danparsonson

Also less recently Asianometry: https://youtu.be/_xPhxtuA_Qc

You understand there's a difference between how the law is, and how you think it should be, right? Only one of those things will actually help you in court.

To add to the other replies, when it's an argument, it's pronounced like "how" not like "no".

Yes - that specific point was not about this situation but a pattern of behaviour.

I don't want to split hairs but I posit there is a difference because 'how I want you to think about things' could be a mixture of lies, truths, and half-truths.

'Lying', to me, implies some relationship with reality - I'm lying if I know there's no orange in my bag but I tell you that there is. What we're talking about is someone who might not know or care whether the orange or even the bag exists at all, and is just saying things to get some specific response out of the audience. The deception or not is irrelevant really.

That's how I shall think of it from now on ^^

Ah, got it. Maybe 'gaslighting' cuts more to the point?

For certain values of 'admired'... It is impressive, in a diabolical way, and seems to be very effective.

> This was either a) a plain lie b) they're actually at their last resort.

That's thinking like a normal honest human :-) My point is that it was likely not a statement about reality (true or false) at all, but rather a phrase designed to elicit some response in the listener, such as the idea: 'Sam Altman isn't the kind of CEO who would put ads in his products unless he really had to'.

He's not describing how things are, but how he wants you to think about them.

No, I suspect that "I kind of think of ads as a last resort" was doublespeak for "ads are coming eventually".

I would tend to think of someone like him as a person who uses words to achieve a specific goal, rather than someone who speaks whatever is truly on their mind. Whether those words are lies or truth or somewhere in between is irrelevant; what matters to them is the outcome.

It's likely a waste of time trying to unpick the meaning, because there is none. "But Sam Altman said..." to me has about as much value as "ChatGPT told me...".

The robots use it a lot because it's a common construct in their training data, because it's a common construct in text written by humans

Yeah the battery life is off-putting for me - watches I had as a kid would last for years on a single battery so I think I have that expectation stuck in my mind. I do like the biometrics angle - something that tracked my health condition and alerted me to any problems would be interesting, but it would have to be fully under my control; I'm not keen on my health data disappearing into who knows what databae.

Yeah fitness tracker is a good one; I'm the opposite with my phone though - I get a ton of utility out of mine, but not things that I think would be improved by shrinking them down onto my wrist.

Alarms I can understand, but that's been a feature of standard watches almost since they were invented I think. I'm looking for something that requires a programmable device to be meaningful and so far the ideas are fairly thin on the ground :-) I get the techy appeal though, and agree wholeheartedly about the value of having total control over a device if possible.

Thanks, I'll take a look!

Yeah, we used to have MP3 players for that - I still have my Sansa Clip+ in a box somewhere...

Yes, you and I are making the same point :-) There's lots of useful advice out there about how to be a better conversationalist but it's exhausting for those of us who have to constantly think about it, and disheartening when we get it wrong despite all the effort.

I've loved the idea of a computer on my wrist since I was a kid but now they're available, I find myself unable to think of an actual useful application for them apart from telling the time - what do you smart watch people actually do with them? Are they mainly toys or do they enrich your life in some practical way?

Nearly 50% of US voters voted for the current administration last time - do you think all of them are on board with the Iran war? There are multiple reasons for a person to vote for a political party; China is a big issue in Taiwan but it's not the only issue.

The kind of reaction described by the GP is probably trained by a lifetime of bad experiences. One can end up going into every interaction thinking about which parts of oneself to dial down in order to have some semblance of a normal conversation, and inevitably that over-thinking just makes it worse. Ask leading questions, smile, listen careful, don't interrupt - you know, all that sort of thing that comes more naturally to some than to others.

Features arise out of the composition of fundamental units of the system, they're not normally first class units themselves. Can you give an example?

I've been using it recently and haven't noticed any stability issues at all so far.

I see, thanks. What do you use the history for? Is it just curiosity about your own evolving interests, or do you have some other use for it?

> ...scrobbling to MusicBrainz...

Forgive the out of touch questions, but, doing what? And why?

Haha dammit - I'll lose my membership at this rate

Thank you, I understand what fuzzing is; that test harness was presumably provided either by the blog post author or generated by Claude somehow, and therefore would not have been part of the proven code, nor part of the original claim by the Lean devs. That's what I meant by saying there is no program as such.

> The bounded, specific codebase that you refer to is typically the library and all of its dependencies, which in this case includes the Lean runtime.

How does that work? I thought the main idea is to write code in the Lean language which has some specific shape conducive to mathematical analysis, along with mathematical proofs that operate on that code. How then does a system like this handle a third party dependency? I've searched around and I can't find any information about how it works. I assumed that the boundary of the proof was the source code - surely they can't also be proving things like, say, DirectX?

> This is why formal verification is difficult: the proof chain needs to extend all the way down to the foundations.

The difficulty is not in explosions of computational complexity due to problems of incompleteness, decidability, the halting problem, those kinds of things? As I said this is not something I know much about but it's surprising to me if 'analysing all the code' is really the difficult bit.

Noted, thank you

To be clear, I think the article was fine and the author did some useful work (finding a bug in the runtime of a supposedly provably correct system is indeed a valuable contribution!). I don't agree that it's pedantic to explain why the title feels like a bait-and-switch (and thus does a disservice to the article itself). It's just a bit of feedback for future reference.

I take some comfort from being technically correct though; it's widely accepted by all of us pedants that that is the best kind of correct ;-)

Yes but, without wishing to be snarky, did you read the article? There is no program as such, in either sense - the announcement from Lean only mentions "a C compression library" (zlib). Not only that, but since we're talking about formal verification, a programmer would likely understand that that is about proving a bounded, specific codebase at source code level, and not operating on a binary along with its associated dependencies (again caveat my limited understanding of these things).

My feeling is that if you told the average non-technical user that some person/organisation had produced a formally verified version of a C compression library, you would likely get a blank look, so I think it's reasonable to assume that both Lean's intended audience, and the audience of the blog post linked here, correspond with number 1. in your list.

Further to my earlier reply - a more succinct way to look at it might be:

- When they fix the run time, bug A goes away. So the proof still holds and the zlib code is still correct.

- When they add a system of proofs for the parser and modify that, then bug B goes away. So the proof still holds and the zlib code is still correct; and now more of the library is proven correct.

The formulation of the title is "I was told X but that's not true"... but that's not true. You were told X, and X is true, but you found Y and Z which are also important.