I don’t think it would be possible if the machine is running on it, since it would be constantly trying to write logs, and errors would arise when that write operation failed.

Dehumanisation?

It’s all just that bit less human.

You can usually configure the kept history interval, if it helps.

Also the between of important enough you’d like to come back to it/review it, but not important enough to put into the bookmark hole.

It’s a bit early, isn’t it? Their election’s only just been.

What’s the downside? Learning a language, with a completely different structure, would be an amazing thing to have in the toolkit.

It’d be quite interesting to compare it against English and see where the similarities and differences are, for something with very different roots.

Also agriculture.

There aren’t a lot of industries who have not only overthrown sovereign nation-states, but have done it so often, and are known for it enough that there’s a well-known term (banana republic).

They should bring back computer classes in school.

The entire point of a school is to learn, and learning how to computer is an excellent thing to have.

Want something faster? Look into business small PCs they are about the size of the small ps2 and will have Intel i3s to i7s. I’d just recommend looking for one that is at least the 8th gen Intel if you care about power efficiency. If using eBay search for used auctions and use a bid sniper that uses the official api

Are they still available? When I looked recently, most of it was Haswell/4th gen, and/or the seller had taken nearly all the parts out, like the memory, drives, and CPU.

In the sense it digged its own grave, yes.

I don’t know about that, mostly because the structure of LLMs at the moment make them incredibly bandwidth-hungry, and also quite sensitive to latency.

You do have some small-scale versions of that, where people link a few smaller computers together to run a larger model, but that usually exists in the form of things like connecting machines via USB4/Thunderbolt, or Oculink, which are much faster and lower latency than a typical internet connection.

There is something a bit like that, with the Kobold AI Horde, but that is still people hosting an entire model and sharing that over the internet, rather than something like Folding/SETI@Home, where you’re doing part of the actual calculations yourself.

Hardware is used way more efficiently in the data center, and any sort of deficit is best resolved in a centralized system due to more efficient resource allocation. The same setup could be used like 2 hours a day for your personal needs, or 24 hours a day to serve many people.

Only for a rather narrow definition of efficiency, though. For raw compute power, yes, but you’d also be adding a whole bunch of overhead from all the network traffic and routing that would be necessary for that sort of thing to work in the first place.

If people used their desktops like in the old days where you’d connect to a mainframe or master computer, you’d end up needing computing hardware on both ends anyway, since the user side would need to be enough to be able to connect and send all the interactions to the master computer. At that point, you could just buff up the client computer, and do all that in the user’s home.

There’s also an argument that some of the basic problems would still apply anyway. A data centre does not fully utilise 100% of its capacity. There’d always be some reserve to have failover if a server dies, or there’s a sudden spike in load, like with a home machine.

There’s also an argument to be made that they could have just had an integration to let an iPad be used as a touch screen interface.

You can already use a Mac’s keyboard and mouse inputs on an iPad, it being able to go the other way wouldn’t be particularly strange.

It does not, but you can block/unsubscribe from communities that have too many of them on the main feed, or block particular users from showing up, which can help with the repeats. That otherwise takes more logic on RSS, and isn’t really possible on the website.

It’s not entirely wrong, in that Reddit is so full of bots and junk that a sizeable chunk of the site is unusable. Any animal photo subreddit often has more bots than a farm.

It’s either personalise or have the deluge.

Coincidentally, Reddit has just announced plans to shutter their RSS feeds, so if you were using those to personalise your feed, good luck.

Probably the pollution. A lot of AI companies are trying to find ways around it, by including watermarks in their model output that their tools can detect, and so on.

We already know that training models on themselves causes model collapse, so OpenAI deciding that they’ve had quite enough wouldn’t be surprising.

On top of that, a lot of websites are clamping down on access because of scraping concerns. Reddit shut down a lot of their third-party APIs, and nicer features, like RSS, Old Reddit (if you haven’t used it in 6 months, you lose access to it, and you need an account for it), etc. Google seems to want a captcha on every search. Twitter famously ramped up the price of its API immensely,.and now has ridiculously stringent measures for accounts, like needing users to download the app, and give hand-scans. It wouldn’t be at all surprising for OpenAI to decide its not worth the effort at this point to try and get around all that, nor pay for it, when they’ve already got all the previous data. Training doesn’t destroy the data used.

A lot of AI training data these days is generated synthetically anyway, and that can be done in-house, whether by their own models, or downloading an open weights model and running it themselves.

And even if you take one down, it could just be uploaded again as another person.

It’s basically like the problems with mobile games or YouTube channels. People are clearly using a template/framework, and can release billions of them. So when one gets taken down, another can just pop up.

And if Steam is too hard on small devs, they start locking out indie devs or people who genuinely put effort in. There’s no good answers.

Supposedly a lot of the big model companies are putting in watermarks, but they generally don’t publicly release the exact things they’re checking for, and it differs per model, so it’s not like there’s a good reliable method of picking it out either.

There’s also a practical element, where Steam is not in a position to vet every game and developer. They could have the community do it, but that then opens up a whole bunch of abuse risks.

There’s also an argument that it wouldn’t stop some of the issues in the article. Where do you draw the line with asset flip, or AI generated? If the developer used AI to make placeholders in the early access of a game, and then swapped it with removed it later on, is the game still AI generated? What if they used assets that were unknowingly AI generated, like stock sound effects?

Or if someone is ESL and used a translator that turned out to be AI powered/they just write with a similar style.

It also won’t stop people doing it.maliciously. if you’re threatening to clone someone’s game off a demo, and release it before they do, you’re probably not going to stop because you need to put a little “AI generated” flag on your game, if the rest of it didn’t stop you.

Even at worst, where your candidate doesn’t win, it wouldn’t hurt anything.

At least you’d be doing something, compared to doing nothing.

> The large language model chatbots use is not a path to sentience.

We don’t know that and won’t know that until there are sentient machines.

You’re right, in the sense that we wouldn’t know what is a path to sapience until it happens, but we can only work with what we know, and what we do know is that the transformer framework language models use is not what a brain is, and a brain is, as far as we know, the only path. The artificial neural network technology transformers are based on is rooted on a mathematical representation of an extremely simplified model of how neurons work. Human minds don’t work in tokens, and brains have chemical signals and things that also have an impact. The physical wiring is part of the story.

We’d need to objectively define what sentience/sapience is to begin with, anyway. We can’t even do that for people and animals, expanding it to machines seems an exercise in futility. It wasn’t that long ago that babies and animals were considered to be little more than automata as far as consciousness goes.

It would be amazing to be proven wrong, because the implications for philosophy and neurology would be staggering, but I digress.