𝙲𝚑𝚊𝚒𝚛𝚖𝚊𝚗 𝙼𝚎𝚘𝚠

I have no issues connecting to my server when using my local DNS and self-signed certificates with the normal app either, or perhaps I’m misunderstanding you.

If producing an AGI is intractable, why does the human meat-brain exist?

Ah, but here we have to get pedantic a little bit: producing an AGI through current known methods is intractable.

The human brain is extremely complex and we still don’t fully know how it works. We don’t know if the way we learn is really analogous to how these AIs learn. We don’t really know if the way we think is analogous to how computers “think”.

There’s also another argument to be made, that an AGI that matches the currently agreed upon definition is impossible. And I mean that in the broadest sense, e.g. humans don’t fit the definition either. If that’s true, then an AI could perhaps be trained in a tractable amount of time, but this would upend our understanding of human consciousness (perhaps justifyingly so). Maybe we’re overestimating how special we are.

And then there’s the argument that you already mentioned: it is intractable, but 60 million years, spread over trillions of creatures is long enough. That also suggests that AGI is really hard, and that creating one really isn’t “around the corner” as some enthusiasts claim. For any practical AGI we’d have to finish training in maybe a couple years, not millions of years.

And maybe we develop some quantum computing breakthrough that gets us where we need to be. Who knows?

This is a gross misrepresentation of the study.

That’s as shortsighted as the “I think there is a world market for maybe five computers” quote, or the worry that NYC would be buried under mountains of horse poop before cars were invented.

That’s not their argument. They’re saying that they can prove that machine learning cannot lead to AGI in the foreseeable future.

Maybe transformers aren’t the path to AGI, but there’s no reason to think we can’t achieve it in general unless you’re religious.

They’re not talking about achieving it in general, they only claim that no known techniques can bring it about in the near future, as the AI-hype people claim. Again, they prove this.

That’s a silly argument. It sets up a strawman and knocks it down. Just because you create a model and prove something in it, doesn’t mean it has any relationship to the real world.

That’s not what they did. They provided an extremely optimistic scenario in which someone creates an AGI through known methods (e.g. they have a computer with limitless memory, they have infinite and perfect training data, they can sample without any bias, current techniques can eventually create AGI, an AGI would only have to be slightly better than random chance but not perfect, etc…), and then present a computational proof that shows that this is in contradiction with other logical proofs.

Basically, if you can train an AGI through currently known methods, then you have an algorithm that can solve the Perfect-vs-Chance problem in polynomial time. There’s a technical explanation in the paper that I’m not going to try and rehash since it’s been too long since I worked on computational proofs, but it seems to check out. But this is a contradiction, as we have proof, hard mathematical proof, that such an algorithm cannot exist and must be non-polynomial or NP-Hard. Therefore, AI-learning for an AGI must also be NP-Hard. And because every known AI learning method is tractable, it cannor possibly lead to AGI. It’s not a strawman, it’s a hard proof of why it’s impossible, like proving that pi has infinite decimals or something.

Ergo, anyone who claims that AGI is around the corner either means “a good AI that can demonstrate some but not all human behaviour” or is bullshitting. We literally could burn up the entire planet for fuel to train an AI and we’d still not end up with an AGI. We need some other breakthrough, e.g. significant advancements in quantum computing perhaps, to even hope at beginning work on an AGI. And again, the authors don’t offer a thought experiment, they provide a computational proof for this.

https://github.com/cheeaun/phanpy?tab=readme-ov-file#easy-way

It’s fairly literally just a download-and-run kind of deal it seems. Does seem pretty trivial.

I won’t pretend I understand all the math and the notation they use, but the abstract/conclusions seem clear enough.

I’d argue what they’re presenting here isn’t the LLM actually “reasoning”. I don’t think the paper really claims that the AI does either.

The CoT process they describe here I think is somewhat analogous to a very advanced version of prompting an LLM something like “Answer like a subject matter expert” and finding it improves the quality of the answer.

They basically help break the problem into smaller steps and get the LLM to answer smaller questions based on those smaller steps. This likely also helps the AI because it was trained on these explained steps, or on smaller problems that it might string together.

I think it mostly helps to transform the prompt into something that is easier for an LLM to respond accurately to. And because each substep is less complex, the LLM has an easier time as well. But the mechanism to break down a problem is quite rigid and not something trainable.

It’s super cool tech, don’t get me wrong. But I wouldn’t say the AI is really “reasoning” here. It’s being prompted in a really clever way to increase the answer quality.

It’s not a direct response.

First off, the video is pure speculation, the author doesn’t really know how it works either (or at least doesn’t seem to claim to know). They have a reasonable grasp of how it works, but what they believe it implies may not be correct.

Second, the way O1 seems to work is that it generates a ton of less-than-ideal answers and picks the best one. It might then rerun that step until it reaches a sufficient answer (as the video says).

The problem with this is that you still have an LLM evaluating each answer based on essentially word prediction, and the entire “reasoning” process is happening outside any LLM; it’s thinking process is not learned, but “hardcoded”.

We know that chaining LLMs like this can give better answers. But I’d argue this isn’t reasoning. Reasoning requires a direct understanding of the domain, which ChatGPT simply doesn’t have. This is explicitly evident by asking it questions using terminology that may appear in multiple domains; it has a tendency of mixing them up, which you wouldn’t do if you truly understood what the words mean. It is possible to get a semblance of understanding of a domain in an LLM, but not in a generalised way.

It’s also evident from the fact that these AIs are apparently unable to come up with “new knowledge”. It’s not able to infer new patterns or theories, it can only “use” what is already given to it. An AI like this would never be able to come up with E=mc2 if it hasn’t been fed information about that formula before. It’s LLM evaluator would dismiss any of the “ideas” that might come close to it because it’s never seen this before; ergo it is unlikely to be true/correct.

Don’t get me wrong, an AI like this may still be quite useful w.r.t. information it has been fed. I see the utility in this, and the tech is cool. But it’s still a very, very far cry from AGI.

This is true, but it’s specifically not what LLMs are doing here. It may come to some very limited, very specific reasoning about some words, but there’s no “general reasoning” going on.

I probably change the brightness setting the most. Why is it all the way at the top of the settings, the furthest out of reach it could be?

Both WhatsApp and Signal show the same amount of chats to me (9 for both). WhatsApp does show a small sliver of a tenth chat, but it’s not really properly visible. There is a compact mode for the navigation bar in Signal, which helps a bit here.

From what I can see there’s slightly more whitespace between chats, and Signal uses the full height for the chat (eg same size as the picture), whereas WhatsApp uses whitespace above and below, pushing the name and message preview together.

In chats the sizes seem about the same to me, but Signal colouring messages might make it appear a bit more bloated perhaps? Not sure.

Shareholders can demand external audits under threat of selling the stock. There’s plenty shareholders can do (and have done in the past). They don’t just sit idle and not do anything you know.

Shareholders seek to maximize profits. If that includes a lawsuit to squeeze out even more investments, then why not?

They never bothered to check if Boeing did what they had to do security wise. Only once it threatened their profits they sprang into action.

⛤

I think the current logo would work fine as a unicode character. I dislike the three anuses for a logo.

If I understand correctly they already can. It’s not user-facing, but votes are federated if I understand correctly.

That’s a very specific usecase though that the majority of programmers likely will never have to face.

You’re mixing the definitions of sideloading media and sideloading apps. Sideloading media follows your definition, e.g. transfer via another local device. Sideloading apps refers to the installing of apps outside of the (pre-)installed app store, e.g. by installing an APK directly.

I don’t think we are, at all.

I mostly say it because SMS is so ancient. Not encrypted, messages are storied by the carrier and can be requested by the government, etc… In that sense, even a corporate-controlled messaging system that offers E2EE would be a step up. After all, SMS is pretty corporate-controlled too, just different ones. But again, this is very much a European perspective, I can see why in the US this might be different.

iMessage is by far and away the most popular chat platform here, and is largely responsible for Apple’s local dominance in the smartphone market.

Ah true, iPhones are much more popular in the US. Quite interesting actually how that happened, iPhones aren’t all that popular here at all and Android phones dominate the market. I wonder why Apple hasn’t managed to copy their dominance here as well?

I’ve never heard of such a thing.

Looks like Tello, Cricket, MobileX, US Mobile and T-Mobile can offer it at least. Apparently it’s often marketed as a Tablet plan, which I suppose makes sense, but it seems a lot of carriers allow you to disable SMS in their web portals these days. I thought it’d be more niche in the US but it seems a more common option than I thought.

It’s been interesting to hear from you about your perspective on this, thanks!

I personally occasionally experience delayed sending/receiving of messages, or messages suddenly coming in in bulk. Only very rarely do messages not come in at all thankfully, but mostly the occasional delays in sending/receiving I think led to the reputation of poor reliablity for SMS. But it makes sense that the US would try to keep those issues to a minimum if so many people still use it, whereas in Europe perhaps it’s less of a priority?

Ah, I see. Your point isn’t necessarily reliability but availability. It’s an interesting perspective to hear that the US appears to be so behind (at least from a European perspective of course) when it comes to messaging apps. As far as SMS reliablity goes, I have occasionally had messages not send, or have messages come in delayed considerably. Or stuff like 2-factor auth texts not coming in, requesting a new one and then suddenly receiving 3 at a time. Not deal-breaking or anything, just the occasional annoyance.

I don’t think WhatsApp allows you to send a message to someone who doesn’t have the app. So WhatsApp would just inform you. Although I don’t recall the last time someone did not have either WhatsApp or Signal installed. But again, that appears to be far more common in the US?

Do you ever miss the extra features that web messaging brings, like in-chat polls, voice messages, etc…? I’m not sure how much of that RCS supports (because almost nobody uses that here). To me it seems like the convenience of web messaging outweighs the “does person x have app y” question, but that’s probably because I never really have to ask myself that question.

I also just realised that you state that everyone has SMS messaging. There are phone plans available here that don’t offer SMS messaging anymore. You can still receive them, but sending them either doesn’t work or costs a high premium (obviously this disadvantage is offset by a lower price for the rest of the plan). I wouldn’t be entirely surprised if SMS eventually just gets phased out.

If you send someone an SMS you know with great certainty that it will be received.

Not to get too bogged down in this debate or anything, but this European is surprised you say that SMS is reliable. One of the great motivators to use web-based messaging apps is because SMS is so notoriously unreliable, with messages occasionally not receiving or sending. Has SMS reliablity been improved much in recent years? Or is web-based messaging less reliable in your experience?

Genuinely curious btw, I’m not in the same party as the troll elsewhere in this thread.

In the Netherlands 112 is fine, most critical systems are. It’s mostly airports that are getting fucked by this it seems.

Banks and PSPs are fine here too.