moonlight

https://www.youtube.com/watch?v=YJg02ivYzSs

Well, it falls apart pretty easily. LLMs are notoriously bad at math. And even if it was accurate consistently, it’s not exactly efficient, when a calculator from the 80s can do the same thing.

We have setups where LLMs can call external functions, but I think it would be cool and useful to be able to replace certain internal processes.

As a side note though, while I don’t think that it’s a “true” thought process, I do think there’s a lot of similarity with LLMs and the human subconscious. A lot of LLM behaviour reminds me of split brain patients.

And as for the math aspect, it does seem like it does math very similarly to us. Studies show that we think of small numbers as discrete quantities, but big numbers in terms of relative size, which seems like exactly what this model is doing.

I just don’t think it’s a particularly good way of doing mental math. Natural intuition in humans and gradient descent in LLMs both seem to create layered heuristics that can become pretty much arbitrarily complex, but it still makes more sense to follow an exact algorithm for some things.

I considered this, and I think it depends mostly on ownership and means of production.

Even in the scenario where everyone has access to superhuman models, that would still lead to labor being devalued. When combined with robotics and other forms of automation, the capitalist class will no longer need workers, and large parts of the economy would disappear. That would create a two tiered society, where those with resources become incredibly wealthy and powerful, and those without have no ability to do much of anything, and would likely revert to an agricultural society (assuming access to land), or just propped up with something like UBI.

Basically, I don’t see how it would lead to any form of communism on its own. It would still require a revolution. That being said, I do think AGI could absolutely be a pillar of a post capitalist utopia, I just don’t think it will do much to get us there.

The math example in particular is very interesting, and makes me wonder if we could splice a calculator into the model, basically doing “brain surgery” to short circuit the learned arithmetic process and replace it.

Depends on what we mean by “AI”.

Machine learning? It’s already had a huge effect, drug discovery alone is transformative.

LLMs and the like? Yeah I’m not sure how positive these are. I don’t think they’ve actually been all that impactful so far.

Once we have true machine intelligence, then we have the potential for great improvements in daily life and society, but that entirely depends on how it will be used.

It could be a bridge to post-scarcity, but under capitalism it’s much more likely it will erode the working class further and exacerbate inequality.

The US isn’t the only country, and I bet people using the fediverse skew heavily toward android in the US anyway.

The flip style ones are pretty useless other than getting a preview on the main cameras.

The fold style, especially a trifold, seems pretty useful, and lets you basically carry around a decent size tablet in your pocket.

I definitely wouldn’t call it a gimmick, although the price and durability tradeoffs are nowhere near worth it for 99% of people.

coffin_dance.mkv

Same, it’s really easy to keep them separate mentally.

I think QWERTY may even be better for typing on a phone, because common letters are more spread out, leading to fewer errors.

except bifold and trifold are established terms for wallets, brochures, etc.

I think it makes more sense like this, anyway, especially considering that those words don’t really refer to the folds themselves. (trifold can just mean triple)

And here I am on GrapheneOS trying to get rid of Google. Although I personally would rule out Huawei for privacy concerns. (and the price, of course)

But if you don’t care about that, what’s stopping you from using Aurora store and your bank’s website?

What scifi movies/books do you like, then? Because I would say most great scifi is political.

The video shows how Honey doesn’t actually get you the best deals, with ‘approved’ discounts. I don’t care about affiliate links either, but I certainly don’t want to be making Paypal so much money for no reason.

There are color e-ink screens for things like comics, although I don’t think there’s any foldable e-ink yet. And yeah, lack of decent video playback means it would be kinda niche.

This isn’t about “AI”, it’s about greedy landlords.

It’s a waste of hardware, and a waste of energy that could be doing something useful.

I definitely think the newer design looks better, although it is less functional and has more unnecessary crap. I definitely prefer the compact view over the the card view. Back when I still used Reddit I used the old design with custom css though.

Ah yes, let me add some Microsoft spyware to my Google spyware

Yeah kitty is really good. I did mostly choose it for the name, but it might be the best terminal emulator out there. It’s fast, with a lot of great modern features.

I’ve been going to the gym 4 days a week (often just 2, but ideally 4). 2 upper body days, 2 lower body days.

For me, going to the gym means that I’m much more likely to do a full workout than if I stayed home. It’s also easier to properly target all of the important muscle groups. Machines are a good place to start. Use lower weight than you think, and really focus on feeling the muscle and doing slow, controlled motions.

I just do protein shakes, which is maybe not enough. I definitely don’t get enough calories, as I live alone and have a very hard time preparing food and eating a proper amount. Calorie surplus is important for building muscle. I’ve made very little progress because I keep losing any weight that I gain :(