@TheDag's banner p

TheDag

Per Aspera ad Astra

4 followers   follows 12 users  
joined 2022 September 05 16:04:17 UTC

				

User ID: 616

TheDag

Per Aspera ad Astra

4 followers   follows 12 users   joined 2022 September 05 16:04:17 UTC

					

No bio...


					

User ID: 616

I don’t have perfect answers, but I like the odds of a network state where people use technology to set up large tribes without geographic proximity or ties to a nation state. Basically the nation state order will break down into smaller and smaller units, as we get enclaves of groups that have niche interests joining together online.

To be clear I still think the nation state will exist for a long time on inertia, but it will slowly become rendered obsolete as people start associating with smaller circles.

Local politics are indeed quite impactful in the US. In fact back in the day that was the explicit goal - the founders wanted us to care more about our mayor than the president.

Unfortunately the massive rise in federal power and the hero worship of federal politicians with radio and tv has killed local politics. Even if the power is still nominally there, it’s just too boring to entertain the average person compared to the massive waves of attention grabbing content available everywhere.

Peasants dying out due to a lack of fertility is very different than being actively culled by elites in my book.

Don’t boo out group is a reporting bucket.

I largely agree with @2rafa, but another important consideration is the dysgenic spiral we’re seeing intelligence in many first world countries. The Yuddite argument is generally to take it slow. However if you see our civilization and intellectual capacity going into decline due to stagnation, why would you argue to slow it down? What makes you think our children will have a better ability to align AI, in the counterfactual where we lock it all down?

I’m always surprised that folks in the AI doomer camp seem to be so tech positive, but don’t see the downside of restricting one of the most useful technologies we’ve ever created. If we slow down the economic engine too much, we’ll have a much harder time with AI alignment in my view.

Perhaps I should’ve chopped off the “better than others” part, which is probably largely true. I still doubt the rich would be okay with people actively dying in droves, if it’s cheap and easy for them to prevent it. Which I expect it to be, if we can keep developing AI and reach a less scarce state.

Hell it’s relatively cheap already to just pay for someone to have an internet connection, computer, food and rent. It’ll only get cheaper as we go.

That is my contention in a way. I think the models that can work in multiple modes, which are already out can be scaled up to something approximating AGI.

My bar is probably quite lower than others here, but even if we just scale up the current models, and work out the kinks and plug-ins, as well as web access, I’d argue that qualifies as AGI. The average human is not actually that smart or economically productive, and if you can get something that does even a good approximation of the work but runs thousands of times faster at his order of magnitude it’s cheaper you’ll start increasing productivity and advancing the pace of AI research quite quickly.

I think the way people focus on timelines with regards to tech development is a bit narrow. I don’t fully buy into the singularity, but what makes it a compelling argument is that we will recursively continue to increase the speed of improvements. Even if we don’t get self recursion, we humans can do a good job leveraging the tech we have now.

I’m at least partially convinced of Venkatesh Rao’s idea that sociopaths create bureaucracy within organizations in large part to be able to shift blame or reap praise as the winds shift. Essentially large bureaucracies exist to make things move slowly and give people scapegoats at different steps.

This setup leads to slow-moving public statements generally, and combine that with a culture that follows the precautionary principle, you get large institutions that generally don’t say anything controversial.

Venkat Rao has done some vague mumblings about how AI and crypto are 'two sides of the same coin,' and I'm curious to see where he goes with it. I like to think there is a sort of synergy between the two technologies that goes deeper than "AI creates fake stuff" vs "crypto tells us what stuff is real or fake".

Agreed here. I think that most major institutions will be massively blindsided, at least from the public perspective. With a technology this volatile that's so hard to predict, I think the risk assessment will make these hide bound juggernauts try to avoid taking a public stance until the chips fall decisively to one side or another. Which is precisely when their statements will cease to really matter.

This is, of course, by design.

If they're not needed and can't use violence to effectively overthrow the system then why would anyone need to pay any attention to them whatsoever?

You're drastically overestimating how cold blooded those at the top of society are. I find this sort of rhetoric so infuriating - do you really think elites all walk around with a view that they are better than all others, and the peasants can just die if they're useless? No, that's not how it works.

Most rich or wealthy folk want the approval of the masses. Status is arguably more desirable than wealth for many. They conceive of themselves as popular, someone that others look up to, and that helps give elites their sense of self. Even if there were absolutely no point to "peasants" living, which I still doubt, elites would let them live out of a sense of care for other human beings, and a need for adoration.

Regardless, humanity still produces quite a bit of important work, and will continue to into the era of AI. We can judge things, laugh, provide companionship, and generally instill meaning in a meaningless world. Even though we have a foolish scheme of employment vs unemployment that disregards vast swathes of human work, I'm optimistic AI will help us reimagine what it means to work. We very well could have people building community, caring for themselves and others, child rearing, discussing novel ideas, etc. after AI rises to swallow most of the economy. It depends on what we want to see happen.

Do you think the existing elite is likely to leverage AI to entrench themselves, or are they behind the curve on this new tech adoption?

It seems like elites weathered the storm of the internet and the social media/mobile revolutions pretty well, although I think the latter was more of an ‘on paper’ revolution in many ways. Then again mass technological shifts. are ripe times for regime change.

Do we see the techno-capitalist dystopia controlled by a few that the left loves to imagine? Or an open source freedom driven spread of LLMs on local hardware, as I think @DaseindustriesLtd would prefer?

The employment regulation with regards to benefits for full time employees also creates a stagnant market. You’d imagine without these sorts of regulatory requirements on hours worked per week, there could be much more flexibility and experimentation with different working hours in different business types.

In my view the amount of hours needed per week should vary drastically by industry, but unfortunately that’s not possible because we’ve decided all full time work is 40 hours, period. This has massive switching costs for the economy and labor market especially.

The future of AI is likely decided this week with Sam Altman's Congressional testimony. What do you expect?

I expect nothing to happen for another few years, by which time it's too late. As @2rafa mentioned below, I'm convinced AI research and development is already far ahead of where it needs to be for AGI in the next couple of years. Given the US's embarrassing track record of trying to regulate social media companies, I highly doubt they'll pass an effective regulation regime.

What I would expect, if something gets rushed through, is for Altman and other big AI players to use this panic the doomers have generated as a way to create an artificially regulated competitive moat. Basically the big players are the ones who rushed in early, broke all the rules, then kicked the ladder down behind them. This is a highly unfortunate, but also highly likely future in my estimation.

It's ironic that we've entered into this age of large networks and systems, yet with the rise of AGI we may truly go back to the course of humanity being determined by the whims of a handful of leaders. I'm not sure I buy the FOOM-superintelligence arguments, but even GPT-4 optimized with plug-ins and website access will be a tsunami of change over the way we approach work. If there are more technical advancements in the next few years, who knows where we will end up.

What annoys me most is that this doomer rhetoric lets politicians act like they're doing something - stopping the AI companies from growing - when in reality they need to face the economic situation. Whether it's UBI, massive unemployment benefits, socialized housing, or whatever, our political class must face the massive economic change coming. At this rate it seems neither side of the aisle is willing to double down on the idea that AI will disrupt the workforce, instead they prefer to argue about the latest social issue du jour. This avoidance of the economic shocks coming in the next five years or less is deeply troubling in my view.

One suit. Will have worn it five times by June.

It helps too if you can discuss these things with other people, online or IRL. There are a bunch of SSC meetups and effective altruism meetups if you’re into that crowd.

I know what you mean. I felt this way somewhat when I found the rationalist space years ago. I basically just put my non work life on hold and read all these posts from slate star: https://slatestarcodex.com/archives/

With a lot of googling and other research as well. Took me about six months but it was worth it, my world view changed entirely.

Thanks!

Really?

Do you have a text argument similar to this podcast? I am interested but prefer to read.

I'd rather see this a bit more fleshed out for a top level comment, personally. This feels like it belongs more in SQS.

Well how do you make parents more brave and conscientious etc? I tend to agree that government programs have issues, but even a broader social movement needs something to coalesce around.

I wish I had an explanation here because I think this is mostly true, but I'm also sure our local historians could come up with a lot of notable exceptions. I'm thinking of things like Victorian gentleman-scientists who could only do their experiments or adventures because they were landed gentry

I think the difference is clearly based on self perception combined with status in society. Because welfare (rightly) has such a negative social connotation, and because there are a lot of bureaucratic procedures around getting on welfare that can function as transformative rituals, I’d posit that people who go on welfare see themselves as fundamentally worth less than they were before. When you have no strong belief in religion, the modern state becomes the most powerful force in your life. If that force is continuously telling you that you can’t productively contribute, you’re likely to believe it.

IMO much of postmodernism could only be dreamed up by people insulated from the consequences of their thoughts.

I actually think we have the opposite problem - there are tons of issues like HBD, the difference between equality and the real world, etc where there are plenty of consequences for those thoughts. Postmodernism also has a wealthy tradition of important insights behind it, but runs into the problem of being inscrutable. Because it’s inscrutable many grifters just parrot the language for status without understanding it at all. Doesn’t help that numbers of published papers are rewarded far more than the quality of someone’s writing.

On the side of risk, I think we generally agree from what you’ve said. I also am working to increase my risk tolerance, and there’s really no substitute for just taking risks even if it terrifies you.

Systematizing it does tend to make it too big to solve and removes the responsibility of anyone to solve it.

This is the fundamental paradox of a massive, globally connected society. We need systems to coordinate, but humans are built to live in systems. Definitely makes for interesting discourse.

The idea seems to be "to keep the homeless person from assaulting people, you must first solve drug addiction and mental illness".

Part of this is also the nature of democracy. It's hard to win a campaign being 'tough on crime.' Generally more feel-good solutions will appeal to a broader majority, as they're far easier to justify and seem less morally fraught. Even if these types of solutions have worse outcomes overall.

Great comment, reported for quality contribution.

The problem isn't guns, the problem is that there are millions of disaffected people living in a country founded on the idea of individual human rights. That works when the people are hyper-invested in their families and the future that they'll be living in; that doesn't work when everybody is depressed and hates each other.

This is an underrated point. So many times when societal ills come up in rationalist discourse, people hand wave away, are ignorant of, or flat out ignore the fact that at least 1/5 (probably more in actuality) of the US population is depressed or has some mental disorder. Even given that our modern psychiatric framing is largely faulty, this mass wave of disaffectedness means that traditional solutions, things that worked for our forefathers, need rethinking.

Even though past societies had plenty of times of upheaval, they had different ways of fixing things. Revolutions, massive aligned religions, cultural processes and holidays like the Roman Saturnalia which acted as a pressure release valve for hierarchical resentment. We've been increasingly preventing a release of the pressure, and it will only get worse as we continue to do nothing.