They are remarkably useful. Of course there are dangers relating to how they are used, but sticking your head in the sand and pretending they are useless accomplishes nothing.

It models only use of language

This phrase, so casually deployed, is doing some seriously heavy lifting. Lanuage is by no means a trivial thing for a computer to meaningfully interpret, and the fact that LLMs do it so well is way more impressive than a casual observer might think.

If you look at earlier procedural attempts to interpret language programmatically, you will see that time and again, the developers get stopped in their tracks because in order to understand a sentence, you need to understand the universe - or at the least a particular corner of it. For example, given the sentence “The stolen painting was found by a tree”, you need to know what a tree is in order to interpret this correctly.

You can’t really use language *unless* you have a model of the universe.

they, in fact, will have some understanding

These models have spontaneously acquired a concept of things like perspective, scale and lighting, which you can argue is already an understanding of 3D space.

What they do not have (and IMO won’t ever have) is consciousness. The fact we have created machines that have understanding of the universe without consciousness is very interesting to me. It’s very illuminating on the subject of what consciousness is, by providing a new example of what it is not.

They absolutely do contain a model of the universe which their answers must conform to. When an LLM hallucinates, it is creating a new answer which fits its internal model.

privacy on that site was horrible, and I stoped de-selecting vendors who want permission to track me after two minutes.

Just open the page in a private window at that point, and click the “yeah sure track everything bro” button.

I cannot wait until architecture-agnostic ML libraries are dominant and I can kiss CUDA goodbye for good

I really hope this happens. After being on Nvidia for over a decade (960 for 5 years and similar midrange cards before that), I finally went AMD at the end of last year. Then of course AI burst onto the scene this year, and I’ve not yet managed to get stable diffusion running to the point it’s made me wonder if I might have made a bad choice.

Same. I had an Nvidia 960 for about 5 years on arch with very few problems. Maybe twice over that time I had to rollback to an older version temporarily due to some incompatibility with wine or such like.

Towards the end of last year I finally decided to upgrade (mostly to play RDR2) and I went with AMD. I love the feel of using a pure open source gfx stack, but there is no real functional advantage to it.