jprete

I think you're arguing in good faith, so my response is that I think the legal-piracy world would make it very easy to download things, on par with going to YouTube, something anyone could and would do. I don't think Netflix would be the same in that world. They might well have been the premiere free movie distributors!

Not that Trump is directly trying to solve it, but TSMC being in Taiwan has been a geopolitical problem for a long time. It's even more of a problem in a situation where China catches up on fabbing, which feels less like an "if" and more like a "when, if not already".

I was talking about a world where piracy is legal, which this one is not.

China doesn't care about climate change either.

Cancel culture was systematically supported and promoted by the most powerful people in our society.

If it were legal to download movies and music, Netflix and Spotify would absolutely not exist.

Steam is an unusual case, because games are running software and can't be trivially reproduced in their unencoded form. The publishers can include copy protection, network connection requirements, or even run essential parts of game logic on their own servers. So free downloads became a much worse experience over time.

None of those services could exist today if copyright didn't exist, because streaming services wouldn't be able to compete with free downloads. I think Patreon and Kickstarter are how creative work is funded in that world.

If you're looking for that kind of content, you could remove the minus sign?

Your comment about looky-loos got me thinking about real-world socialization. People who hang around the edge of a group in the real world, looking like they're trying not to be noticed, get pegged as creepy - it looks like predator behavior in the wild! The group then has a reason to either invite them to participate or keep them out - otherwise they're making the space feel unsafe.

The edge-hangers have two good choices: They can be very visibly present and then join the conversation when an opportunity presents itself, or they can leave.

Social-media edge-hangers don't have to do anything, they can just hang there forever without clearly joining or leaving the group.

A little off-topic, but have you found spam or fake content to be a problem in the last year or two?

I don't think the memo mentions training data sources; it's about usage and impact.

I think people are looking for the term "freeware" although the connotations don't match.

I don't think anyone would not give the new administration credit or blame after they came out of the gate wrecking-ball style.

There's no meaningful inspection of LLM code, because the real code is the model weights.

The way to get better is to do it a lot. Every time you dig through a problem to solve it, you're not just learning about that problem - you're learning about every problem near it in the problem space, including the meta-problems about how to get information, solve problems, and test solutions.

In a sense you're slowly building the LLM in your head, but it's more valuable there because of the much-better idea evaluation, and lack of network/GPU overhead.

I definitely think it's doing more than that here (at least inside of the vector-space computations). The model probably directly contains the paper-wood-log association.

Codenames is absolutely dead-center of what I expect Large Language Models to be good at. The fundamental skills of the game are: having an excellent embedding for word semantics and connotations; modeling other people's embeddings; a little bit of game strategy related to its competitive nature.

I was disappointed to realize how thin the veneer really is. And sometimes a music-carrying video will be unavailable in the music app because YT music and YT as video service apparently aren't subject to the same region rules.

Why would developers do that?

The developer's incentive is to control the experience for a mix of the users' ends and the developer's ends. Functionality being what users want and monetization being what developers want. Devs don't expose APIs for the same reason why hackers want them - it commodifies the service.

An AI-first app only makes sense if the developer controls the AI and is developing the app to sell AI subscriptions. An independent AI company has no incentive to support the dev's monetization and every incentive to subvert it in favor of their own.

(EDIT: This is also why AI agents will "use" mice and keyboards. The agent provider needs the app or service to think they're interacting with the actual human user instead of a bot, or else they'll get blocked.)

The link doesn't say that. The phrase you use is a reference in the Wikipedia article to the DOJ's characterization of Dmytry Firtash, "a Ukrainian oligarch who is prominent in the natural gas sector", not Giuliani.

Alexander the Great made his conquests by building a really good reputation for war, then leveraging it to get tribute agreements while leaving the local governments intact. This is a good way to do it when communication lines are slow and unreliable, because the emperor just needs to check tribute once a year to enforce the agreements, but it's weak control.

If Alexander could have left perfectly aligned copies of himself in every city he passed, he could have gotten much more control and authority, and still avoided a fight by agreeing to maintain the local power structure with himself as the new head of state.

Quite a few have succeeded in conquering large fractions of the Earth's population: Napoleon, Hitler, Genghis Khan, the Roman emperors, Alexander the Great, Mao Zedong. America and Britain as systems did so for long periods of time.

All of these entities would have been enormously more powerful with access to an AGI's immortality, sleeplessness, and ability to clone itself.

A long time ago I had the idea that maybe Guinness started a "book of world records" precisely because it answers exactly the kind of question that will routinely pop up at the pub.

In the absence of a definition I'd read it straightforwardly - it means that someone stops making an effort to learn better ways to learn. I.e. if they start using chatbots to learn, they stop practicing other methods and just rely on the chatbot. (EDIT: I realize now that this probably isn't news to the parent!)

I've heard stories of junior engineers falling into this trap. They asked the chatbot everything rather than exposing their lack of knowledge to their coworkers. And if the chatbot avoids blatant mistakes, junior engineers won't recognize when the bot makes a subtle one.

I get your point, and my first question was "what operations are even atomic here for this problem to be well-defined?".

But I think "language operations are atomic and everything else isn't" is a reasonable inference for a puzzle.

I'm also intrigued by locking being done through atomic assignments; I thought it required at least something like "write if equal to value" or "write new value and return old".

Totally fair. I was thinking more in terms of the rate at which people can consume it; if your primary interest is following a sport, or current reality-TV shows, you can only consume content as quickly as it is released.

The GP's statement doesn't work with reality TV or televised sports. Both of those are produced with a lot of human effort, and the cycle time for new content is way too large to form addictions.

Gambling, alcohol, and gacha games are clearly addictive and frequently are not set up to be in the best interests of the users.

I'm imagining exactly who makes these calls, if a human is even involved, as someone spending five seconds on a moderation app looking at screenshots and immediately pushing the "gambling" button so they can go on to the next task more quickly.

It doesn't have to be a human, either, at this point it could be AI trained on the same bad answers.

Those are two different concepts hiding in similar words. You can't [numerically or precisely] measure productivity, but some engineers are vastly more productive [such that you can easily tell the difference without a formal measurement].

I love Factorio, although I haven't gone far in the expansion. I don't expect any more from them. They've done a lot for this game and they've done it for a very long time. I assume they're now interested in doing something new and I hope they find success at that.