Based on the description it also includes images and maybe video. So deepfake porn of people without their consent, and without adequate regard of age.

Yes, they toss some stuff into system prompt to 'promise to be a good boy', but as an *enforcement* strategy, that's been demonstrably a poor mechanism that gets worse with nuance.

Sure, it's about making sure they aren't getting *too* much screen time and *not at all* about trying to audit that they are doing as much screen time as the managers expect them to be getting..

Funny that they list 'passkeys' as a proof of human. Peel it back and a passkey is like an ssh keypair. They *could* try to employ attestation to limit to 'blessed passkey vendors', but it's going to be a tough scenario at all.

If folks are determined to 'bot' it up, a pretty legitimate passkey can be part of that. It was never designed to serve the purpose of proving 'human' interaction.

Blackwater would like a word...

Of course you also just described most corporations too...

Question is do they *not* have those? Looked up a couple and they seemed to be equipped on that front...

I think the expense is more like the long standing manufacturers accustomed to charging a boat load of money for a resistive heat loop in the seat. Then optimizing it so they *always* ship that loop, but enable/disable it to still charge a boat load of money for it. Then when they realize they can turn it on/off at will, then they want to charge *monthly* for it. All for a resistive heating setup that they deem really so cheap as to put in all the cars, but they want the 'premium' pricing to continue.

Rinse and repeat for a great deal of stuff in cars they do to drive massive margins. For the low margin stuff, they just make it disproportionately worse than they need to for cost, almost to punish the entry level buyer so they'll know better if they ever get enough money to pay for the premium stuff.

This is consistent with what I've heard second hand, that in Meta they don't really have any vision so instead they are just telling as many people to vaguely 'do stuff' with it as much as possible, in hopes that someone lucks into a hook for Meta to actually "get in the game" in a way similar to OpenAI, Anthropic, Google, or even Microsoft have found an "in".

There's no particular actionable ideas, so throw everyone in random directions and hope you end up owning someone's "hobby" effort that catches on in an unexpected way.

For simplicity, we say it "requires" a Microsoft account because the UI does not give you an option unless you keep track of today's "trick" to skip the Microsoft account. And this "magic" has evolved over time, so it's not even that "an advanced user is given a clear supported way to bypass", it's that the loopholes shift to make sure this remains a pain to make people give up and just make the damn account already.

It's plainly stated, this is a business direction. Microsoft wants to use your Windows interaction as an upsell and subscription opportunity. The revenue associated with the license itself barely matters.

I doubt that AI will particularly drive changes, since the whole point would be that the LLMs don't care and can work with the languages as-is.

The languages evolve and go away because of human motivations. Maybe a toolchain had a proprietary lock and associated with a business that went south. Maybe the language was a bit overly verbose and people didn't feel like dealing with it. Sometimes it's more like fashion than reason with how they go unpopular and back, maybe because some rabid fan implements a massively useful framework or something.

But LLMs don't have preferences to really assert. To the extent that you have any LLM consideration, it's perhaps to be terse for sake of preserving tokens, but comprehensive library function goes a lot further than syntax concerns to help that, and I don't think the actual language output is a huge portion of the expended tokens anyway.

A theme in the LLM dealing with code front, a *whole* lot of people who never got into coding saying a whole lot about the matter of LLM dealing with code...

To some extent, I get it, a manager who was always intimidated by code manages to get a little project to come out almost like he wanted without the manager actually knowing how to code, and that's exciting. But they always start pontificating on what their success means on projects that they themselves can't make happen even with an LLM, but imagine that whatever it does for them, it should work even more wonders for the coders, or else that if it can do a basic project now, then within a couple of months it'll just do whatever they imagine.

Why bother? LLMs doing the job they are supposed to works just fine targeting a human readable language, what do they imagine the gains to be? The whole point is that the LLMs deal in human friendly material.

This seems to be a pointlessly 'futurist' ambition for showing enthusiasm for LLMs rather than some goal with practical implications.

Even when I see someone having reasonably 'good' luck with their LLM usage, they end up with something a little wrong that, if they would just go manual for a second would be a super quick tweak. But they are committed to the prompt interaction, they keep trying back and forth with prompting and not getting exactly what they want, or something else changing at the same time. Leaving it human readable means you can do that quick manual change without trying to try to induce an LLM to make a precise change, which LLMs even at their best are prone to not manage.

I think he suggests how they operate and perhaps their resultant product quality, not their relative performance business wise.

There were the accounting violations before, and now we see that a significant chunk of that revenue was allegedly on the back of ultimately illegal activity.

They tend to play fast and loose with various facets of running their business compared to others, and it shows in their quality, which isn't exactly great.

However, they do tend to come in much cheaper, and if you deem 'white box' type systems adequate, they are the only ostensibly American company to be found in that game.

Sure, that's one thing, but the impression I got here and in a few examples in my personal life is people think they can resell the use of an LLM to get money for a service they didn't otherwise know how to do.

Specifically, for example, someone saying they could now sell software to people despite not knowing how to code by just typing the prospective customer requirement into an LLM and sending back the output and making money. It doesn't really work that way, but there are people who seem to simultaneously believe it can and yet also believe there is a business opportunity to sit between the customer and a chat prompt without any additional expertise..

I am shocked at how many people think they can resell the services of an LLM agent as a side hustle. If it works out, then there's zero reason to go to *you* for the same LLM they can just use themselves. If it *doesn't* work, well, you've got problems.