The overall model is similar to tap-to-pay debit cards. They're only used for consumer-to-business payments. When you tap the card, the card sends over an account number / signature / etc, which the merchant sends to a central clearinghouse to finalize the transactions.

The main difference is that the card itself keeps a running balance of how much money the customer has available to spend. In many cases, this gives the merchant enough confidence to e.g. let you through the train turnstile without actually waiting for the central clearinghouse to confirm the transaction. (I think in practice they usually send all the transactions in batches, daily or weekly or something.)

The readers do some trusted-computing/secure-enclave type stuff but are not especially hard to obtain; I think there have even been cases where companies like Nintendo have built them into consumer products, so that you could e.g. tap your card to your Nintendo 3DS to buy a video game.

I imagine there's a bit more security on the machines that let you load money on the cards, but it's probably not completely impossible to make a fake card. But the low value limit (usually a couple hundred dollars, depending on the card provider), the inability to get cash out of the system (often you can't even buy things like postage stamps), and the fact that you'll get caught relatively quickly (once the central clearinghouse notices the transactions don't match up) make it unattractive to do it in practice.

> What makes it aggressive?

The parent comment already contained the answer to your question (the multiple links are what makes it aggressive, in GP's opinion). Your comment might have been seen as more constructive if it engaged with that directly.

Well, you're allowed to write it, but the browser will just ignore the marker. The range will end at the tag, the same place https://github.com/WICG/declarative-partial-updates/blob/mai... says it goes by default if you don't write the at all. (That default wouldn't make a ton of sense if patch ranges were allowed to break out of their parent elements like this.)

Jake Archibald thinks it'd be cool to use a similar syntax to do CSS highlights, but there are other problems with that (like not being able to create overlapping ranges), and he's not on the Chrome team that's pushing this proposal.

Truly, a design only an LLM could love.

That team got acquired by Amplify, and eventually the "Desmos Classroom" tool used to make the activities got acquired by Amplify too, leaving Desmos with just the more consumer-focused graphing calculator stuff. Amplify also acquired this other "Polypad" product and kind of merged the two together into what's now called "Amplify Classroom".

I think Amplify is still selling/distributing a lot of the activities that were originally made by the Desmos people.

Hacker News hides the reply link on deeply nested replies for a little while to try and prevent flamewars. https://github.com/minimaxir/hacker-news-undocumented#hidden... says you can work around this by clicking on the comment's timestamp.

The second prong of the argument is basically that, when you invest in Anthropic, you can't just invest in one model and then collect the profits from that model. You're investing in a whole company in the hopes that they can be profitable overall; at some point they'll need to stop spending so much money on training and give it back to the investors instead. Zitron argues that this isn't going to happen because training is actually something that companies need to do to retain customers at all. An analogy here might be the fact that Microsoft has to spend a certain amount of "R&D" budget fixing security vulnerabilities in Windows Server just to retain their current customer base; if attackers found out about a serious security hole but Microsoft didn't fix it, everyone would need to stop using Windows Server. LLM companies do the same kind of thing to fix "jailbreaks" and other unexpected model behaviors.

The third prong of the argument is that, in general, there's a long history of companies using creative accounting to try and make themselves look profitable and then collapsing because they're not actually profitable. For example, WeWork's "community-adjusted EBIDTA" figured claimed the company was profitable using very similar arguments to Dario, and then the company went bankrupt. If you're already cooking the numbers, you have almost arbitrary flexibility to report whatever "margins" you want by excluding some of your costs from the calculation.

The TL;DR is that Dario likes to talk about imaginary/hypothetical companies a lot in interviews, and those companies' financials don't have a direct basis in reality.

It never really took off. I think because computers already require users to read and type Latin letters in lots of other situations, and it's not that hard to learn what a few keywords mean, so you might as well stick with the English keywords everyone else is using.