did something. There was never any other plausible outcome.

Models don't come up with anything useful on their own. That's why they invented RAG in the first place.

This is by design. Remember, Google has a deal to read and train on Reddit. It makes their AI slop seem legitimate.

Agreed. But Marvel / MCU also shows that "Stargate Syndrome" is all too easy to catch. The heroes and villains just keep getting more powerful and dangerous, no wonder the movies are flopping.