That's apparently already changing in the EU, where software vulnerabilities mean the company is liable for damages. The only way out is to straight up not make any money (not just from direct sales) from the software.

That is a misrepresentation. You are obligated to actually put effort into securing your products, which is the only sensible stance to take.

How so?

I'm thrilled that companies are liable for crap that ends up hurting other people. I don't think they should get an easy way out, and I also like that there's a carve out for people who aren't making money off of software (like OSS devs.)

Is the burden of proof on me, the developer? Do I need to prove in perpetuity that I didn’t get a job or a free flight to talk at a conference because of my free software? (Which had a flaw that hurt someone)

I have no clue.

But I do think that this is a much better start than letting companies ignore the impact to software consumers or having open source devs be on the hook for volunteer work.

Only referencing America, but professional liability for doctors, engineers, lawyers, etc isn't based on perfection. It's based on a reasonable effort.

So Google could, for example, switch from a tiny "this could be wrong!" byline to having the AI be less overconfident every freaking time regardless of whether it's spouting made up crap or actual facts.

The scale doesn't sound like a way out. If your company expects to get away with doing the wrong thing where smaller companies can't, then the solution isn't to continue getting away with it.

Are "your guys" a guerrilla force or something?

Because the military doesn't give soldiers rifles with guard rails. They give the soldiers intense, rigid training, and then try to enforce discipline and correct use socially.

If an LLM is going to be important in that way (this seems like a very contrived way,) then it's in the interest of the LLM's host to make sure it doesn't have guard rails that would get in the way _that_ way.

The whole thing stemmed precisely because of how they wanted to use Claude, and Anthropic was uncomfortable with it. Which to me screams that the models guard rails shouldn't be applicable to military use, or the outcome could wind up problematic, as we integrate AI more into military use, it sounds absurd now, but I will not be surprised if it starts being used in unexpected ways where a model needs to be fully unlocked from any sort of guardrails outside of guardrails that prevent it from imploding its own systems.

There are simpler ways to disrupt a flight.

Yeah. You should have seen the line to the bathroom when I named my WiFi hotspot "Free mile high club - meet me in the bathroom".

Are there? Setting a device name might be the lowest effort thing I can think of.

Requires you to be on the plane.

Just call the police and say you have a bomb planted on flight XYZ and want 100000$ or you'll detonate it.

The vast majority of people aren't aware of open versus closed protocols. If enough people they want to communicate with are using it to counterbalance how frustrating it is, they'll use it. It happened because businesses realized there's profit in lock in, and they threw resources at it.

Open protocols are still there and still used, but we're sad because the smaller userbase is frustrating. Just like how people still publish human written content to personal blogs, but they're proportionally non-existent.

That's the past.

Why does Google think it's a good idea to make that the case even if you don't block their crawlers?

A not thoroughly thought out response:

Those people would heavily incentivized to protect their ability to vote.

People in America die from preventable illnesses constantly because they cannot afford access to care but I guess they forgot to protect the ability to not die or whatever.

No, they're already being suppressed. They'll take the easiest action possible to ease the pain, which means voting for whoever does away with the fines.

If you sell something to someone and they do computer crimes, you're going to have to prove that you couldn't've known that they're a computer crimer.

It's the same thing with selling general offensive security tools. You have to proactively make it clear that it's for testing and not criminal use. Otherwise, cops are going to assume you're complicit and make things shitty.

I don't think that using LLMs for medicine is an appropriate fix for the US's healthcare issues.

Unless healthcare businesses decide to improve patient care with AI instead of increasing patients per day, I think it's going to make things even worse.

Doctors using AI will probably just increasing the number of patients they see. But for me as patient AI is super useful to get a good handle on the situation before I see a doctor.

I'm not suggesting it as a fix. I'm saying it's the only option to get medical answers for many people.

Calling people dogs by analogy is not great.

That aside, corporations and groups don't make decisions. People do. We can understand and empathize with what led them to that decision (and sometimes we might be looking at the wrong person), but they're still responsible.