The size of improvements over time is diminishing. We’re not in “big bang” territory anymore and we’re about two years into the “incremental refinement” period. We’re about to enter the next AI Winter unless somebody comes up with a new architectural component as revolutionary as transformers have been for ML models.

The models are also getting extremely big as well, since the big improvement currently seems to largely be stuffing the model with more parameters, and making that work.

I can only imagine that the training cost has also been skyrocketing.

The core problem is that LLMs do not create. Full stop. All creativity is borne by the human inputs. Until that changes - until the model gains the capability to truly create new information, we’ve hit the limits in raw capability.

The models suppress outliers by design. Statistically speaking, the most novel thing is a garbled mess of random words, but random noise is useless, so it ends up being suppressed. You can see by fiddling the samplers, or increasing the temperature.

The Library of Babel is the most creative thing in the world, containing every possible combination of English words and letters. You can basically act like an LLM by trying to find a new coherent sentence in it, but also one that hasn’t been said before. It’s basically impossible.

But that is what an improvement is supposed to be. Compare that to finding a sentence that has been said, or something close to it.

Most gyms already have a “no photographs/videographs” sign, don’t they?

Both could be true at once.

People are also allowed to have contradicting opinions. People aren’t perfectly ordered logical reasoning machines.

MetaMuseAl

It could also just be immaturity. They’ve not yet figured out how to draw the boundary between fantasy and reality, and separate their idealised character from the VA. Or what things may be better kept to themselves.

There’s a game called War frame, maybe they’re talking about setting up their controls?

Like how you used to be able to get macros for fishing in runescape.

Copycats might also be an issue. I remember that when Douglas Wreden published his book, he had to deal with numerous AI-generated copycats popping up one after the other, using his name and face, trying to trick people into buying their book, because his had become popular.

They usually do. They just do the same thing a lot of nuclear reactors do and also evaporatively cool one end of the heat exchange loop.

I would be curious how that will work with anti-trust legislation. At least currently, there’s an argument that Sony does not prevent people from buying games from other sources, or other publishers from publishing physical games for their consoles.

That’s no longer the case if they switch to a purely digital model, since they become the sole source of all the games on PlayStation, and have unilateral authority over them. The only place you can get a game on a digital model is Sony.

Such a unit exists and it is also called tokens, that can measure the capability of a model and the size of a running operation in a model.

I think you might have it mixed up with parameters, rather than tokens. Parameters are how big the model is, and are an indirect measure of how capable it is. Bigger models tend to be more capable.

But what they use for calculating your bill is something different today.

The tokenizer varies a little, but I don’t think it’s changed measurably from tokens. You pay an amount for a million tokens worth of processing. The tokeniser difference just alters how text is converted to tokens, but the tokens themselves don’t change all that much.

If anything, I’d honestly put the issue more with reasoning chains in models, where they basically babble to themselves inside of a <think> tag, that most interfaces hide/collapse. It makes them work better, but vastly increases the amount of tokens per operation.

They have been getting longer and more sophisticated with newer models. So you might have a model now that basically repeats the output multiple times whilst refining and drafting the non-reasoning output.

If you’re making it generate a lot, that’ll balloon the usage, and thus price.

We had hardware piracy back in the day, when there were CPUs you may have been able to unlock extra cores on.

Then the companies started lasering them off so that was no longer possible.

I’d be surprised if it was. The man was a symptom, not the root cause of the problem, and he certainly didn’t do it all by himself.

He was the stereotypical man in the underworld who could get you anything, as long as you could afford the price.

Others would almost certainly have filled in the gap he left by now.

They do, but there may be better local alternatives, since the shipping can be quite high.

It seems more like a bad editor/insufficient proofreading.

That’s the kind of comment I sometimes make if I rephrase it in the draft, but don’t check to make sure I remove all of the original sentence before posting.

The case recently resolved in the plaintiff’s favour. though Google intends to appeal, so it’s up in the air how things go.

And if it’s this easy to poison the AIs, imagine how easy it is for someone with an actual agenda to mislead people in ways that aren’t as fantastical and quickly spotted.

Equally concerning is that these systems are now seeing use in a range of things. There are lawyers who use it to file suits when they shouldn’t be, and a US lawmaker was recently found to be using AI to draft laws. What happens when things like that make it into the models training data, rather than just being pulled in by RAG/web tools? They’d become part of the base knowledge of all the models of that line going forward.

It’s funny when it’s outlandish. The question becomes what happens when it isn’t? Even without an agenda, what happens when it cites an outdated/incorrect source, or assumes that someone making a joke was correct, and ends up drawing from that when filling a lawsuit/drafting a law?

I don’t think people would have minded very much if they had taken one of their existing chassis and swapped it for an electric drivetrain, but kept the design mostly the same.

This thing looks fine as a standard passenger EV

I’d honestly disagree there. It looks like someone wearing oversized clothes, which isn’t helped by its colour scheme.

It stands out in a bad way, compared to a modern passenger EV, which arguably looks nicer because it’s less awkward.

But ask a mainstream manufacturer to make an EV car and they look stupid half the time.

IMO, it’s probably more that they overdesign them. They want the EV to look futuristic and unique compared to their regular cars, but their cars already look like that to some degree, and so they overcook the design into looking like some science fiction vehicle. Take the Ferrari, for example, they tried to make it have a floating arch where the hood would normally be.

That’s fine for a movie or video game, but in real life, coupled with the practical limits, it just doesn’t look very good.

Poorly optimised is also extra bad in today’s climate, where the cost of hardware is just through the roof. People simply can’t afford a new, more powerful computer.

Lasers would be quite risky. Anything that powerful would also instantly obliterate peoples’ vision if there’s a reflection and it gets them in the eye.