I’ll be honest, the real pictures would fool me. The era where you could tell by the number of fingers or toes is gone, apparently. She does seem to change military branches a bit too often to be realistic. And, her colleagues seem a bit too unbothered by someone taking a picture of her with her shoes and socks off and feet up on a desk in an apparent military setting. But there aren’t any pictures I glance at and think “oh, that’s obviously AI”.

LLMs are an obvious dead end when it comes to actual “intelligence” or understanding how the world works.

But, this sounds like a “draw the rest of the owl” situation.

“JEPA learns abstract representations of how the world works, ignoring unpredictable surface detail.”

Oh, it’s that simple is it? Just have it “learn abstract representations of how the world works”. Amazing how nobody thought to do that before!

I think I understand the distinction they’re trying to draw. Current models are trained on billions of pictures of cats and billions of pictures of dogs. You feed it an image of Fido and it finds a point in 2500 dimensional space and knows whether that point is in the “cat space” or “dog space”. It can be very good, but it doesn’t have any “understanding” of what makes something a cat vs. a dog. Humans, OTOH, aren’t trained on billions of images. But, they learn about things like “teeth” and “whiskers” and “snouts” and “eyes”. Within their knowledge of eyes, they spot that vertical slit pupils are unusual and different, and part of what makes something “catlike”. AFAIK, nobody has ever managed to create a system that learns abstract features without intensive human training.

I like that they’re trying something new. But, are they counting on a massive breakthrough on a problem that has existed since people first started theorizing about AI? Or, is it just a matter of refining a known process?

I hope whoever this was recorded these snipers for as long as possible.

I have an old YouTube app on my iPad, and it still works fine. One of the more responsive apps on the device. I get nagged nearly every time I use it to update to the newest YouTube release, but that’s impossible. I’d first have to upgrade my OS, and Apple no longer releases new OSes for this generation of iPads. So, I’m stuck with an old YouTube, which mostly works fine, and an occasional nag message.

I’m sure within a year or two mine will be like yours and YouTube will simply no longer work. But, for now it’s in a relatively good spot where I can use a version of YouTube designed for this particular hardware that doesn’t feel sluggish.

You do really feel this when you’re using old hardware.

I have an iPad that’s maybe a decade old at this point. I’m using it for the exact same things I was a decade ago, except that I can barely use the web browser. I don’t know if it’s the browser or the pages or both, but most web sites are unbearably slow, and some simply don’t work, javascript hangs and some elements simply never load. The device is too old to get OS updates, which means I can’t update some of the apps. But, that’s a good thing because those old apps are still very responsive. The apps I can update are getting slower and slower all the time.

In my experience, LLMs suck at making smart, small changes. To know how to do that they need to “understand” the entire codebase, and that’s expensive.

Amazon saves the wages of a senior dev by doing that, but then they get outages costing them the wages of that senior dev for decades. I doubt the goal is to blame senior devs. If they wanted them gone they could easily fire them.

What is AI good at? Creating thousands of lines of code that look plausibly correct in seconds.

What are humans bad at? Reviewing changes containing thousands of lines of plausibly correct code.

This is a great way to force senior devs to take the blame for things. But, if they actually want to avoid outages rather than just assign blame to them, they’ll need to submit small, efficient changes that the submitter understands and can explain clearly. Wouldn’t it be simpler just to say “No AI”?

Putting young-person slang in a movie has been a bad idea since before Gen Z existed.

Nature is healing.

My hate of the copyright-ownership side of Hollywood / Nashville / Atlanta, etc. has been burning white hot since the days that the RIAA was suing people using P2P networks. But, I had to admit that at least they could probably make a valid claim for copyright infringement. But this?!

It’s interesting how it’s the “Performing Right Society” (which I’ve never heard of). The “performing” part of that suggests that maybe they have an issue with people sharing clips containing music, or live streaming games where they share music. But, again, why Valve? Sure, people can share clips with friends. And, occasionally you see developers streaming their games. But, nobody is really “performing” live streams on Steam. I suspect they just think Valve is rich and so they can strong-arm them and Valve will settle to make them go away. I hope they bit off more than they can chew. Valve is indeed rich, and they have a tendency to be stubborn. I think they might well fight, and fight hard.

I wish a possible outcome was that the PRS ceased to exist. But, I suspect they’re like a flea or something, and even if you knock them off from this attempt to suck someone’s blood, you can’t kill them, and they’ll just find another victim.

To adults? Or teens? Or 6-year-olds?

which was probably written by AI

That’s the key bit. You would expect that one of the first tests for a custom OS would be whether it runs doom, so of course the AI is going to write the text that you’d expect to see: “It runs doom”.

The number 69 has staying power. It was hardly new when it was used in Bill and Ted’s Excellent Adventure, and that was in 1989, 37 years ago.

How long will 6-7 last? I’m guessing not more than a year. I bet even now it’s being included as part of a script for a kids’ movie, and by the time the movie comes out the kids will all think it’s “cringe” (or whatever term replaces cringe).

deleted by creator

Except real money is used for taxes and government spending. Fake money is needed only by criminals.

Bitcoin actually has legitimate uses

Ransomware.

“Expensive cartoon monkey pictures”

Do you need a microscope?

D4s are annoying. They’re the only die without a top face, so it’s hard to read the number, and since the faces slant inwards it can be hard to pick them up.

On the other hand, it’s satisfying that the typical RPG dice contain the 5 platonic solids. If I had to play a game where a D4 was used often (unlike in say D&D where it’s rare), I’d probably go with one of the alternate D4 shapes. Even if you prefer platonic solids, you can just use an 8-sided die and renumber 5-8 as being 1-4 again.