And reading comprehension is not your strong suit.

My point was that the other part of your comment, "They just regurgitate random stuff they found on the Internet", is clearly false.

Makes me feel like maybe, sometimes, you can't trust what is on teh internet tubes.

"Lots" and "one that is widely accepted" are not the same thing at all, which is my point. There's no point in discussing intelligence without first nailing down which definition the interlocutors are using.

Also, nearly all of the definitions are extremely fuzzy.

Indeed.

The claim you made was not about whether machines can currently perform like a brain, nor about complexity. The claim you made is that the human brain cannot be simulated by a Turing machine, which is a much, much stronger claim.

So I repeat, why do you think that?

38 of the 50 US states are "one-party consent" states, which means that as long as one person present (e.g. the person wearing the glasses) is aware of the recording, it's legal. Roughly half of countries around the world either allow any party to a conversation to record it without telling the rest, or don't have any restrictions at all.

Since there's no clear definition of what "intelligent" means, much less one that is widely accepted, that's a pointless argument.

You clearly have not used them much. Try using an LLM (one of the top commercial models, e.g. Claude Opus 4.7) to debug code that has never been on the Internet. I don't think anyone knows what "intelligent" means, so arguing that point is a waste of time, but LLMs clearly can observe results, reason about them, form hypotheses, devise ways to test those hypotheses, perform the tests, evaluate and reason about the results, etc.

My guess is that the face blurring happens on the car, before the data is even uploaded to Waymo, as a way to protect the privacy of random passersby from Waymo and Waymo employees. They likely never had possession of imagery with unblurred exterior faces, so there was no way they could provide it.

There's no way that Waymo is just defying a court order.

Why do you think that?

That's a fantastic idea. Thanks!

"And itâ(TM)s probably easier to hike in a gallon or three of gas than the equivalent electricity. But having back up solar panels could solve that in some situations with an EV too."
Grok says about 2 days to charge up 10 miles of range from a 4m2 reasonably portable solar panel in decent days at 45n latitude.
Yeah, I think I'd rather just fetch some gas.

Thanks for the update, I was unaware of the project at all.
Fascinating stuff, I wish them the best & sincerely hope the Canadian government-as-it's-going doesn't see this as a potential source of funds that can be drained off for more norcan.

As he already said, haters gonna hate.
Despite being variously a know-nothing alt-right trans-denying fascist stupid head, he's worth a $trillion and running several successful businesses that individually would be massive, groundbreaking success stories.

That's gotta just burn, doesn't it?

SMRs that utilize 1950's designs and require intense operational oversight and maintenance could never be economical. SMRs that use better, safer designs that can safely operate with no active operational oversight and little or no maintenance, which can function without intervention and without refueling for 20-30 years, then be inexpensively disposed-of and replaced, and which can be manufactured in large quantities to bring the unit cost down, promise to be very economical.

Will the new designs actually achieve all of those goals? On paper it looks good. Whether that theory will translate into practice is something that can only be discovered by trying.

The risks are a lot smaller than you think they are, because of new reactor design. Nearly all of the nuclear reactors in the world are still using a design that's 70 years old, that requires active cooling and doesn't fail safely. We have much better designs now, at least on paper, designs that simply can't melt down, whose failure mode is to simply stop. But no one builds these new designs on industrial scale because they're unproven, and there hasn't been much funding for doing all of the engineering and research needed to develop them into fully-functioning designs that can be.

I'm skeptical that small reactors are really the best way to actually deploy nuclear power on a large scale, because of security concerns, but starting small is the best way to validate and refine new designs. And modularity is clearly a good strategy for making deployments of varying sizes cost-effective. If you can develop a cost-effective module that can be manufactured in large numbers, you can build large plants by clustering them.

The new designs shouldn't actually need much operational oversight -- if something goes operationally wrong, they just stop functioning -- but they'll still have highly radioactive cores which, if extracted, could be pretty terrible weapons. Not to make nuclear bombs, but to greatly enhance the damage done by conventional explosives, by adding radiation hazards that linger for years. So, security will remain an important consideration, and the SMRs should only be deployed where security can be assured, which will in practice mean that most are deployed in large clusters.

This all assumes that the safety, effectiveness and cost-effectiveness of the new designs proves out, of course. The only way to find out whether that will be the case is to try.