TheAntipopulist

TheAntipopulist Formerly Ben___Garrison 1mo ago

I don't think we'll ever see eye-to-eye on the previous bill, but at this point it doesn't matter. That bill is dead, but Republicans have a trifecta, so there's no reason not to get a real bill done now. There will likely be no better time to pass a bill in possibly decades than the present moment, and it would significantly ease a lot of the issues the courts had with EO's and practical enforcement.

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago

Oh, that's right, you were seemingly the one person on the entire forum that agreed with me back then. Thanks for that, by the way. It was nice to not feel like I was going completely insane.

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago

In my post 9 months ago I agreed with the notion that Biden was fundamentally at fault for the surge of immigration, and that he could have reduced it significantly with powers he already had. This was borne out when illegal immigration numbers plummeted towards the (very) end of Biden's presidency. That said, immigration is still fundamentally broken in a lot of ways that only the Legislature can fix. From my post:

The issue with this idea is that even if Biden were to reimplement all of Trump’s executive orders, they still amounted to little more than a bandaid on a bullet hole. Critics of the bill are technically correct in pointing out that there was less blood before Biden ripped off the bandaid, but it’s ludicrous to then assume that the bandaid was all that was ever needed. US immigration law and border enforcement is fundamentally broken in a number of ways, and this bill would have gone a long way in addressing the worst problems. Recall that Trump himself tried to go after asylum laws directly, but his efforts mostly fizzled in the courts.

The fact no Republicans voted for the bill after Trump told them not to is just an indication that R congressmen are utterly beholden to Trump. Remember that Lankford, a Republican immigration hawk, was the chief architect of the bill in the first place.

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago

Funny, the Democrats explicitly wanted to pass a very conservative immigration bill last year before Trump sabotaged it for cynical reasons.

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago

I really don't see why employers would see that as a big problem. Filtering out junk is not expensive or hard. ATS exists, and this is one of the tasks HR uses to justify its existence. Big companies like being able to grab talent from all over the world.

-1

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago

A lot of the current immigration debate is wondering what increasingly exotic interpretations of statutes can be made to sneak immigration reform through the backdoor without the Judicial system interfering. But this only needs to happen because the statutes the Legislative passed decades ago say specific things. Why not just pass a new bill? I wrote a post (copied to substack here) that detailed the Republican sabotage of the compromise immigration reform bill back in 2024, and a lot of the discussion was predicated on Trump winning a trifecta in November. I thought that was excessively risky... but they won! Why not do what they said now!?! Stop with the judicial cloak-and-dagger, and pass some freaking immigration reform!

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago

It's really only dysfunctional for employees, who have to spend the effort applying to dozens or hundreds of jobs filled with broken interfaces and astrology quizzes. For employers everything's working fairly well, or at least it's not materially worse than it's ever been. It costs next to nothing for them to filter out more candidates, especially if the process is partially automated.

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago

Somethings gotta give. Regulation, or something.

Why do you say this? I mean, I wish it would, but why do you think someone is coming to save us?

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago · Edited 1mo ago

There's certainly been low unemployment for the most part, but that's not the issue (though it could exacerbate things if unemployment spiked). It's the fact that the internet has made applying for things (jobs, dating, schools) so much easier, which led to a proliferation of applications. But applications are mostly a zero-sum game, so employers, schools, etc. have responded by ratcheting up expectations.

This could theoretically be solved if the government cracked down on the most abusive practices (like ghosting after a formal job offer) and instituted a well-designed tax to counteract application spam, but that would probably be as unpopular as congestion pricing, so I doubt it would pass in our populist-addled age.

Context

TheAntipopulist Formerly Ben___Garrison 1mo ago

From college to dating to jobs, no one in history has been rejected more than Gen Z

This is an interesting article about the trend of mass-applications that has become increasingly normalized across many areas of life. If you've applied for a job in the past decade or so, you'll know that the signal:noise ratio is very bad, and as such you're kind of expected to mass-apply to dozens or hundreds of jobs. Each job will get bombarded with something like 1000 applicants in the first few days, and while many of those applicants will be junk, there will probably be at least a few dozen high-quality candidates that you're competing with. This has led to companies becoming extremely picky. In my specific area of tech, its led to an expectation that you need to do dozens of hours of "leetcode", which are little toy problems that are ostensibly used to make sure you actually know how to program, but which actually do a terrible job at this because real programmers will usually be somewhat bad at these, while people who grind leetcode but know little else can do quite well. There's also a further expectation that you might be asked to do other ridiculous feats like have 8+ rounds of interviews for an entry-level position, and you might be ghosted at any point in this process, even after you've interviewed with real people. Heck, you might even be ghosted after you've received and accepted a formal job offer, then if you show up to work the company will just lie and say they have no idea who you are. While there's theoretically some recourse by suing for promissory estoppel, it's almost never worth the effort so it rarely happens. The accepted answer is "that's just part of the game now, swallow your pride and move on".

Dating, and to some extent college applications are also like this. Young people live in a world where they constantly have doors slammed in their face. While I think a little bit of rejection can be good to build resilience, I doubt humans are psychologically well-equipped to handle the barrage of rejection that's become commonplace. Getting rejected hurts even if it's just a small annoyance from not receiving a response. It makes you feel like you're being treated like garbage a little bit, which would almost certainly prompt some amount of nihilism after a while. It might also lead to some amount of risk aversion. I myself simply refuse to deal with online dating at all, which has dramatically limited my romantic options. But if dying alone is the price required to remove this nonsense from at least one aspect of my life, that's a deal I'd gladly take.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

I don't know what's happened to Russian nuke numbers in the short term (i.e. since Ukraine), but they've reneged on arms talks which indicates they'll almost certainly build a lot more once the war is over. So I guess I could have been technically incorrect when I was talking about Russia having increased stockpiles already, they've just signaled they want to in the medium-long term.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

Who said you have to have "utilon maximization as a terminal value"? This is a basic element of any societal critique. "Here's what's bad... and here's a better alternative". Without the better alternative you can often just make a bad thing worse, like how the right has spiralled into being the party of Catturd.

Part-time youtubers are literally better than the mainstream media.

Random content creators are so much worse than say the NYT. The worst youtubers or substackers or whatever are so, so much more awful than the worst NYT columnists in terms of bias and adherence to truth. The best content creators can be about on par with the best NYT has to offer when it comes to op-eds and analysis, but they have big blind spots when it comes to reporting original facts in many places.

-1

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

You can maybe squint and see Russia's Crimea adventure in 2014 as being somewhat similar to Yugoslavia, with having a simmering civil war and all. But their 2022 invasion was in no way similar since they went after the entire country, not just the parts that were in a civil conflict.

Plus there's the big difference that Russia was seeking to annex the land directly to itself in both cases. This might seem like goalpost-moving from my previous response, but it was more an issue of me not properly articulating in the first place.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

The US certainly sparked some prolif itself with foreign policy (mostly Iraq + Afghanistan, Libya was more of a European-led conflict). But on net, the US has been the biggest leader of antiprolif by far. Very few middle powers had nukes under American unipolarity, but that's almost certainly going to change over the coming decades.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

This post is a perfect encapsulation of the totalizing nonsense of negative partisanship.

Is it good? Who cares! It makes the outgroup angry, so it must be good in some way!

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

This is indeed a pretty banger article by Hanania. The biggest problem with disempowering the elites is that you need to have an alternative in mind to make things better. For all the hissing and screeching people do at the mainstream media, the right has been utterly unable to build anything close to a good replacement. Every attempt they've made is either nonexistent or is so much worse, like Catturd's Twitter feed.

I don't know what the heck happened to Elon Musk to make him spiral in the last few years in particular despite being pretty sane for most of the High Woke period, but it's become a problem that governments effectively can't ignore.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

Of course we have seen that this doesn't seem to apply to the US

When did the US ever try to change borders through war? The US engaged in a lot of regime change, but it wasn't the only one doing so (look at Russian interventions in Georgia and elsewhere).

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago · Edited 2mo ago

Hard agreed on this. Nuclear weapons are a classic prisoner's dilemma, where it's better for nobody to have them, but if one side gets them then that's really really bad if the other side doesn't have them. It was America that underpinned most antiprolif efforts since the end of the Cold War, partially through providing its nuclear umbrella, and partially by strongarming those who thought of getting nukes themselves away from doing so. This has led to the dictatorships of the world playing defect-bot, where no democracies have gotten nukes since the 90s, but Russia and China have increased their stockpiles while North Korea has nukes now, and Iran isn't far behind. Russia's has effectively been using nuclear blackmail in Ukraine as well, breaking the previous norm of nukes giving a strong defense-only shield. Now, anyone with nukes apparently has more right than other nations to offensively invade their neighbors too

It's a sad fate, but Poland, Japan, and South Korea all need nukes now. If Taiwan and Ukraine had nukes, then it's likely they wouldn't face anywhere near the level of insecurity they do. They probably wouldn't be able to get them now since China/Russia would freak out.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

I just saw a map of Silph Co, and yeah, that seems like it would absolutely destroy the current iteration of Claude. I'm looking forward to another 100 hours of "Blackout Strategy".

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

This reminds me of a joke from CS: "The first 90 percent of the code accounts for the first 90 percent of the development time. The remaining 10 percent of the code accounts for the other 90 percent of the development time.".

So if the definition of AGI is 'can beat the elite four deathless' we're on track for ~2027 as expected.

Certainly not! I'd want it to beat Pokemon Red in <40 hours (which is 1.5x the average human time to do so), not merely do it deathless. And I also don't think that'd be enough for AGI -- it's necessary, but not sufficient for it to do stuff like that.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

I think by 2027

Thank you for writing a specific year. I think that's wildly optimistic. What you're effectively describing is a fast-takeoff singularity. I think any sort of singularity is at least 5-10 years away, and probably a lot more. I think you've been reading WAY too much hype, but we'll let the next few years judge one of us correct.

I've added your prediction to my notes to keep track of.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

I'd want to limit the LLMs to screenshot input and tools/extensions it writes itself

These are decent parameters, although I'd put a bigger emphasis on "they can clear the game in a reasonable amount of time". Since the average time for a human to beat it is 26 hours, the AI should be able to do it in less than 40.

I'd say 2 years is a bit optimistic. I'd put the 50/50 chance at 3ish years myself.

If LLMs had sufficiently advanced in long-term planning such that the navigation issues in Pokemon were solved, then that would qualify as more than a marginal improvement in my eyes. It would be at least "moderate", if not moreso depending on how generalizable it was.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

Drawing straight lines on graphs, then zooming off to infinity rarely works well. In 1995 computers could barely play chess. In 1997 they beat the best player in the world. But that was hardly a harbinger of imminent AGI as we now know. AI basically went back in the cupboard for a few decades.

but I'd be even more surprised if 3 years from now they didn't play flawlessly.

This is an interesting timeline. I'm assuming you're talking about general AI playing Pokemon and not just task-specific ones (which can already play it flawlessly, e.g. TAS's) which wouldn't be interesting. I'd say there's a 50-50 chance there's a general AI that can play Pokemon like this in 3 years.

Context

TheAntipopulist Formerly Ben___Garrison 2mo ago

Almost all improvements since late 2022 have been marginal. ChatGPT 4.5 was a disappointment outside of perhaps having a better personality. There's no robust solution to hallucinations, or any of the other myriad of problems LLMs have. I imagine there will be some decent products in the same vein as Deep Research, but AGI is highly unlikely in the short term.

I made my prediction 2 years ago and am increasingly sure it will come true.

Context