15-25 kWh drawn by the utility is quite a bit! I hope they're paying him well for that.

They're bringing outside scrutiny into their processes; that's going to slow them down, too.

Not "at the same time". You're conflating what AI can do now with what it will likely be able to do in a year or five, or 20.

Amodei and others are worried about what AI that doesn't yet exist may do, not about what current-generation AI can do. AI is getting better, very quickly, and unless there's some blocking obstacle that no one has been able to identify will at some point become far, far smarter than humans, individually or collectively. So before that happens, we need to figure out how to make sure that superintelligent AI wants to help us, and wants to do it in the right way.

You're not supposed to sell it. (Does anyone sell euros?) You're supposed to spend it. But hardly anyone takes it.

at some point most people want to sell the Bitcoin for dollars or pounds or euros

The fact that this is even significant (and it is) 15-20 years after the introduction of Bitcoin, shows that it has been a complete, miserable failure.

I think Bitcoin was a neat idea. Shame it didn't work.

Why would they need a warrant? Flock consents to their customers accessing Flock's data.

[various anticipated objections] Yes, but this is in America.

So many Hollywood scriptwriters who lazily decided, "I'll just use Nazis for the bad guys, they're universally-recognized evergreen enemies," are facepalming these days.

It is possible to put laws into place that require how public-private partnerships work. It would be an imposition upon Flock and upon the police. And that is just fine, and as it should be.

Law enforcement contracting to the private sector should not be allowed to be a bypass of our rights.

The only place I see Al in the legal profession making sense is to do basic caselaw research, where it can cite the sources and the references upstream and downstream to those sources. The lawyer should be responsible to review the claimed caselaw citations to determine the merit of their support for the argument that the lawyer is making, and to pick the particular citations to use that are the strongest for the particular case.

Likewise the same sort of search but inverted, for one's opponent's most likely arguments or rebuttals might be performed, again with the lawyer doing the review based on the citations that the Al tool provided.

There was a television series in the 2010s called Person of Interest where the premise was that one of the main characters had managed to achieve Al, but in order to satisfy legal constraints, the Al was only allowed to spit-out a single phrase about a person, rather than any detail of any sort. It was up to the limited number of persons associated with the Al to figure out why this 'person of interest' had been referenced, whether they were the perpetrator, the victim, a witness, or someone who would suffer downstream effects. While soft science fiction, and pretty clearly taking inspiration from the older series Quantum Leap in having to figure out why they were there, the concept had some merit as to how Al should really be used. It should be treated as untrusted, it should be treated as requiring thorough review. It should be treated as another tool available, but no more trustable or authoritative than any other, and arguably a bit less trustable, a bit less authoritative.

I've seen firsthand in meetings people saying, "Well Claude says..." or "Well ChatGPT says..."

What's bothering me about this is that those running the meeting aren't shutting that crap down immediately.

Except that in basically all of the examples that you cite, the transition was from one proven technology slowly into another technology that was at least approaching some degree of maturity by the time it it mass-market adoption.

In most of those examples. early adopters were nearly all ultrawealthy who were using it for themselves. Those who weren't incredibly rich were technology-enthusiasts of some fashion or another. Those groups initially worked-out most of the showstopper-problems with the technology before widespread adoption, and in many of those industries either compatibility or outright law added extra constraint over time.

AI for the masses has taken a different development path. It's not ready and is misbehaving, to the detriment of many. "Hallucinations" should not be tolerated at this scale of deployment.

Not if you're the police.

Or as it was put nearly two millennia ago, "Quis custodiet ipsos custodes?"

This is why systems like this require human oversight. Arguably if there's a judicial review process for allowing searches, then this system needs to be required to use that process. With a system containing precinct, local court district, municipal, county, state, and federal district jurisdictions, the nature of setting up proper permissions alone should be a mandatory herculean effort, probably the most difficult part of the entire system.

Without such a system requiring judicial oversight to confirm that searches are authorized, all we have is a further slide into a police-state.

(I'm just speculating.) People don't need help with selling fun things; they want a 'bot that can sell the crap which humans have given up on. Anthropic doesn't want to sell widgets; they want to sell stores which sell widgets (sort of like selling pickaxes instead of getting into mining).

Maybe the problem it has been asked to solve is "I have some crap that nobody wants. Sell it."