avaer

8 часов назадdiscuss

Why not a "stop killing AI" movement?

If a company deploys a paid AI model and makes people depend on it, they need to dump the weights at EOL.

avaer

вчераdiscuss

I'm impressed. It feels like a faster Fable (probably due to the more efficient token usage). It performs roughly the same job, just with 4x less steps (gamedev).

Remains to be seen how the "shorter prompts" advice translates to homogeneity/collapse though.

avaer

позавчераdiscuss

The thing that rubs me the wrong way about WP (and one of the core issues in this controversy) is "taxation without representation".

Wikipedia is perfectly happy to promote the grand narrative that they are the thankless sacred keepers of humanity's knowledge. They'll take your edits. And they are perfectly happy to take your money for this divine goal. Just take a look at the marketing they plaster onto every page during the fundraising drive.

But the second that someone commits the grave sin of "canvassing" outside of Wikipedia to bring attention to an issue they care about, they get banned, and the discussion is ended. There's just a total lack of accountability for anything unless you play by the thick codex of online governance rules they made up.

It's why I haven't ever donated anything to Wikipedia.

avaer

позавчераdiscuss

If the data and code is all there, why not release the 5B weights?

avaer

позавчераdiscuss

Seems like an application of Goodhart's law; measuring worth by degree or grades stopped measuring learning or ability.

This was a lot harder to cheat before AI, but now the floodgates are open and grades and degrees earned post-AI are showing that they mean little.

Cheating on college tests should be a jailable criminal offense (similar to computer fraud) so that there is dignity in the degree again. Considering the money involved, I don't see why not.

But this probably won't happen, because many rich people are very happy to buy their degrees. See also [1]

https://stanforddaily.com/2026/04/09/the-real-reason-student...

avaer

3 дня назадdiscuss

"The cars have all have cameras checking for bad behavior, why shouldn't your phone and laptop?" said the esteemed lawmaker.

"Oh course there will be exceptions for politicians and authorized individuals, for national security reasons."

avaer

4 дня назадdiscuss

Why would that comparison be bad if it's accurate?

Also, "sufficiently advanced technology is indistinguishable from magic". Both can be made good and evil.

avaer

4 дня назадdiscuss

Those are not compelling arguments.

If the best we've got for convincing people to learn to code is that it's like math notation (the most hated part of math for the uninitiated), or pretty like a violin (useless for a new grad), then coding is in serious trouble.

IMO a better argument is it helps you "think like a computer". But if you wanted to learn that there are many video games I'd recommend mastering instead of learning to code. For most people "learn to code" is like telling programmers to "learn asm".

(I've been coding ~30 years)

avaer

4 дня назадdiscuss

It really depends on what a word is...

If a WORD is 16-32 bits, you could pack in quite a lot. If a word is a Unicode word, that's even more bytes. Or you could do this in German and go nuts.

avaer

5 дней назадdiscuss

I don't know, compute is compute. Arguably making complex software with LLMs isn't all that different from training a model to do a thing. You're throwing a lot of compute at the problem and hoping for a stochastic solution. The distinction will become even blurrier with time.

Though I agree it might be informative to split it by industry sector.

avaer

5 дней назадdiscuss

I wonder if this is going to come to the West too. It's not like the harms or problems are any different, only the flow of money and power and willingness to enforce is different.

Considering that gaming is the world's biggest entertainment industry, I don't know how you'd fairly enforce which manipulative virtual interactions are ok and which ones are not, unless you just pick and choose by fiat. Which China is glad to do (they've been doing this for a long time with games), but it probably wouldn't fly in the States.

avaer

5 дней назадdiscuss

It should be pretty obvious any frontier AI company worth its salt will steal any and all data it can, by whatever means necessary. They've all been doing it so far. It's a tragedy of the commons.

I wouldn't even limit it to "AI research". They are looking to expand into other areas (see: OpenAI's superapp ambitions), and they will take your other business ideas, research, and code too. If they get caught they will blame it on the AI and lawyer up with more dollars than you.

I think you might be underselling the scope of the problem, and I don't think it's going to be fixed by some person deciding not to call the LLM today.

avaer

5 дней назадdiscuss

If you architect your protocol cleanly, you should be able to run the simulation client side too without much effort (certainly the web platform has everything you'd need). This is how modern game engines do it, it goes back to the Quake VM and probably beyond.

You'd still get multiplayer "for free", but it could be turned on and off. You could do it with zero code change for the actual games, they don't have to know.

It seems some of the games could do with a singleplayer mode that doesn't depend on the backend having free slots.

Just an idea from a fellow web games person!

avaer

5 дней назадdiscuss

Does everything need a "session"? Is it possible to have fully static games/exports?

avaer

5 дней назадdiscuss

Maybe you think gaming/media companies are greedy and should make less money. But if you want the industry to continue in the current state, keep in mind the price of games needs to be adjusted:

  price *= (total game sales / avg player count)

If there's a million sales but only 10 thousand people playing at a time, the price of the game needs to be multiplied by ~100x, because these copies can be shared, and the sales would be divided by 100x if the copies can be efficiently shared. The modern internet would make this buttery smooth to do (companies that make this easy will pop up overnight).

I'm not arguing either way, but this is the back of the napkin math to consider, and how that would ripple across the industry, for better or worse.

avaer

6 дней назадdiscuss

There's also a long list of government (or subpeonable) entities on your certificate trust list.

Without which TLS is not gonna work.

The article is arguing that in practice you could just send your "encrypted" communications to the browser vendor, or one of the governments on the certificate root list, or someone else in the distribution chain, and have them be the middle man. The security properties of your communications would be the same. Hence "snake oil".

Things like stapling don't change this much, or reduce to TOFU.

avaer

7 дней назадdiscuss

A funny piece, but what if all of the following are true:

  - AI *is* how real research is done
  - Incentives are such that this is how researchers must work
  - Academia institution does not have a structure to accept this reality

Academia did this to itself. By hiring based on volume of output/citations, the way for researchers to win is to game the system better than the next person. AI just exposes that the old power structure doesn't serve much academic purpose.

Maybe we should take away the grants from the universities and put them into autoresearch loops + human reviewers?

avaer

9 дней назадdiscuss

The blast radius is far worse than any "malware" Google could protect you from.

TFA is playing it up, but it is arguable that this is a real virus, except the shady hackers are Google.

avaer

10 дней назадdiscuss

They aren't competing with Steam. The console market is a closed cabal where console makers sell the machine at a loss and make up for it with locked down software where publishers pay a significant proportion of the sales to the console maker, who controls supply and dealflow with private contracts.

They might lose, but it's nothing like PC.

avaer

10 дней назадdiscuss

Totally valid.

If someone thinks they're building better open source with their AI, let them fork; their AI can maintain downstream. If it's really better, people will join the fork. Good luck.

In all likelihood anyone attempting this will realize the value that a maintainer provides. On the odd chance they discover a new working model and produce better software, all the better, everyone wins.

Submissions

Decache

Telescript (Programming Language)

Industrial Policy for the Intelligence Age [pdf]

Basilisk Collection

Zelda's Z-Targeting

WebSR: Real-time AI Upscaling of videos and images on the web

YC: The AI Agent Economy Is Here [video]

Unicorn Jelly

Waypoint-1: Real-Time Interactive Video Diffusion from Overworld

EmulatorJS

WorldPlay: Real-Time Interactive World Modeling

comments