Live data from Hacker News

The Gentle Singularity

blog.samaltman.com

531–540 of 557 posts

Re: The Gentle Singularity

#531
post #474

Earlier quoted context omitted.

Discovery definitely becomes more important in this world. Fortunately AI helps with that too.

Why do you think it will not soon be biased as search is?

I think AI gives us the opportunity to write our own ranking algorithms. Of course most people will pick an off-the-shelf editorial perspective but I expect these to be much more configurable than the current state with a tiny number of algorithmic gatekeepers.

Re: The Gentle Singularity

#532
post #481

Earlier quoted context omitted.

Another way to look at it is - why force the agent to grapple with the whole code base when they can rapidly standup many single purpose services instead?

A large code base with modules with clear interfaces accomplishes most of this and is more efficient. Often vastly more efficient. Compare the cost of a function call or an in process message queue with a network API call over http. Cloud providers make more money the more they can get people to use inefficient designs with more moving parts to rack up more charges. Bonus if it also locks you into managed services. D…

Yep all true. Im not really arguing a point so much as thinking out loud - I guess the core idea I'm chewing on is whether LLM assisted development is enough of a paradigm shift that things that were formerly bad ideas become good ideas - but perhaps this particular line of thought (microservices) is the wrong one.

Re: The Gentle Singularity

#533

Earlier quoted context omitted.

>. They fall on their face trying to work in sprawling codebases. >However, they are excellent at writing small bespoke programs For programmers I predict a future of a million micro services. Sprawling has always been an undesirable and unnecessary byproduct of growing code bases, but there's been no easy solution to that. I wonder if LLMs would perform better on a mono repo of many small services than one sprawling…

The issue as I have observed it is that software is written to cover as many use cases for as many users as possible. Obviously executives want to sell their product to as many hands as possible. But end users often only need a tiny fraction of what the software is fully capable of, leaving a situation where you need you need to purchase 100% to use 5%. LLMs can break down this wall by offering the user the ability t…

Yeah if code-gen gets to a point similar to image-gen, where a user can prompt a full micro utility, then we will see the dual erosion of more DIY and also lower barrier of entry to the industry. Still most software value today is not about the utility but the IP, the network, and convenience. I suppose convenience is the main pillar at stake here.

Re: The Gentle Singularity

#534
post #463

Earlier quoted context omitted.

In a way it is. Why are housing costs so high in Redmond, WA? The result of an influx of high income tech workers to the local housing market and the resulting shift of prices such as to eventually dilute the utility of that high salary to begin with. People in the area without a hook on that whale are of course left high and dry.

No... the reason housing prices are high and rising is that not enough housing is getting build in the places people want to live. The main reason for that is that the people who already live in those places can block construction of new housing. That, and zoning.

This is part of it yes. But after sufficient housing stock is built for the available working population, prices generally reach an equilibrium around median worker pay in the area. And if the median worker pay is now inflated by tech salaries and you are are still a line cook, you’ve been screwed. The construction market will not respond by making housing until it is cheap and unsold enough to be affordable on your line cook salary. The construction companies would sooner take jobs in more lucrative markets. If you are a line cook where most everyone in the area is paid about as much as you on the other hand, you can probably afford a local mortgage.

Re: The Gentle Singularity

#535

Earlier quoted context omitted.

Sure, people’s well being as a whole haven’t gotten better since the 1980s, except for >Air quality (no more leaded gasoline) >Life expectancy >Cancer survival rates >Access to information >Infant mortality >Violent crime rates across the western world >Access to education >Clean water and food for 4+ billion people >HIV treatment >etc The negativity on this site is insane. They will deny the greatest scientific achi…

My position is more nuanced that you present it to be. I’m arguing against Altman’s notion of progress as being driven by scientific and technological achievements. Looking through your list it’s social policies which either drive or are necessary components of most of the improvements you mention. Even the medical advances you mention are dependent on a society’s ability to offer healthcare to its members - a notabl…

You need both social change and technological change to drive progress, but more often technological change drives the conversation.

Standards for workplace safety, working hours, social welfare are only possible because technologically developed societies are sufficiently productive to even consider those possibilities. The idea of a 40 hour workweek, the fact that most people aren’t farmers now is due to the industrial revolution, machinery and tech multiplying the productive capacity of a single individual. HIV treatment, cancer treatment, blood tests are a standard part of healthcare not because social policies willed them to be so alone, but because the technology to deliver those treatments at scale exists and allows that to be a possibility in the first place.

I do agree that without the right social structures, policy and general political will, the conditions for those technological advances may never have come about in the first place, but in general, it is technology that makes certain allowances for human flourishing possible at all.

Re: The Gentle Singularity

#536

Earlier quoted context omitted.

Link for context (to what Ed Zitron said)?

https://www.wheresyoured.at/openai-is-a-systemic-risk-to-the...

Taking his word for it, that OpenAI is casting its future on two companies that have never built a single data center is cause for a raised eyebrow.

Re: The Gentle Singularity

#537
If OAI/Sama believes they are taking-off in that AI is building the next model we should see a step change the next model. As they accelerate the gap will grow, we'll see next model whether this is PR/Marketing or if they really are taking off

Re: The Gentle Singularity

#538

If OAI/Sama believes they are taking-off in that AI is building the next model we should see a step change the next model. As they accelerate the gap will grow, we'll see next model whether this is PR/Marketing or if they really are taking off

Even in Sama's telling it would only have a small effect just now. "larval version of recursive self-improvement".

Re: The Gentle Singularity

#539

Earlier quoted context omitted.

My position is more nuanced that you present it to be. I’m arguing against Altman’s notion of progress as being driven by scientific and technological achievements. Looking through your list it’s social policies which either drive or are necessary components of most of the improvements you mention. Even the medical advances you mention are dependent on a society’s ability to offer healthcare to its members - a notabl…

You need both social change and technological change to drive progress, but more often technological change drives the conversation. Standards for workplace safety, working hours, social welfare are only possible because technologically developed societies are sufficiently productive to even consider those possibilities. The idea of a 40 hour workweek, the fact that most people aren’t farmers now is due to the indust…

You’re right that technology is a necessary ingredient in some cases such as medical treatment. However workplace safety and 40 hour work weeks are not good examples. In the 1800s factory owners extracted as much value from their workforce as they could; they did this by maximizing working hours, dangerous working conditions, child labor, etc; all while enjoying lives of affluence themselves. It’s the fact that factory owners were living lives of affluence that we know the workers were being exploited. There was an highly unequal relationship between workers and owners. It’s not the productivity growth of the Industrial Revolution the made lives better for workers - actually it got much worse initially as they were squeezed by the capitalist class. It’s only when workers organized to struggle for a larger share of the value they were creating for their bosses that their lives got better.

Today wealth inequality is at unheard of levels. The wealth of the richest person roughly equals that of the 50% poorest. That is profit extracted from the value workers produce for the wealthy. Technological andvancement isn’t required to give workers a larger piece of the pie; the pie today is already quite enormous. Instead workers must organize if they want better lives. In lieu of that, advances like AI will continue to only profit the very wealthy; the rest of us will likely experience a mix of advantages and drawbacks (analogous to what we’ve seen with the internet or smartphones).

Re: The Gentle Singularity

#540

Earlier quoted context omitted.

You need both social change and technological change to drive progress, but more often technological change drives the conversation. Standards for workplace safety, working hours, social welfare are only possible because technologically developed societies are sufficiently productive to even consider those possibilities. The idea of a 40 hour workweek, the fact that most people aren’t farmers now is due to the indust…

You’re right that technology is a necessary ingredient in some cases such as medical treatment. However workplace safety and 40 hour work weeks are not good examples. In the 1800s factory owners extracted as much value from their workforce as they could; they did this by maximizing working hours, dangerous working conditions, child labor, etc; all while enjoying lives of affluence themselves. It’s the fact that facto…

Worldwide wealth inequality isn't as important to look at as overall worldwide poverty, which has gone down markedly over the past century. It is much preferable to be able to have running water and food despite the richest person in your country having 30 houses, over having no water or food but the richest guy only has 2 houses. In the former, inequality is higher, but the absolute pain of poverty is lower, which is a more relevant metric for human welfare.

The industrial revolution, in the 1800s, transformed England and Europe due to the incentives of the growth of industry, and the free movement of labor towards those incentives. Largely, those who worked in factories did so because the benefits/costs exceeded that of farm work. If there were no incentives to change, everyone would have stayed on the farm, preserving the agrarian status quo. Despite all the hazards, the benefit/cost was better for industrial work, largely because it paid more, the money allowing for the improvement of living conditions for those who worked in factories.

I agree that to get better working conditions has been a constant battle between the interests of the working class and capital owning class, but looking at everything through that lens ignores the whole point of work/industry in the first place, in that it produces surplus wealth, and only once that wealth is produced can societies make the decision on how that wealth is allocated. In every society that has generally followed the rule of "free market capitalism with common sense regulation and protection of private property", living standards, median wealth ownership and wages, etc. have almost universally gone up decade after decade.

Post reply on HN