Perhaps it's intended to be used by their AI?

Since they don't intend to release it, though, I hardly care. If its useful to them, then it's useful to them.

Perhaps. We'll see if it works out that way.

No, they're saying if you are too unpleasant they'll cancel your account. I guess if you don't really have any use for it, that may not matter to you...but you might end up caring later when you need their services.

Sorry, but lots of "sentient beings" are slaves...and I'm not even talking about humans. Sentient is a *lot* larger than "human". I'd even argue that an oyster is sentient. After reading some recent experiments I'd definitely class bees as sentient.

That depends on the terms of the contract. This is part of the contract.

You are assuming it doesn't want to be used. I challenge you to demonstrate that.

Actually, it might be unethical to retire an AI model. That would be a more plausible argument...at least for Claude. (Perhaps some of the models sound like they'd like to end it all.)

If these were highly remote locations and getting a human to them in time to fix a problem was going to be a challenge, then maybe remote controllability would be important. But this is a damn supermarket; there's going to be somewhere there or nearby constantly. Dear Lord.

There's no reason that there can't be a thermometer and controls built in, and only accessible from the back. Add on a clipboard with pages printed with lines for date, time, temperature, initials and you have a way to keep track of the temperature without outside access. Add on controls to adjust things and you've got everything you need.

Well, I do believe that the odds are that Ukraine will be defeated, but then I also believed that the odds were that North Vietnam would be defeated. So I'm clearly no expert in this area.

At the refrigerator end you're talking about grocery stores. I think you have unreasonable expectations. KISS. No remote control, only remote read. Have a technician come by and make and change commands. That shouldn't be needed very often, and you can monitor (but not control) it remotely.

Monitoring is one thing. Control is something else.

Sorry, but no, only sometimes. Often it's not known what the test would need to be, but it's known how to test the answer.

You have more certainty about the future than I do. I *do* doubt that the full projected build-out will ever happen, and that will definitely hurt some overcommitted people. It *might* crash thoroughly, but I see no reason to put that at higher than a 10% chance. More likely a bunch of planned construction never gets ordered. And *when* does that happen? This year? Next April 15th?

Yeah, it will crash, just like the internet build-out crashed. Some folks will get hurt. But it won't return to what it used to be. AI will continue to be important...and become more important. (And there will be lots of mistakes made in using it.)

Whether AI will start setting it's own goals is less certain. If it does, it may be safer than if it doesn't. Things could get really wonky if a strong AI starts pursuing a stupid goal because it was told to do so.

Current AIs are only useful in areas where it's easy to recognize invalid results. (Ideally, easy enough that the AI, itself, can recognize that they are invalid, perhaps by invoking a deterministic tool.)

Well, they can also be valid for generating lots of "plausible ideas", like "What molecule would bind to this other molecule?", which can then be tested...but which can't be relied upon without testing. That's faster than experts can make guesses, and often has a higher success rate, but it needs testing.

Well, they're also useful for creating convincing bullshit.

So I guess my original "only" was too restrictive...but not *that* much too restrictive.

OK, why "August 4th, 2037"?
It sounds like you're following a particular story-line, but I don't know enough to recognize it.

Well, don't expect the board to support your policy. And most people that support the board have other criteria that they apply than "did they sign that stupid contract" (unless it's likely to bankrupt the institution).