"...do you discover who's been swimming naked."

AI mostly comes to us from California, the original land of "Fake it till you make it". AI fakes intelligent results, without "knowing" anything, with what Gary Marcus calls "a world model". A world model means you "know", that you "understand", that you "comprehend" that:

-- the cited-cases part of a legal filing should be only drawn from the list of existing cases that the thing was given
-- that laws mentioned should exist in documents called "Title Code" that the thing was given
-- that there are laws, these are supreme "orders" above others given
-- that one of them is "killing human beings is against supreme orders that supercede all other instructions" (sorry, not on topic: it's just HAL didn't "know" the 3 laws)
-- that human hands have just the five fingers.

They just fake it, producing documents that look like good ones. They can't make it. Not in them. No world model.

And the tide is going out.

> You can hardly blame Google for that

You can, actually, blame Google for implementing an unnecessary code signing requirement that only works if Google is allowed to do business in your country.

Also, well done Google! You've just given millions of people entirely legitimate reasons to find workarounds your attempt to lock up Android, and entire Nation States a legitimate excuse to hack your operating system, or even ban it and support free forks.

It's not illegal, but it will almost always result in a mistrial and a referral to the Bar if done by a lawyer.

Sometimes the best thing to do is make popcorn and enjoy the show...

I love the fact that this article claims the specification fixes the "main barrier to Enterprise Adoption (of AI slop)" right after an article saying basically AI will never be secure and you need to stop pretending it ever will be.

So it's unreliable, removes business knowledge from the employees of a company and puts it in the hands of a hostile third party, and now it turns out it's insecure, but adoption of the slobtasystem is only being held back by some specification? Really?

Name an alternative to a spreadsheet that's simple to use and doesn't require an overworked and usually unenthusiastic IT department to set up?

The issue here isn't that they used a spreadsheet, it's that spreadsheets are still treated as toys despite being the exact opposite.

How is a couple years of apprenticeship where you are paid any worse than 4 years of college plus debt?

Maybe a second responsible party should examine everything before something is published? This is how it works in many places.

> No one doing anything complex with software is writing their own code anymore

You are not everyone.

The fact you're lazy and do not care about the quality of your code or its future maintainability doesn't not mean anything unless you're reflective of the entire industry. And you aren't.

> The problem is that LLM-generated code has no copyright. Anybody is free to copy it without any limits whatsoever. That does mean that the GPL can potentially go out the window if too much code is LLM-generated or LLM-derived.

It is far too early to make this statement. It is sort of true in the US as a judge or two have ruled something like this, but copyright is a world wide issue, with a patchwork of completely different laws. It may not be protected by copyright in the US, but it might conceivably violate the copyrights of every single person whose works were loaded into the LLM without their consent in Europe, for example.

And honestly, I think that would be right.

That's a question, but an even thornier question is whether LLM code is itself a copyright violation. Europe and the US appear to be moving, slowly, in opposite directions on this, so there's a risk that using LLM code will render your entire project copyrighted by someone else in various jurisdictions if they can show that the LLM you used was built using their code.

For example, CoPilot, IIRC, is trained on the entirety of Github. If someone finds their code on Github, they may co-own your CoPilot-authored project in Europe, at minimum.

But LLMs are designed to use attractive, dopamine maximizing, phrases and workflows, so people use it anyway in the same way they gamble away their kid's college funds at a Las Vegas casino. We're so fucked.

I don't understand your problem with the way apprenticeship is done. It's always been like this! Even in IT.

This "AI" seems to be severely lacking in training, in that no superior/invincible force (e.g. a party simulating law enforcement) showed up and demonstrated the downside to taking certain actions.

We should but only in parts of the world where car usage isn't mandatory. US suburbs, for example, are not built for or served by anything approaching adequate transit, nor are they even walkable. Those areas not only should not need repeated testing, but you shouldn't even need a license to begin with.

And if you think that's a really terrible idea, well, maybe you need to support more walkable communities with adequate transit, even if you, yourself, wouldn't use any of it.

Also note I'm not proposing not punishing people for making driving errors. Just... suggesting that banning people from driving when that's actually practically a death sentence isn't appropriate. We used to shoot horse thieves because we believed in freedom of movement, and knew the consequences of taking someone's sole form of transportation. Somehow we don't now.

The ease with which a license should be obtained and kept should be proportional to how mandatory driving is.

Dammit, can't anyone make some stylish cyber-shoes? That looks like shit! Anyone who goes around upskirt-snapping with shoes like that, is going to get laughed out of the ballroom by all the snobs.