elevatortrim

How can you get the agents to do anything useful without giving them meaningful access?

If it only lives in an isolated sandbox, it can only act within the sandbox, then I would have to manually move what was done in the sandbox to real-life.

I am not saying it should have critical access, but this is more of a question: How can you get value out of AI if it can only act in a sandbox?

I just had chatgpt draw a table of wealth distribution by year based on best data sources it can find.

It shows that inequality has been on the rise from year 0 (top 1% has 45% of wealth) all the way until WW1 - top 1% had 65% in 1910. It then drops to 45% again post WW2 and has been on the rise since. 2026 shows top 1% own 62-63%.

What is interesting is, the bottom 50% has never been poorer. The table starts from 3% for bottom 50% and fluctuates between 1.8 and 5 all the way until 1970 (5%) which marks the beginning of a sharp decline. Today, bottom 50% has 1% of the wealth -a historical low- while the top 1% is almost at a historical high. The wealth distribution has never been more unequal.

Obviously the total wealth kept increasing and an average person today would have much more than an average person at any point in history, but people usually compare themselves with others alive today, not others who lived 100 years ago.

https://chatgpt.com/share/6a0b9a2a-4c6c-8394-8a66-7f86c510c8...

Exactly. And this is why ideas like post-scarcity or universal-high-income are not realistic. Because we can already make life heaven on earth for everyone alive. But we won't. Not because we do not have the resources or means of production, but because we do not have the willingness and systems in place. If we do not fix these first, AI-abundance will bring more suffering, not more prosperity.

We should fix this and feed, house, and cloth everyone. We should create the systems so people are taken care of, and critical mass of people have enough culture and education and good incentives so it sustains. Once we know we can do this and the culture and the systems are irrevocably change in humans' favour, we should then look at AI-abundance.

Just learn, sure. But the difference between my efficiency of using it on my day 2 and month 6 is significant. Yet I feel I am barely scratching the surface of it.

The reason for food shortages is not scarcity of food, nor the reason for shelter, clothing, or transportation.

We have enough technology and resources to make it a heaven for everyone except for unpreventable diseases or similar.

Yet we do not because we did not crack the human-alignment problem.

That’s the problem we need to solve, not the resource problem.

If we solve the resource problem before the human alignment problem, we will cause unimaginable suffering.

If the assets are public and not associated to my account, how could they ever restore access if they made them inaccessible?

My guess is that if they take down the public hosting, most clients would lose access to work they paid for and fiverr has no way to put these back behind an authorisation. It is just a public list of files, either everyone has access to your file, or do not, including you.

I can’t elaborate on your example as I’m not very knowledge-able on the smallpox vaccine bit that depends on how close we were to inventing the vaccine anyway - I’ll take your word on we were not and Edward Jenner had a revolutionary advancement. Then we can say we have the vaccine thanks to him.

But when it comes to space technology, if it was possible to produce the same technology by targeting the required technology directly, we can’t say it was because of war only because some of the inventions of war were re-used. It eould be like saying we have a 45th president thanks to Trump - it would be absurd as we’d have a 45th president anyway.

So I do not think there is enough grounds to attribute this mission’s success to war - with some of the war’s budget, NASA could have invented the required technology anyway.

No.

Because success is individual, inequality is statistical.

It ia true that AI gives ordinary people a lot more chance to be successful.

But do not forget that success depends on lots of factors that are not in one’s control: knowing the right people, time being right for what you are doing, and lots of others. So while the mechanics of success is a lot different to lottery, it does not work much differently: 1 in 1M attempts are successful.

Yes, AI gives everyone more lottery tickets, but it gives rich people a lot more tickets.

I would rather you did not do that. You would create a shit tone more global transfer of goods accelerating global warming, and make societies dependant on unsustainable dirt cheap production practices.

Even if yourself could argue that you’ve done a good thing overall, I’d rather not take your word on that and would rather not have you decide something so extremely impactful.

Is what you say really controversial?

Except for commercial flights (which I would easily give up for a hopeful society), I do not find anything on your list remotely relevant to my happiness or well-being.

Imported cheap goods are obviously something all of us consume a lot, but we only need them to feel good in comparison to our neighbours.

As long as we keep them for hospitals and medicine, the rest going away would be just fine. Children would play with whatever they can find instead of cheap plastic toys, we would have to learn to multi-purpose our tools instead of having a specific object for every minor purpose.

To say “because of war”, you would also have to prove we could not do it without war.

You are right.

I think there are quite a lot of people who are only interested in playing and never deliberately practising. They do not get that far (they do not have to!).

And then the vast majority of aspring guitar players who frequent learning online material (including me) spend all of their time practising and learning, and too little of it playing for fun and performing. Most are constantly frustrated about their progress.

Then there is a small group of people, who spend a lot of time playing for fun and performing, but also a good amount of time deliberately practising. In my experience, those tend to be the ones people think are great players.

If you can afford it and have the time, your local conservatoire.

Piano is very different than the guitar, it is much more technique and repertoire focused (as opposed to improvisation and interpretation of guitar).

In comparison to guitar, there is much more and much higher quality material available for piano as it is being taught for hundreds of years, but most of it is accessible through classical teachers.

There is definitely opportunity for someone to create the justin guitar of piano though, I do not think it has been done yet.

Putting in the hours is important, but not I would not single it out as what matters because there is so much varienty of talent between musicians who put in similar hours. What you do in those hours at least equally matters.

I can attest. After 7 years of practising guitar, the gap between my ability and my taste is even greater compared to when I started.

Actually I can say the same thing for programming, I can build most software I would think of building when I started 20 years ago, but there is still a large gap between what I can build but what I discovered later and now would like to be able to build (I'd need to learn lots of maths in addition to other things).

I think there is an expectation that it should not matter, and there is a reality that it does matter, and there is lots of discussion because the expectations and the reality do not match.

I think being conventionally attractive gives you a lot more chance practice socialising and my observation is that, people who use that chance get so good at it, they remain very good at relationships even at old age.

That world is not necessarily bleak.

We currently have two broad mechanisms to equate people's value.

*Employees:*

Easy to replace = Low Salary = Gets Few Resources

Hard to replace = High Salary = Gets Many Resources

*Entrepreneurs:*

Output consumed low = Low Pay = Gets Few Resources

Output consumed high = High Pay = Gets Many Resources

(Resource consumption ignored)

In a world where machines do everything, aspects of these change:

*Employees:*

Easy to replace = Gets whatever resources

(no-one hard to replace)

It is up to us to define whether 'whatever' is bleak or not. If we decide that resources need to be shared fairly, it could be heaven, not hell.

*Entrepreneurs:*

Resource consumption: Whatever

It is up to us how much resource consumption we allow. If we decide that resource consumption need to be sustainable, it could be heaven, not hell.

I would not want a quadrillion dollars in the first place, I would first try to reduce it to an amount that I can maintain relatively hassle-free and under-the-radar. But even for that perfect amount, I can't think of an age where I'd want to spent 10 years in prison, no matter how comfortable it is.

Not sure if that would work better:

1. The idea behind UBI is that it is near-zero effort, the cost to operate UBI should be minimal. UBW cannot be low overhead I suppose. 2. What motivation do I have to do the work if I can’t get fired?

Everything you say is still the idea and I agree but where does the idea that progressive taxation would go away coming from? What does it have to do with UBI?

Nah - I've also spent decades trying to become the best software developer I can and now it is giving me enormous power. What used to take me 5 days is now taking me a day, and my output is now higher quality. I now finish things properly with the docs, and the nooks and crannies before moving on.

What used to take incompetent developers 5 days - it is still taking them 5 days.

I’d do that in a heartbeat for IWC!

There are of course people who buy Rolexes or whatever for the brand. There are probably more people who buy Mac Books for the logo.

But most people are rational. Most people would pay that much money for a watch only if it does not impact their finances in a meaningful way, but within that, they would want to buy the highest quality they can. And for some people that means an IWC, Omega, Longines, or whatever. If they could buy the same quality from a less known brand, they would. Lots of people buy Grand Seikos at Rolex prices. I buy normal Seikos at $300. We all pay for the quality we can afford.

Brand is important because it is an insurance that you are not being scammed. Besides that, it is not the main factor let alone “era defining” as the article is trying to make it out.

Considering he is one of the most prominent tech writers, PG's articles are likely among the most impactful to have influenced AI writing styles.

Fair enough.

This particular article, I would say compherensive, but lacking substance.

The very promise of this article, that people primarily paying for brand name is simply wrong and it does not take too much time to spend in watch circles to know this.

What people pay is not performance/features/value, obviously. But the prime factor is also not the brand name. With the exception of Rolex, practically nobody knows if your watch costed $100 or $100,000.

What people pay for is artistry, reliability, and level-of-polish. These all have a cost that justify the price of the watch. Nobody pays $10,000 for a watch that could be produced for $100 just because of its brand name.

Of course, they look like normal watches, that’s the point. However, if you paid for, you would get an extremely polished watch, rares/high-quality materials, hand-checked for every imperfection etc. as opposed to an almost the same looking watch for a brand, say Orient, which you would be able to find minor imperfections even as a non-enthusiast.

These particular watches are not my taste as well (I do not find them ugly though) but these are some of the most popular examples watch community find very pretty.

I would argue Garmins are better value for the money for fitness features but I agree. I excluded Casio because they build “brand era” watches as well as extremely cheap watches, not to mention all the other things they do. So I’m not sure where Casio is to average consumer when it comes to watches.

Actually most watch collectors do not wear their most expensive watches, they have daily or travel watches to wear that carry less risk.

Also, watches are status symbols to a really narrow niche. Vast majority of people cannot name a non-Rolex expensive watch and would assume Seiko (and maybe Tissot) are the best watches after Rolex, followed by Swatch (and maybe Timex).