svachalek

Some of this stuff is ludicrous. I finally tried /loop last week and discovered every loop iteration passes the entire context history to the model. So pretty quickly you're running a full 1M context window, without cache, likely just to check if something is ready or needs to be done. It's miserably terrible engineering unless your entire and only goal is to burn tokens.

I haven't seen either of these running outside their creator's services yet, but typically you can watch services like openrouter or nano-gpt for it to show up at a (usually small) discount.

You've got to consider how much power that uses though. Depending where you live, some of these providers can serve it for less than you pay for power.

I'd imagine it has to be immense. Most of the engineers I know are spending $200/day, let's say 20 workdays per month, $4000/mo. Whatever % of revenue it is, it's got to be close to 100% of profit.

The Claude-isms are so heavy in this it's painful to read: "Next: what an actual structural reset requires, once the cheap corrections are gone, the workarounds are load-bearing, and the people who would have made it easy have already left."

And if they did, you still don't need to be developing a Twitter-bot LLM and/or nudify image model to support your rocketry projects.

From your description it may be acting as a placebo -- the act of putting the pill into your mouth is what tells your subconscious to activate a sleep cycle.

Well it's not the dev who would want one, it's the agent, for the same reliability/security reasons that a dev historically would have.

Because absolutely no one is marketing apps as being superior, but instead this is just the experience of users.

I guess you haven't used the mobile web? Practically any website you use covers half the page with a banner saying don't use the website, the app is SOOO much better.

It's about tracking. Most of the money in the industry comes from knowing precisely when you were on the toilet every day and they can't get that from a web app.

You can implement pagination exactly the same way. It's a UX decision that has nothing to do with underlying queries, although it typically maps.

The typical infinite scroll that I've seen implemented does not work the way you describe though, it's just pagination without controls. The reason it works is because it's pushing content you never asked for anyway and it just keeps pushing. Without any sense of pages you'll never know the difference.

With the amount of gerrymandering and campaign finance going on, there's a question of whether this is a democracy in the first place. If it is, it's clear it can't be trusted to elect an executive who follows the laws.

What data? I've never seen data showing people approve of their Congressperson. Especially if that person is a Democrat.

Maybe they were playing Gobblet Gobblers instead which is a much more fun game on the same size board.

The Netrunner card game is very asymmetric and very modern. I guess you could say it's balanced but nothing asymmetric can be perfectly balanced.

Depends on if they are launched serially or in parallel then.

It's not a legal obligation, no, but neither should you as a customer accept a vendor that treats you like that.

Yeah IIRC the latest Pro is $12 and Flash is $9 which is not the usual 2X-3X multiplier we see separate model grades. It also puts Flash now about 2X GLM 5.2, which is a highly capable open weight model.

I'd say the prediction is correct, as the acquisition is more or less just a better way to capitalize on the bankruptcy.

Yeah I remember some studies showed this with overly sexy ads. They were very memorable to the audience but all they remembered was hot chicks, they couldn't recall the product.

There's practically nothing you can get 100% agreement on, and yet we can still find that there is a right answer.

More importantly, the download is out there. You can download it yourself today, and if it's that important to you, you can buy the hardware too.

If you don't think Anthropic and OpenAI are multi-trillion dollar militarized behemoths you need to catch up on some news.

Both are planning $trillion+ IPOs this year. OpenAI is collaborating with the Department of War, and Anthropic is under intense pressure to do the same and their top model is being held hostage right now. This week, the Department of War wrote a statement that xAI should not be held accountable for environmental laws because Grok is a vital weapon system of the US and was used to fire over 2000 missiles at Iran. The pentagon's statement mentions there are 3-4 such models so you may be able to guess which they are.

What does further progress get us? Mass unemployment? Extinction? Pick your dark future science fiction?

The happy ending where we're all living in a garden of eden cared for by benevolent AI is hardly worth considering when you look at the cast of characters who are in charge of the world right now.

Is an attempt to copy all or parts of a model an attack, when models have very questionable copyright status? Maybe? I don't think most people have much sympathy here though.

Dang, where do they pay 2x?

I think we need to disqualify humans as well. Their brains have been shown to operate on probabilistic chemical interactions and even quantum effects.

There’s a difference between having normal levels of difficulty and bad luck, and having people blame those on the wrong thing, vs having extraordinarily miserable quality and having people find the obvious difference. Potentially yes, they might have terrible wiring in their office or a crippling fondness for vim. But if I were their PR department I’d be talking about that if it was the problem.

I keep getting this ad for... Meta? glasses that shows a guy at a concert I guess filming it with his glasses. It zooms out and everyone around him has their phone in the air but not him because he's super cool.

It's like one of those puzzle images where you're supposed to find 20 things wrong with this picture.

The yield triangle is super common here (California and everywhere else I've been in the US). It's even on your image.

No parking is usually a red curb or striped out area, which is different but also nonverbal. The complication is that many places have a EULA on parking spots that reads something like "No parking, 9-5pm, except on Tuesdays and full moons, or in a yellow vehicle, or by written agreement with a minimum of two signatures not including Bob". Good luck putting that in an icon.

That "no stopping" sign is unfamiliar to me, I guess we write it out but it's a pretty rare thing here to not allow stopping.

Most of the rest is familiar and/or obvious except the ones that have German words on them. Many of the concepts in writing don't exist on the German signs, I don't know if you don't have these signs or just not listed on your chart. Some of the English ones are just tooltips; it's never ok to stop on a railroad track but someone thought a reminder would be nice.