autuni

I have considered getting into Vim a couple of times because it always seemed like the thing to do, but never did.

It does not make sense to me that I should use an editor I need to study and practice for. It can't possibly that much more efficient, especially given that usually speed is not the limiting factor for what I do.

Maybe one day I will be tempted to find out why it's supposedly worth it again.

great products only take them so far, the question is are they great enough so people would be willing to pay the prices they'd need to ask if they would want to at least match their spending with their income

that's not the issue, it's about trust.

People who are careful with AI code that isn't audited will not suddenly trust AI audited AI code

I think most people see the issue there with the "well audited" part. You didn't know which developers just copy pasted stack overflow code together before either, but the rate at which that code came in was manageable for reviews, and supplementing reviews with AI doesn't really lower the concern for people concerned about the quality of the code.

The alternative would be to rely on self-policing, which is a bad idea. Not even because you can't trust people to not be lazy, but simply because there's also actively bad actors involved.

If the code is well reviewed/audited, of course it doesn't really matter or not if it was hand-typed or generated, it's (at least for me) more of a question on how to manage quality control.

can't wait for the ad-fueled cheap tiers right when some games publish as streaming only

that's a fair point and that part is indeed interesting to a degree, and I would fully agree if it wouldn't be stated multiple times that the author did not actually read, i.e., properly engage with, any of the involved papers, but only used an LLM to talk about them. The author also emphasizes that they did not check the proof, or fully believe that the result is feasible. (Yes, I did in fact go back and read it)

> I have not checked most of the details, but I think that the claimed o(n log n) bound is plausible.

Stopped reading right after that. Why is this so high up on HN? It's literally nothing.

I'm not an expert, but in my mind if you actually want to make the human species interplanetary, flying small groups of people to the planet won't do anything for it. Send robots to build a base first, and yes that will take a long time. Otherwise you'll just be a species that visited other planets. (which would already be cool though obviously)

I'm quite excited by planetary defense missions. While I don't think that anyone has any reason to be fearful of the constant danger of being wiped out by an asteroid, in my mind it marks a big milestone for us as a species to be able to (in not so distant future) protect ourselves from something that wiped out dinosaurs. Just feels like a mission that is not driven by global power struggles/projections, but is just a positive for everyone.

It seems weird to think about how much more developed we are as a species, and yet so far we are still vulnerable to the same incident that ended their reign over this planet, completely out of their (and our) control.

Forecasting is just trying to identify recurring and long-time trends in data, and you're absolutely right that with a narrow observation window it is more difficult to do so, however if you have enough historical data (i.e. several years), these models are remarkably good at picking up on patterns leading up to e.g. bank holidays, the information is contained in the data.

That being said, more information is of course better, assuming it doesn't introduce bias, like focusing on specific events like holidays too much, when the information is already contained in the historical data, without having to make assumptions about what an upcoming holiday means.

The biggest challenge last time I worked with time series forecasting (which is ~5 years ago) was learning long-term shifts in the baseline of the data (e.g., the general use of electricity always goes up, a lot more solar panels and EVs influence the baseline, etc.).

I can only assume (hope) that there has been some substantial progress in that area though.

(Edited the formatting for a nicer reading experience)

I don't think you realize how little tax some of the richest individuals actually pay. They can dodge taxes and exploit loopholes in ways that aren't available to the "average" person. Of course the average person shouldn't try to dodge taxes, and neither should the ultra wealthy, because if they are applied in a fair way, taxes are good for the community. The problem is just that it doesn't work in a fair way in practice.

I agree this is more of a "jumping car" than anywhere near an actual "flying car". My main issue with both is that if ground based cars have an issue, for example an engine failure, in most cases you can still get the car to safely stop at the side of the road. If you're flying, that's a very different story.

Even if we would accept the premise that flying cars will be the norm at some point, I'm not sure what people imagine that would do for them. Technically you could solve traffic jams to some degree, but still not entirely because I can't imagine that at any point flying cars would mean that you just fly wherever you want. It would give us multilayered traffic, which is cool to reduce traffic jams, but does it really do anything else?

idk about the safety of involving jumping cars into car crashes, I'd still prefer the brakes (obviously not always enough time). The one useful situation I could see is if you're driving (accidentally I hope) off a cliff.

It's a tech demo marketing presentation with a jumping car in it, not a flying one. It does look pretty cool though. Let's first wait and see if/when this will even be available and not just a marketing stunt.

It's a cool demonstration, I'll give them that. From what I've seen online, people don't even seem to consider whether or not they would actually want this, just because the novelty of it. I'd even ignore that and call it good marketing, if that's what they would actually try to sell here - but you need some kind of special invitation access to even get those, while they use this demonstration to boost the pre-orders for a car without the features everyone is talking about (or at least trying to, we'll see how many will actually go for it).

To a degree I think it comes with the rational approach. If you want to be a rationalist, you need to consider many perspectives, and somehow bring in the nuance of as many of them as possible instead of ignoring the parts that don't fit your narrative. It's very easy to be brief with one-sided takes, but nuance is boring and lengthy. That of course doesn't mean that the authors conclusions and arguments will be correct in the end.

if you're thinking purely about game loops and game mechanics, yes a lot has been done before and used over and over again. but that's by no means all the programming going on.

just look at how much people complained about poorly optimized games in the past decade. the most impressive game dev feats in terms of programming were not the game mechanics, it was getting their game to run as well as it did on (very) limited hardware resources, a problem we are hitting again now. and it is very obvious looking at modern games that this problem is nowhere near solved, and in fact very hard.

I don't think the rationale makes sense. If you want to provide a free resource why not just provide it for free for everyone, considering it doesn't even cost anything because downloads/online lookups are not a limited resource. Not everything has to be in "competition" with other countries. The whole mindset that if it doesn't give enough of an advantage to one countries people over everyone else means it shouldn't be done at all is very weird.

What stopped you from getting into the industry? Was it just the lack of opportunity? Unless I'm living under a rock I don't think there's too many companies in the industry in the first place.

They should have posted the ones they actually confirmed. They don't have any time pressure to release hundreds of new proofs, they could do 10 a week for all I care, as long as they make sure everything is actually sound and properly checked. You know, do the due diligence with the results. Not only would that prevent them from having to do retractions, it also wouldn't be any less impressive. Just not as much shock marketing value.

It's really not that hard to understand. My issue is NOT with the fact that they publish results. It's purely about how it's done and what the misleading claims are that come along with them.

You may want to read beyond the first part of a sentence. Yes, preprints have always been a thing. The claims and publicity around the potential findings are handled differently here. That's where the issue is.

this is not entirely related to the tweet but to the topic in general, this prompted me to check their repo again and saw this:

> The vast majority of results were obtained with the same procedure using an unreleased internal OpenAI model. On average, each result used three hours of ChatGPT Pro thinking compute with that model. Over the course of the evaluation, the model was posed approximately 4,000 problems. Aggregating the output into result families and manuscripts and requiring an appropriate level of significance led to the catalog outlined above.

seeing the full list of problems would be the most interesting part of this whole situation. it could give some insights into what kind of attributes of problems cause issues / are easy to solve for LLMs. (edit: they posted results for ~700 of the 4000)

The difference is that human mathematicians wouldn't post it online, claim they have achieved some proof, and then check after publication and announcing the results to the world. You would always check your work first, then publish it. It's not about preprint vs peer-reviewed. That of course is normal practice, it's the high-profile claims that are being made that are the problem here. They just blindly published results produced by the LLM, with 0 due diligence.

No, in their original blog they wrote:

> As part of our GitHub repository, we are sharing formalizations of many of the proofs in Lean, a programming language that allows mathematical proofs to be checked by a computer. We will update the repository with more formalizations as we obtain them.

Meaning they published all results before checking all of them, and intended to add more Lean proofs later. In the linked post they state ~42% of the posted results now have formalized proofs, some were added, some verified, and I assume this means that some results turned out to be wrong.

I think your second bucket would need to be changed, as it's not really undecided for some and Tao in fact argues (in a way) that there is no need for the field to decide. There is a need for results, but there is also a need for humans to fully understand why the results are the way they are and how they were obtained. This doesn't replace the human in the loop it's more of a cooperative effort.

I'd end up with the buckets

a) Fully human

b) Hybrid human-AI

c) Fully AI

Neither the US nor Europe has such a unified housing market that this could be a reliable statement. In Europe it differs greatly within countries, and there are many countries. In the US it also varies greatly within each state and there are many states.

Also at this point both have seen such negative upward trends in housing prices that the assumption that someone placed in the middle class owns a home at all (without any debt attached) is questionable. (But that's just a comment and besides the point)

It feels odd to me that they wouldn't prefer applied problems. Seems like an easy way to profitability. Probably based on what attributes they're looking for in a problem when picking them.

I think right now it should definitely be a race to profitability, because revenue will not save you from unsustainable finances, no matter how large the values are.

I think OpenAI is currently far enough ahead that they could just cut training and R&D for a while (maybe even 1-2 years) and rake in some money before they'd need to start again, but if a smaller company finds a way to keep doing R&D while also being profitable, it doesn't matter if they're a little behind in quality, they will win in the long term.

But does that change the training requirements specifically? You're right of course for inference, unless I understood that part wrong about active parameters (which is entirely possible).

> Simultaneously, Europeans are notoriously arrogant. It's a mockable combination.

Funnily enough, you can just change European to American here and that's how many Europeans view Americans. Mostly in terms of the wealth distribution though instead of the overall wealth, and the fact that Americans just blindly claim they're the best in everything without hesitation.

Maybe we should focus on huge, exploitative companies that suck out the wealth from everyday Europeans and Americans instead.