Live data from Hacker News

Moore's Law for Everything

moores.samaltman.com

401–410 of 554 posts

Re: Moore's Law for Everything

#401

This is definitely good thinking, though it seems unrealistic. The stuff about AI is just science fiction. We'll see major advances in the next 15 years, but it's unlikely that there will be a compounding intelligence situation where robots are able to program and build other robots without human intervention. There's no scientific evidence that this will happen, but it's certainly possible on a long enough time fram…

It's not a property tax (which taxes the building and the land), it's only a land value tax. So if you have a $300,000 house, it's not going to tax 2.5% of the $300,000 house, just the $50,000 worth of land it sits on.

Re: Moore's Law for Everything

#403
post #346

Earlier quoted context omitted.

The 2.5% tax doesn't mean country B is growing 2.5% faster. The 2.5% in country B could be sitting in a vault doing nothing or being spent on yachts. Conversely the 2.5% in country A is being invested in making the workforce more productive.

In general it is right to assume that free capital will be used as productively as possible.

> In general it is right to assume that free capital will be used as productively as possible.

Only if you define 'productive' as 'generating more capital'.

Yes, over the very very long term, externalities that aren't accounted for (it doesn't matter whether we're talking about negative externalities not mitigated, or positive externalities not invested in) will eventually exact their toll on capital, but in the short, medium, and even moderately long term, any externalities that can possibly be dismissed as irrelevant to the fundamental mission of turning large piles of money into larger piles, will be.

It is also worth noting that by ignoring externalities until some crisis forces the issue, opportunities are opened up for capital to (quite profitably) address the resulting crisis. Whereas addressing an externality in a more timely manner, while reducing overall costs, even more sharply reduces those opportunities.

Re: Moore's Law for Everything

#404
post #280

Earlier quoted context omitted.

I don't necessarily disagree but that's really not a good analogy. Horses aren't people. When their jobs disappear no one gives a fuck. No one speaks up for their interests. People will not go so easy. They are able to create amazing things if given the opportunity. No one wants mass human unemployment.

People have an advantage that horses don't have: people can own property, and they have legal rights, like the ability to vote in elections. When we're talking about AI, though, it's not too hard to imagine computer programs eventually being allowed to legally own things and accrue wealth. It's harder to imagine them being allowed to vote (how would that even work?) but something like it could happen. We can also ima…

> It's harder to imagine them being allowed to vote (how would that even work?)

They don't have to actually vote, anymore than corporations or wealthy humans do, in order to exert their will on the political process.

That's what campaign contributions, PACs, lobbyists, astroturfing, think-tanks, etc. are for.

Re: Moore's Law for Everything

#406

Earlier quoted context omitted.

> About 96% of people used to work in agriculture. Today it's about 4%. Let's say the 96% number is from 1750, which I think is roughly right. World population then was ~800 million. Today it is ~7700 million. Most have jobs. The human workforce has expanded almost 10x in this period.

“World population then was ~800 million. Today it is ~7700 million. Most have jobs.” I doubt this. In the USA, we have an unemployment number, “able bodied, but not working, or looking for work.” If the 50’s, it was basically housewives. Today it’s a lot of people. (This might not be the exact wording the government uses.) My point is this number is huge, and never talked about much in the United States. I don’t thin…

I don’t know how the rest of the world treats this number.

India deliberately chose not to automate agriculture. They still have about 50% agricultural workers. Recent attempts to change this have resulted in riots.[1]

[1] https://www.facebook.com/Reuters/videos/242972323968171/

Re: Moore's Law for Everything

#407

Earlier quoted context omitted.

> About 96% of people used to work in agriculture. Today it's about 4%. Let's say the 96% number is from 1750, which I think is roughly right. World population then was ~800 million. Today it is ~7700 million. Most have jobs. The human workforce has expanded almost 10x in this period.

But the food needs (and other needs) have expanded 10x. It makes no sense to look at the absolute numbers instead of the percentage. If there had been no technological advancements at all, the workforce would have still expanded (in absolute numbers) to match the population. Without machines and technology, we would need far, far more people today in agriculture. The need for all those jobs has been destroyed by tech…

One way to look at this is to compare the basic capabilities of humans and machines. The human list remains roughly constant, but the machine list increases over time.

Actually, what's happening in automation is not what people expected. We still don't have good robot manipulation in unstructured situations. Machine learning gave us automation of "higher" functions first.

The current result is a growth in really dumb jobs, like Amazon warehouse pickers. The computers do almost all the thinking. Humans just pick up things where told to do so and put them somewhere else. Such jobs have no promotion path.

This is the "machines should think, people should work" revolution.

Re: Moore's Law for Everything

#408
post #207

Earlier quoted context omitted.

I can't get myself to agree with this. Indeed, it's true, horses don't create demand, but neither do unemployed humans. Our economy isn't centred around humans, it's centred around capital. Unemployable humans are as useless to it as horses, as they can't exchange their labour with capital owners (or become capital owners themselves). Unemployable people might want more, but will they get more? No, they won't.

I think there's two distinct ideas being discussed. Can people be unemployed because of structural, societal, or political reasons? Sure, and I think most people would agree. We have plenty of example of that happening - depressions, recessions, mass refugees, etc. But can people be unemployed because we run out of things for humans to do? That's what the video is arguing, but it won't be the case unless we've fully…

People can be unemployed because there's nothing they can do that generates enough value to pay them enough to live on.

Re: Moore's Law for Everything

#409

I don't buy this whole idea of technological progress being exponential, let alone monotonic. The average piece of software made in 2021 does less, runs slower and is more difficult to maintain than the average piece of software from 2001. Consumer-grade CPUs haven't seen a drastic increase in performance for around a decade too. I'm still running an i7-2600k that was released in 2010. A current-tech Ryzen 7 has mayb…

Compare voice recognition in 2001 to 2021.

Re: Moore's Law for Everything

#410
post #285

Earlier quoted context omitted.

Maybe I don't understand the question, but you can see that in population statistics - certain populations reproduce at an above replacement rate, some below. In a lot of the developing world, however, and no different 150 years ago in the developed world, you'd have a lot of kids because you'd expect several to die early. Once a certain level of prosperity and peace is reached, this doesn't become as much of a biolo…

I am aware that some countries reproduce above replacement rate and some below, and that this rate is normally tied with country economic prosperity. My point was an evolutionary one: wouldn't those couples who choose to reproduce more (assuming you can afford to raise your kids) be selected for, evolutionarily, vs couples who chose to remain childfree? Thus, even in a few generations, I would assume a moderate cultu…

> My point was an evolutionary one: wouldn't those couples who choose to reproduce more (assuming you can afford to raise your kids) be selected for, evolutionarily, vs couples who chose to remain childfree?

Keep in mind that the incentive to have kids, evolutionarily speaking, is weak enough that we have had to be bribed with sex to keep things going.

Once sex and reproduction were decoupled, you're definitely looking at much more than a couple of generations before biological selection pressures start having the effect of an increase in the desire to have children per-se, as distinct from the desire to have sex.

OTOH, cultural incentives are an entirely different matter: whether we're talking about religions decrying or forbidding the use of contraceptives, or advocacy for those who want children at all to have larger families, these can have a snowballing effect especially since culture is passed down to offspring even more readily, though not as reliably, than genetic traits (eg. you can pass your culture on to an adopted child, or to a spouse, etc.).

This selects for cultures that are good at spreading themselves and good at encouraging reproduction, and selects humans for a general tendency toward susceptibility to indoctrination (at least during childhood) including the imperative to eventually pass the cultural package onward. To a certain extent it also selects for acceptance of authority and resistance to indoctrination with beliefs that conflict with already held ones as an adult.

Actually, all these particular biological traits are selected for and reinforced by so many other different feedback loops, that cultural packages can mostly just assume their presence and focus on leveraging them.

Post reply on HN