Jeff_Brown

Someone in these comments said the paper on Barnette's Conjecture is short. Are any other proofs in this collection short and/or understandable?

Yeah the four samples I heard all sounded kind of formulaic. Too obligated to continue the mold of the last thirty seconds. Impressive and quite pleasant but no big ideas and no thrill of flight.

Could this relate to continual learning? It lets you update without pausing the entire system.

The burning question I can't get any information nn is whether, if they determined an earlier misaligned generation may have transmitted misalignment to the current models, they would roll back to a safe checkpoint to rebuild from there. I suspect they would not unless forced to.

Please elaborate!

The HF hack was a felony.

By itself, "aha" carries no insight, but the insight is probably stated immediately after it. In that case the aha is semantically useful, by identifying the insight it is near.

A question not addressed here is whether there exists any solution to the EU requirement that doesn't suck.

If the same red/green algorithm described in the article is applied to generated code, I cannot imagine how that does not degrade code quality (probabilistically, not at every point).

I love it. I wish I could select materials.

Yes -- power without trust is of no use.

One of the Voyager probes measured the density of the interstellar vacuum at 80,000 protons (and the same number of electrons) per cubic meter. A proton going through a piece of aluminum foil delivers a roughly constant amount of energy regardless of speed; a relativistic proton will pinch through and carry most of its energy with it.

(No punchline; I just think that's cool. I understand that the real problem is the rare dust grain, not the ubiquitous gas.)

I suspect load-bearing is a euphemism for 'not garbage'. Ad in 'most of what you said I can mostly ignore'.

This reminds me of when a friend saw that someone had Liked something he had posted online, and he wanted to Like the Like. There was no interface to do that, so he was "reduced" to merely liking it in reality.

The idea that a producer is at fault and not also the consumer paying them to do that is strange to me.

Can anyone find (or draw) a picture of the construction?

It's really not clear. We might all become unemployable. But as coders become more powerful, they can do more, which makes them more valuable, if they or the businesses empluying them can invent work to do.

If all we can do is compete for the same fixed amount of work, though, it does look bleak.

What an alien preference ordering.

It might be a bad idea to put that in all caps, because in the training data, angry conversations are less productive. (I do the same thing, just in lowercase.)

As a consequentialist who shares the author's concerns, I feel fine (ethically) using AI without advancing it. Foregoing opportunities meaningful to yourself for deontological reasons when it won't have any impact on society is pointless.

This might sound snarky but in all earnestness, try talking to an AI about your experience using it.

They're not saying the election results are in. They're saying Orban's style of rule has been a disaster for Hungary.

I love 3B1B but generally don't have time to watch long videos. Can anyone sum up the punchline?

Yes, without a good experiment (maybe a natural one [1]) we can't know. Even if the study controls for everything observable, there may be unobserved differences that lead to the caffeination difference. For instance, even though two people might have the same job, education, etc. the one who is more ambitious, or creative, or hopeful, or simply healthy enough to feel like working more, might drink more coffee.

https://en.wikipedia.org/wiki/Natural_experiment?wprov=sfla1

Dogfoodung usefully connotes getting off your high horse, getting dirty, getting your face in it. I think it's perfect.

As someone who has received death threats, I can tell you, the comfort from the fact that they're usually not acted on, while real, is not huge.

But does it have network access, or access to other apps (perhaps via shared storage)? There's no reason for it to have any of that, and there exist systems for tracking whether an app can do those things.

When I go to enable it I get 'may be able to collect all the data you type, including ... passwords'.

'may be able to'? How is this not knowable? Do I have to wait for effect systems to gain popularity before installations make sense?

This feels like the most important comment here. Do other Mastodon users feel the same? The OP Madden me want to try Mastodon.

You sound offended. Not my intent. It is linguistically difficult to avoid connotations of intelligence when describing artificial intelligence. What term instead of 'judgment' would you prefer for determining whether a user's request is ethical?

Something I love about emacs is the ability to tab complete the name of a command. I do know a lot of keyboard shortcuts, but I use way, way more commands than I know the shortcut for. Need to rename a buffer? M-x ren-buf TAB should do it. Etc.