infotainment

Yeah I feel like you could run a similar analysis on random non-AI Github projects and get the same result. Most of them never go anywhere.

If you primarily choose to watch educational videos sure, but YouTube can give you just as much brainrot as TikTok, depending on what the recommendation engine decides you might like.

The answer, IMO, is simply banning all algorithm-driven social media, for everyone and not just kids.

This conveniently sidesteps the identity/privacy arguments, makes it much easier to enforce, and would present an even greater net benefit. There is no benefit to algorithmic social media at all, and everyone would be better off without it.

The angle of this that people may not realize is that San Francisco already owns the power generation (ie, the Hetch Hetchy dam). [1]

PG&E simply owns the wires, and overcharges for their use.

[1] https://www.sfpuc.gov/about-us/our-systems/hetch-hetchy-powe...

The major difference is that this is currently the only New Glenn launchpad, which is likely to cause major launch delays.

The arrogance of American tourists is truly boundless. How dare Japanese people not speak English! Who do they think they are?

This attitude is so unbelievably prevalent among native English speakers. "Obviously everyone should speak *my* language -- why should I ever have to learn another one?"

100% this -- westerners love to criticize Japan's justice system, while ignoring the fact that much of it actually works.

Drugs? Petty crime? Homelessness? No other country comes close to managing these problems as well as Japan does, and Japan somehow manages to do this without descending into a 1984-esque surveillance state. Wander the streets of Tokyo at night and you will see zero drug-addicted homeless people. How many western cities could one say that about?

Andrej Karpathy can train a GPT-2 class model for less than $80 now, so at least the environmental cost of training may drop to a point that it's acceptable to LLM vegetarians: https://twitter.com/karpathy/status/2017703360393318587

I suspect that even if you reduced the cost of training or any other real world metric, the goalposts would immediately move. It seems to me that it has never been about those things, but simply about the feeling of superiority one can attain by eschewing something seen as trending.

Of course, which tells you the position from which the author of the linked post is arguing.

built on massive exploitation of human labor and make profligate use of scarce resources

This kind of hyperbole repeated ad infinitum by haters online is not-constructive, IMO. I would be quite certain that the manufacture of whatever computing device the author is accessing the internet on used far more resources and exploited far more human labor than training an ML model ever did.

I suppose opinionated people on the internet and ideological consistency have never exactly gone hand-in-hand.

Like the boring fence-sitter I am, I think it will have a mix of positive and negative effects.

Definitely the correct way to think about AI, but "nuanced and reasonable" is not the way of internet debates, sadly.

Always has been

Online shopping is a pox upon society

Don't worry, GIMP's UI never made any sense to anyone (Graphic Designer or otherwise), save for hardcore GIMP fans.

Did Meta spend around 60Mn lobbying for age verification to be forcibly added to every OS install ?

Of course they would want this -- as long as the OS reports that the user is over 18 via such a system, then Meta is legally off the hook for any COPPA violations.

The proposals, which return to Parliament on Tuesday, would replace juries in England and Wales with a single judge in cases where a convicted defendant would be jailed for up to three years.

Wow, this is literally the plot of the Phoenix Wright: Ace Attorney video games. I'm sure it will go great with no downsides.

This article misses a major point: even without nuclear weapons, NK has always had the option to shell Seoul with conventional artillery and completely destroy it.

That, above all else, is likely why no one tried to topple them before they were able to develop nuclear weapons.

I'd say if one has any interest in writing objectively about space technology, one will likely end up being perceived as having a "pro-Musk bias".

Elon himself is indeed questionable, but you really can't argue with his space-related achievements. Even other eccentric billionaires like Bezos haven't come close.

I do wish they'd release some open model weights though. Their attitude against open weight models in general is what's keeping me from subscribing.

Why should gambling be the exception? One could argue other app-based vices are just as bad, if not worse.

Device based attestation seems like the way to go largely; it doesn't solve the problem, but it's good enough that it would cover most cases.

This is such an entirely predictable outcome that people were warning about ever since the Patriot Act days and the creation of DHS.

Unfortunate, but the inevitable consequence of granting the kinds of powers that DHS was given.

Really slow FPS for me (macOS, Firefox). Not really getting past 1fps...

Should be an interesting test of camera performance!

On the plus side, the actually good mobile Anki client, AnkiDroid, remains out of the hands of this potentially questionable new entity.

(AnkiDroid has always been run independently, which is good, considering the state of the iOS client, which has always been neglected.)

That comment was in response to an article; I feel like they may have a different view about a 40+ minute YouTube video.

I always felt like Prolog's ability to execute programs was entirely accidental.

To me, it feels like a data description language that someone discovered could be tricked into performing computation.

I get OP's point though -- yes, all smartphones do tend to end up as e-waste in the end, but would you buy a smartphone if it couldn't ever be charged?

If, every time you powered on the phone, you knew you were draining from its finite battery life? I would probably think twice before doing any kind of software updates at least!

It shows them just fine -- the issue is that on iOS, only Apple Watches are permitted to perform notification "actions" (e.g., replying to a message, etc).

Recently, EU pressure might mean that Apple will open that ability up to non-Apple watches as well.

This 100% -- the AI features already in Firefox, for the most part, rely on local models. (Right now there is translation and tab-grouping, IIRC.)

Local based AI features are great and I wish they were used more often, instead of just offloading everything to cloud services with questionable privacy.