mahogany

Can you explain more? This statement is already true before this policy change, so I don't understand why this is not a non sequitur. But I know this is a common rhetoric so I am trying to understand what the logic is.

No, it's not how math works at all. Math is not experimental science. Results take years to formulate and are generally known to be true by colleagues via correspondence and seminars before a paper is even written. Retraction is very rare.

I don't know why people are talking about something like this that they are clearly not familiar with.

It's so funny to see people on Hacker News talk about academic communities that they clearly are not part of. To be clear to anyone reading, no, retracting papers is _not_ standard in math. In math, people submit papers after they have checked with colleagues, they give seminars, they submit it for peer review, etc. It is not shotgunning papers and retracting incorrect ones on the regular. Retracting papers is rare and embarrassing.

Foot racing was never a broad source of income for humans. There is a massive difference. The question will be: what economic value can humans produce in the future with AI tools? There is a chance that the answer is "not much". That is the scary part.

The internet did not lead to a rapid diminishing of human economic value across the economy. Whether AI will kill us all is a distraction. Think more practically. Think about the future of economic value given a scenario where AI is capable of everything a human can do. Our entire society is built around economic value. Our individual survival and wellbeing is based on it. What happens when you are not needed by those who hold the resources?

Some people will quibble with the definition, but I agree with the spirit of this. I have pruned all my web traffic down to a few sites, and Hacker News and Reddit are the two that remain where I feel that pull for my attention. Hacker News is better than, say, TikTok, but I still find myself reaching for it to distract me from other things in my life. It is "easy". Like 24 hour news, the FOMO is still strong (you can always refresh and find some novelty), and reading a bunch of internet comments is a guaranteed cheap fix for dopamine (or whatever).

When I have experimented with banning these sites, I have felt a positive boost in my mental wellbeing, where I go back to being bored, which is great and leads to me doing more focused activities like reading more. I just don't know how to solve the "head in the sand" issue. This site has legitimately helped me in my career for example. I think I will land on a middle ground somewhere will be the ultimate answer, like catching up once per week or something.

I've only used the Claude TUI, but it is extremely sluggish. It takes 1-2 seconds minimum to launch on a MBP M4 with 48 GB of memory. I have also encountered plenty of display-related bugs that you can find documented all over the internet. If you had to group this into a software quality bucket, it certainly would not fall into "good". Maybe "mid".

Don't you need to run each model several times and take some average since their outputs can change for a fixed prompt? How do you test for that?

The thread is not really about what's legal; the topic is integrity. It sounds like, based on the fact that you wouldn't do it yourself, you agree that it's not a good thing to do.

Note that SET memory_limit is a soft limit and can be completely ignored for some tasks, so this wouldn’t be the same as having a limit on physical memory.

Taking Anthropic’s whole AI framing to its obvious logical end: if this is true as written, why was Jarred needed at all in this loop? It seems like an utter waste of time for a highly paid Anthropic employee. Can’t Anthropic have a top level agent that is looking at all interesting unsolved problems and orchestrating subagents via the same process?

The difference now is about the scale. What percentage of the labor force do you think were “computers”? It is not comparable to a potential disruption of huge swathes of the workforce in a short time span. That’s what people are worried about.

> No, I think people will demand to not live in poverty

What about other countries in the world where people... live in poverty?

> Capitalism allows people to escape poverty by personal effort

Isn't it possible that there is a future where AI makes the average value of human labor (or, "personal effort") plummet? Perhaps capitalism will lose some of its edge against a technology like this.

The comment is making fun of a Claude-ism where it becomes super “honest” about stuff.

It’s a joke but there is an underlying real effect where this type of language is psychologically manipulative and I would guess makes people believe LLMs output more than if it didn’t use “honest” (or “load bearing” or whatever super serious important sounding word).

You may be right but it seems like a different type of tool when it does the thinking for you. I feel like your analogy is closer to a good autocomplete rather than an agentic LLM loop. Agents feel like a genuinely different technology.

Maybe it’s more like an HVAC installer has a tool which tells them where to position the furnace. After many years, will they forget why the furnace should go there? Or maybe it’s like the HVAC installer has a tool which auto attaches the pipes. Hopefully it attached cold to cold and hot to hot, but the HVAC installer hasn’t had to do this in a while so they kind of forgot how to do it manually, and so will just trust the tool.

I was primarily pushing back on anti-screens in its absolute sense, and being somewhat pedantic. However, I largely agree with you, especially considering how screens are most often used today.

If it were a human (going off of memory as it has been a while), they would probably be using mathscinet and their university library to obtain copies of these papers online. Many old papers are digitized and available by these means. I’m sure the AI companies have it all easily accessible and/or the entirety of mathscinet is in the training data. The “personal correspondence” is possibly lifting from another paper or journal but yeah that is a bit odd that they wouldn’t source where they lifted that from directly.

I can’t say if the citations are accurate because I didn’t check.

I am skeptical of using AI for a tutor and would not give it to my kid, but I find the staunch no-screens sentiment equally ridiculous and similar to saying no-books.

Do you have kids? Are you telling me that you and the other parent spend (or spent) 100% of your time with your 5 year old? We spend a ton of time with our kid but also he watches educational TV. Unless you are privileged enough to have a nanny, having your kid constantly with another person is unrealistic to me. Then it is a matter of what they will do alone, and I don't see how a screen is inherently worse than other solo activities. I played video games and tinkered with DOS when I was a young kid and learned a lot, which led to my current career.

Screens are a medium. I don't see how they are inherently bad, although it is probably correct that screens carry the most bad content out of any information medium. But my kid has learned how crayons and instruments are made by watching Mr Rogers, and I simply don't see how he would have learned that via some other medium in the same amount of detail (visual and auditory together). He has seen jazz and classical musicians explain and play their instruments. He has seen how adults interact with each other and respect one another through Mr Rogers and (older) Sesame Street. I've then seen this translate into his real world behaviors. You think that is fundamentally bad?

You would learn way more by asking an LLM how postgres works. Nobody is even reading this Rust code. There are 7000 commits over 2 weeks. What is being learned from that?

> Just run both pgrust and postgres and compare the output.

The space of inputs and outputs is infinite. You can't prove programs are the same by "just" testing a bunch inputs.

Of course, if the rewrite is all of those things, then it seems like a good option. The whole point is how do we know if it actually is better.

Genuine question - do you feel that you understand a code change well enough by only viewing diffs, enough to push the Claude code as “your own”? I can see this for some types of PRs but for anything that is relatively complex or touches multiple files, I personally have to open an editor and navigate through the code paths to really grok if an AI’s (or a human’s for that matter) change is correct.

Is this something people widely vary on? I suspect a lot of people are not like me given that I do see people saying things such as “I rarely use my ide/editor any more” with AI, hence my question to you. Maybe I’m just bad at keeping the code base all in my head.

I do agree with you that I feel that I am coding less myself and I do miss that for sure. But the job demands otherwise.

Every time people point out a limitation or constraint of LLMs, I see a comment that is to the effect of “but humans…”. I don’t understand why this comparison is relevant to this particular thread. Is it just an amusing similarity?

The point is that there is a 0% chance you can meaningfully understand the impact and consequences of what a 10k line change is doing in a couple hour review, so there’s no way for you to know if it’s “bad” in any sense other than it’s so awful that it’s obvious on a skim (which is what 10k lines in a couple hours is).

The entire PR description is filled with LLMisms. I find this style so hard to read, it’s almost nauseating these days. My eyes start to glaze over and I stop reading pretty quickly after. I’m having to deal with this a lot at work these days unfortunately. I mean, did anyone even read or verify this output? When I read this type of stuff, I have the nagging suspicion that I am the proofreader and I have to do the verification myself. At that point, what are we even doing? I can generate Claude output myself…

I know I’m being overly dramatic but this sort of thing feels so wrong and inhuman(?) to me for some reason.

I think the answer to your question depends on what you mean by a “completely egalitarian world”. Depending on the answer, I would ask: why does the kid desire to charge people to earn money in a society like that in the first place?

But yes, I think a lot of people would say there should be a cap on how much more wealth someone has compared to the median, for many reasons, such as the amount of political power it would yield him. The thought experiment uses a harmless activity for earning the wealth, but criticism of wealth inequality is often based on what happens _after_ the wealth is earned too. If in this example you will claim the world remains perfectly (socially) egalitarian and the kid will be benevolent, then maybe we can let him keep his wealth in the thought experiment. But that’s not the world we live in, and possibly never will be.

> As an AI optimist, I think all forced labor should eventually be done by AI. People can then spend their time pursuing their own hobbies. Just as many people still play Go after AlphaGo appeared, because they genuinely love the game.

And what sort of economic system do you imagine will be in place to support billions of people being able to just play Go all day long? How do you imagine the large capitalistic global powers transitioning into that state?

The sibling has pointed out some things and you can check that, but —

This has AI fingerprints all over it and it was clear to me too. The fact that not only do people not notice this, but also that the comment pointing it out is dead and I can’t even reply to it, is sad/scary/funny. The internet of people is literally dying before our eyes and HN just flags anyone helping to sound the alarm.

> What all these theories have in common is the underlying belief that our brain/mind works as the machines we build

It’s interesting to note that this is not a new phenomenon. If you read writers from the past, they were always comparing the body or the brain to whatever the modern machine, or idea of a machine, was at the time, whether a steam engine or automata.