it is replacing.

If half the population was on UBI, would there still be enough demand (for other than food, clothing, and shelter I suppose) to support the capitalist economy producing goods and services that only half the population could possibly afford?

And if the capitalist consumer-driven economy is floundering on lack of demand, where will the taxation revenue to support the UBI come from?

How does that economic math work out?

by around 2050, the climate might still be livable and our current relatively stable organization of society, food production systems etc may survive.

So I don't care what policy changes you make, as long as your policy is clearly aimed at zero emissions economy by 2050.
Oh, and your policy should probably have metrics that clearly measure the rate of change of emissions say every 5 years til we get there.

That's how I'll evaluate your policy, because anything else is knowingly evil bullshit, as far as your moral obligations to near future generations, not to mention long-lived eco-systems, are.

So how does this latest step accord with my policy evaluation criteria?

That's actually f***ing brilliant. The style reminiscent of a cynical jaded commentator on society. Would be great if all slashdot comment sections had this kind of AI cyberpunk TLDR summary of them at the beginning.

My guess is we will see reductions in the number of people doing each job, before we see total elimination of those jobs for humans.

Because the AI assisted person (or the person just orchestrating AI and keeping it on task or checking for obvious blunders) will be more productive than their AI-less predecessor.

So we should see a productivity increase, followed by economic turmoil and downturn as fewer people earn good income and participate a lot in the economy.

You could argue that there will still be as much money churning around in the economy, just in the hands of fewer people, so no easily measurable economic decline.

But we've already seen that as large classes of people become under-employed (rust belt America for example) we tend to see social discontent, populist misdiagnosing authoritarian leaders rising as ultimately useless and counterproductive champions of the unprivileged, and a bit of a spiral into uncertainty and instability.

AI job reductions will probably make this worse. And all the instability really will severely degrade the economy, and also cause geopolitical instability.

In short, "interesting times" in the sense of the curse.

Don't worry. Most of the training data was actually just dumb web comments.

I just hope the young generation is taking notes for the climate crime indictments of the future.

along with having to suffer their generally inferior because under-invested-in under-refined suite of default applications and information resources.

What's the point?

Well, that's what you get from a severe bout of viral populist policy syndrome.

If TSMC replicates its top-level chip making factories in the USA.

The presence in Taiwan of the only factory in the world capable of making the leading-edge performance chips today (smallest process fabs) is probably a key reason why:

a) China doesn't draw up immediate invasion plans. - China is independently working on replicating TSMC but apparently it will take like a decade, so they themselves are still somewhat dependent on TSMC remaining existing in a viable Taiwan. Plus China knows that the rest of the world would be really monumentally pissed by the loss of the best chips for a full decade or so during a messy invasion.

b) The USA despite Mr. T.'s recent denials, probably still sees Taiwan as an essential ally, critical to US and the world's leading-tech economy.

It is hard to overstate how big a moat TSMC has. They invested many many billions and more than 20 years evolving and constantly innovating to be the only place capable of fabricating the best chips.

With smart robots, the following two economic and political systems are obsolte:
1. Capitalism as a source of trickle-down wealth via employment.
2. Marxism which sees all value as having been generated by human labor (and thus should be returned fairly to laborers).

With general AI and flexibly capable smart robots, resource extraction and production (and management of that, and many service jobs also) can happen with very little human labor. Negligible amounts.

So we are faced with new socioeconomic problems, and need a completely new analysis.

- How should we distribute wealth (means for a comfortable living), if the obtaining and making of valuable goods and services requires almost no labor?
Labor-based socialism fails to answer this. Capitalism floating all boats (by hiring and spreading the wealth) fails to answer this.

- Why are we each here, given that our primary purpose is no longer to strive in the economy to help society (our fellow humans and ourselves) be wealthy and secure? How should we obtain our self-worth, knowing that if we work, usually it will just be a make-work project and AI and robots could do it as effectively or better, and more cost-effectively?

the shameful shit just might be:

- The utter disrespect that the current US administration has for science and academics in general
- The racist, xenophobic, fascist actions of the current US administration with respect to immigrants including legal ones who have the temerity to publicly disagree with right-wing doctrine.
- The fascist disrespect of the current US administration for the judicial branch of government, calling for the impeachment and harassment of judges who impede the ill considered and draconian actions of the administration.

Your country is turning into an inhospitable place for independent thinkers of any kind. Deal with that shameful shit.

But do you understand how quickly the AI technology is being innovated?

It will not be long before the central ideas, themes, narratives, plots for the art or stories can be generated by the AIs alone.

Agentive AI just needs to be turned on the problem of creating art that is likely to be interesting or fascinating to humans, and it will generate itself a todo list of how to accomplish that, then it will study the relevant art and art criticism background from its knowledge base, and get going generating original interesting art.

My guess is this is significantly less than 5 years away.

such knowledge as is expressed in human language communications which are ingested comprehensively into LLMs....

captures and represents the semantics of the entities and relationships (including situations) which the humans are writing about.

LLMs are actually learning some kind of amalgamated and averaged version of the collective human knowledge base about the world. In particular, the aspects of the world that we think are important to express in communications to others.

LLMs are thus "borrowing" our discrimination of the wheat from the chaff in observations of the complex world around us.
We don't describe everything that is out there, or everything going on. We edit to the relevant to our concerns, goals etc.
LLMs also then learn this from our communications en masse.

Thus LLMs learn (capture in their "neural" weights and biases) a human-like conceptualization of the (important entity types, instances and relation types and instances in the) world around us.

I've read your posts enough to know that you're not going to believe this leap from the syntactic and statistical-over-corpus to the semantic.
It is none the less true. If you can't see the actual thinking and creativity these LLM systems are doing, wait a few years while extended thinking chain versions of them are refined. Pretty soon, shouting that it's all just statistics (and bits or whatever) will be ignored, as society moves on to consider how we can coexist with these cross-training superintelligences.

He called the complexity of the maintenance of code in two languages rather than one in a module the cancer. Due to a breaking of uniformity/simplicity/regularity of convention etc. I would surmise). He did not call the new language itself, per se, cancer.