If AIs are to find the solution for us, we need one really smart one, not many AIs that are similarly smart to existing ones. He is proposing building more data centres, ie. the latter option.

If we can spot these trends while working 9-5, then an idiot can probably spot them if they spend 40 hours a week on it.

PieDock

That is the error that the model made. Your quote talks about the causes of these errors. I asked what caused the model to make this error.

Sure, but which of these factors do you think were relevant to the case in the article? The AI seems to have had a large corpus of documents relating to the reporter. Those articles presumably stated clearly that he was the reporter and not the defendant. We are left with “incorrect assumptions made by the model”. What kind of assumption would that be?

In fact, all of the results are hallucinations. It’s just that some of them happen to be good answers and others are not. Instead of labelling the bad answers as hallucinations, we should be labelling the good ones as confirmation bias.

The email says that you can do it. It doesn’t say that you can do it without purchasing the upsell option.

AI developers: your copyrighted work is such a small contributor to the AI’s output that copyright doesn’t apply. Also AI developers: but our AI won’t work without it.

We need to be transitioning to zero carbon as fast as possible, period, and even that isn’t good enough. Moderating our energy consumption is vital. There is a cliff at the end of the road and business as usual means driving on down the road.

I am not saying that we need to turn off our lights and heating. I am saying that we first-worlders use a lot of power on frivolous things that we absolutely can live without.

Your ICE has a significantly longer range, and the road network has evolved so that you can be reasonably confident that you’ll find a filling station when you need one.

Today I’m driving an EV that doesn’t have it, and I’m missing it. Different EVs have different ranges and not every filling station on the autobahn has chargers. On the other hand, there are lots of places just off the autobahn which do have chargers. It’s a different game. Your mileage may vary of course.

The Megane E-tech has functionality in its satnav that lets you plot a route with charging stations on the way, showing how much capacity you will have left when you get to them. Not essential, but very useful for somebody who is new to EVs.

Software that communicates with power companies to allow the car to charge overnight at advantageous rates, or even feed energy back into the grid. Again, not essential, but good for the customer and helps with the transition to green electricity.

As far as I can tell, Microsoft tried to hold off these anti-trust lawsuits by intentionally making the interoperability and feature-parity between its products shockingly bad.

If you make a painting now, it wouldn’t be based on those thousands and thousands of paintings since, although you have seen them, you apparently do not remember them. But, if you did, and you made a painting based on one, and did not acknowledge it, you would indeed be a bad artist.

The bad part about using the art of the past is not copying. The problem is plagiarism.

Inspiration is absolutely a thing. When Constable and Cezanne sat at their easels, a large part of their inspiration was Nature. When Picasso invented Cubism, he was reacting to tradition, not following it. There are also artists like Alfred Wallis, who are very unconnected to tradition.

I think your final sentence is actually trying to say that we have advances in tools, not inspiration, since the Lascaux caves are easily on a par with the Sistine Chapel if you allow for the technology? And that AI is simply a new tool? That may be, but does the artist using this new tool control which images it was trained on? Do they even know? Can they even know?

Maybe the AIs should mix their own pigments as well, instead of taking all the other artists’ work and grinding that up.

I had forgotten how the real OSD starts. That is horrifying. But then, I suppose if you’re going to engage in this kind of disinformation, why not go all-in?

It’s good of the author to extend Futo the benefit of the doubt in this way.

The very first paragraph of their definition is: “Open source just means access to the source code.”. If they are really that unfamiliar with the software industry, then their code must be a horror show. Personally, I think they know exactly what they are doing.

I wish Altman would read Accelerando.

But we do know how they operate. I saw a post a while back where somebody asked the LLM how it was calculating (incorrectly) the date of Easter. It answered with the formula for the date of Easter. The only problem is that that was a lie. It doesn’t calculate. You or I can perform long multiplication if asked to, but the LLM can’t (ironically, since the hardware it runs on is far better at multiplication than we are).

This seems to be a really long way of saying that you agree that current LLMs hallucinate all the time.

I’m not sure that the ability to change in response to new data would necessarily be enough. They cannot form hypotheses and, even if they could, they have no way to test them.

The period when dejanews just started to index newsgroups was a golden age for finding answers on the internet, IMO, and there’s a strong similarity to the fediverse. All we need is for it to be searchable… OK, I see your point now.