@Brainwavez's banner p

Brainwavez


				

				

				
0 followers   follows 0 users  
joined 2025 December 28 04:50:10 UTC

				

User ID: 4102

Brainwavez


				
				
				

				
0 followers   follows 0 users   joined 2025 December 28 04:50:10 UTC

					

No bio...


					

User ID: 4102

The kids aren't alright (continued)

This college graduation season, many commencement speakers are extolling AI, then getting boo'd by the students. Most notably Eric Shmidt, in University of Arizona, after telling students to "deal with it"; also less recognized speakers in smaller universities (like MTSU and UCF).

Glendale Community College received additional boos because it used an AI tool to read students' names, which messed up.

In contrast, Steve Wozniak told students they "all have AI — actual intelligence" to applause.

This reflects multiple overlapping problems:

  • Age gap: Partly because of TFR collapse, old people have more resources, and are catered to more by politicians (who also are usually old themselves)
    • The graduates are Gen Z, the speakers are old (Eric Shmidt is a baby boomer)
  • Wealth gap: The white-collar job market (at least certain fields, like tech and art) is struggling, while top white-collar employing businesses are doing fine
    • The graduates are white-collar employees, the speakers are CEOs
  • AI favorability gap: AI has the potential to make the wealth gap worse and college more useless, to an extent it's already doing so
    • The graduates are against AI (believing it's contributing to their problems), the speakers are in favor
  • Collapsing college
    • College tuition has increased to absurd levels
    • College has become easier, evidenced by grade inflation and more attendees
    • College has become less personal, because there are more attendees
    • AI makes cheating much easier
    • College has become less helpful towards getting a better job, because there are more attendees, and grade inflation & cheating have caused employers to less value accreditations and GPA

Tech students are particularly affected: many were told that if they went to college, they'd be practically guaranteed an easy, high-paying job, like their older peers; but today they graduate to a bad job market. Meanwhile, the companies they planned to join are posting record profits. AI has invalidated some of their learned skills, and moreover, has the potential to worsen the job market and wealth gap.

Although it's not just tech. Liberal arts students have worse job prospects (although some of theirs were never good), and seem to be more against AI. Law and accounting are apparently being impacted, because AI automates their entry-level jobs.

In summary, the speakers have a completely different perspective due to their age, AI outlook, and wealth; and students aren't happy to see their college which has failed them do it one last time, by appointing an out-of-touch speaker (or using AI to flub announcing their names).


Where to go from here?

Undergraduate education is deeply flawed. I think (not an uncommon position): students should only go to college if for graduate education (which is also flawed but for different reasons, and has purpose until ASI or a suitable alternative). Otherwise, they can learn degree skills in high school or on-the-job training: probably a free unpaid internship, which (as long as it demands real skills, not cheap labor) would be an improvement over paying for college; or pursue a trade. But first, employers must no longer prioritize (let alone require) college degrees; I believe this is happening in some fields, but slowly. In the meantime, more students should and will attend cheap online degree mills, possibly alongside an internship (to graduate with job experience and a better resume).

As for AI...I don't really know. It has some great use-cases, and the potential to strictly improve standards of living (why do something that AI can automate?); it and/or another revolutionary advancement is probably necessary to mitigate climate change and TFR collapse. But it also causes some problems, and has the potential to create global catastrophe. Regardless, I don't expect I or the graduates can influence its evolution or effects. For those reasons, I'm not really optimistic or pessimistic about it. At least I'm aware enough not to extol it to college graduates.

If

  • The ticket is legitimate

  • The NGOs suing can themselves be sued for frivolity

The cop has nothing to fear. Although I can imagine a lawsuit for a frivolous lawsuit for a frivolous lawsuit…getting out of hand.


I believe that generally, the law should, only and be heavily incentivized to, be applied in obvious cases. In ambiguity, the justice system does nothing, except ensures the defendant has a way to prevent or at least record future infractions.

For example:

  • If someone runs over the speed limit, they get fined. If someone is wrongly fined and has their own evidence, they appeal, and get repaid with extra. If someone is wrongly fined but has no evidence, better luck next time: they can start recording their speed, so if they’re caught again they’ll have evidence to reverse both cases.

    • Except in the third scenario, precedent itself is evidence: if cops in an area are repeatedly caught issuing wrongful fines, their future contested fines will be presumed invalid unless they provide evidence.
  • If someone has a legitimate claim, they sue. If someone is clearly illegitimately sued, they counter-sue. If someone is illegitimately sued but can’t prove obviousness, both parties waste their time.

    • Again, precedent: if someone keeps filing failed or ambiguous lawsuits, their ambiguous future lawsuits in that category will be deemed frivolous. If someone keeps defending against failed or ambiguous lawsuits, future ambiguous lawsuits in that category will be deemed frivolous (in their favor).
  • Rulings can be appealed a fixed number of times, possibly zero. The appeals themselves will waste everyone’s time, unless the defendant can prove (beyond reasonable doubt) that the prior ruling was clearly wrong (not just ambiguous), or the plaintiff can prove the defendant’s new argument is frivolous (or clearly the same as their old argument).

My basic reasoning is: the more the law is enforced by letter, the more it can be sidestepped (broken in spirit). The more it’s enforced by spirit, the more susceptible it is to corruption, and corruption in the law is more dangerous than corruption in other institutions. If the law is only enforced in obvious cases, both issues are reduced: the letter is too unambiguous to sidestep, corrupting the spirit would be too obvious. Meanwhile, other institutions with softer enforcement can counter non-obvious infractions, either by making them obvious to the law, or using ambiguity to their advantage. The law (by itself) doesn’t prevent crime, so at least some of these other institutions are necessary anyways.

I forget how weakly-qualified immunity is justified.

I understand why state officials should have more ability to use violence than citizens, including ability to make reasonable mistakes (i.e. immunity). But there should be a limit for egregious (either intentional or unacceptably incompetent) mistakes. Why is the limit so high?

We have too much petty crime, yet qualified immunity seems to boost it, by shielding cops from punishment for not handling it. I also have the impression that there are many regions where most people don’t respect their cops. In these ways, reducing qualified immunity would increase enforcement, and (by making cops more respected by locals) improve job conditions and morale.

The college system desperately needs reform, but capping As is a bad solution which can make it even worse.

I took a class where As were capped via a downward curve. If you knew the material, but made a few careless mistakes or struggled early in the semester, you’d end with a bad grade.

Another class, the material was easy, but the exams were difficult in ways that would never matter in the real world, like convoluted phrasing.

In both cases, giving everyone As would’ve better reflected their competence, and not caused unnecessary stress.

See also: the UK trying to eliminate juries ("in cases where a convicted defendant would be imprisoned for up to three years").


I'm generally in favor of the jury. It's "the worst system except for all others that have been tried": other than a few high-profile cases (like OJ Simpson), it seems most juries reach reasonable conclusions. A jury is harder to corrupt than a judge, since it's 12 people who are supposed to be ordinary citizens. They tend more lenient (and if the defendant is clearly innocent and the jury convicts them anyways, the judge can and will vacate their ruling), but I'm generally against convicting someone unless they're clearly guilty. Less efficiency isn't a big issue, because most trials are avoided via plea bargain or dropped prosecution.

I also think civil cases against corporations should only need a certain ratio of jurors, because a wrongful conviction is less severe. Criminal cases should be nearly or totally unanimous.

My understanding is that the judge is very powerful in the courtroom: if jurors are bickering or not acting serious, the judge can sanction or replace them, or order a retrial with an entirely new jury. For example, here the judge should've ordered "no costumes".


EDIT after actually reading the article:

In summary, the focused (costume incident) case is about hospitals suing opioid pharmacies. Two jurors wanted to acquit the pharmacies based on the law, the remaining six wanted to convict presumably based on morals. The verdict needed to be unanimous, but apparently the case can be retried (because the parties plan to do so).

Personally, I really wouldn't care either way how this case resolved. In a criminal case that requires unanimity, if jurors are deadlocked between law and feelings, the defendant can only be acquitted, which I generally support whether law or feeling are on their side. A civil case sometimes (in some jurisdictions) doesn't require unanimity.

I think the main issues here are jurors harassing others, violating court rules (using AI), and complaining about other jurors making them feel unsafe. The judge can and should handle these; it seems like they mostly did, and some jurors are just upset that the case didn't end how they wanted.

Perfect Tides and its sequel.

If you don't at least think the protagonist is well-written, who (from some other media) would you consider a well-written character?

Restating and extending this:

In theory I have no qualms with AI art, in practice I’ve yet to see a (knowingly) AI-assisted work I consider decent.

I suspect because (at least to me) originality and relatability (i.e. taste) are important in art; language and diffusion models especially lack both, and they lack the (predictable, easy) control necessary for the artist to infuse theirs.

An artist with taste can make incredible art in B&W, pixel, low-poly, etc. The medium is less important, and those being easy mediums to control, allow the artist to express themselves more accurately.

I think, like how AI for code is another abstraction, AI for art is another medium. Some people disagree, partly because today’s AI is a poor abstraction; likewise, today’s AI is a poor medium, because of the aforementioned lack of control.

I’m sure a big draw towards vaporware is nostalgia for the old Internet.

I went to a public school, and my experience sounds similar.

Everyone went to the same campus.

We had a separate elementary and middle/high school.

Discipline was uniformly strict

Discipline existed, including suspension and expulsion, but wasn't especially strict. Students acted out, but I never saw anything compared to horror stories I hear today.

Flip side of that, no hallpasses, and in high school you didn't even need to ask permission to go to the bathroom, because if you abused the privilege you got in trouble.

There were honors and regulars classes for every core class except religious ed, with real tracking for math in the secondary grades.

We had honors classes, and math branched into calculus, statistics, and discrete.

Starting around 10th grade finals declined in importance, I think when I graduated there was one final exam(Latin) and that's it, we had essays and projects and whatnot for everything else.

Most classes had finals, but they were a small fraction of the overall grade. Some finals were multi-week projects.

Very importantly, your parents could not get you moved to a more rigorous/prestigious class, although sometimes teachers could, and you could transfer credits in from community college if you wanted.

Electives existed and were unsegregated, although a few of them had pre-reqs which kept it to kids in honors in related classes.


I think it was OK, but I wish there was less busywork, higher-level advanced classes, and more computer classes.

An idea from the old thread: instead of testing students and separating them into different schools, each student should learn at their own rate in the same school, via a vast universal online curriculum.

Additionally, instead of one final exam, there should be many interspersed exams. To save proctoring resources, each student could have a random subset of their exams strictly proctored, unknown before exam day.

Students would graduate either at a certain age or after passing certain lessons, whichever comes first. Then they could optionally keep taking lessons and exams (at home online, or by paying to enroll in an adult school). Enough progress in a specific subject would award them a Master’s degree.

Problems? I think it would still be unfair, stressful, unreliable, etc., but better than the current system. Students who don’t have access to coaches would be less disadvantaged, because the online curriculum would cover practically every subject and be designed by the best teachers. With many exams, students may be less stressed and hampered by exam-day sickness, since one bad exam grade is negligible. With many exams and random proctoring, cheating would be harder, even if proctoring was less strict to compensate for frequency.

Maybe they can with online resources like Khan Academy. They’ve gotten much better very recently (the latest improvement being one-on-one LLM tutors), so schools haven’t yet adapted.

Then, teachers only must ensure students follow the rules and answer rare questions, strictly less than they do now.

Substitute “merit” with “ability”.

It’s arguably unfair, but I can’t imagine a solid argument that it’s unjust that doesn’t also justify Harrison Bergeron.

I’m also sure some students would be over and under placed due to conscious and subconscious bias (because not all assignments can be graded on objective criteria, and everybody is biased), which is one reason I leave open the opportunity for a student to manually enroll in a higher class.

I am skeptical of the whole "encouraging children's natural interests instead of formal education" part

The students would still be required to take core subjects, just at different speeds. Although I also think there should be more electives, by having one teacher administering multiple (with the help of online resources).

At the end of the day, this system will benefit the best students the most, and it seems likely that the students would form cliques based on whether they are in the good or the bad class.

Sure, although I imagine there will be some exceptions. Partly because the less academic students may be more “cool”.

The best students are usually from good socioeconomic backgrounds, so this will easily be spun as discrimination and enforcement of the existing social order. Limiting social mobility, putting disadvantaged groups further behind, etc.

Unfortunately yes, even though it’s supposed to be exclusively based on merit.

However, if a non-disruptive student or their parent really wants to be in a class above their level, I think it should happen. If they struggle, some of their assignments should be replaced with those from their actual level and between, to try to prevent them from falling behind, but if they continue to insist they can stay. That may slightly alleviate complaints, because the students in the lower sections are there partly by choice (albeit partly by encouraged default).

I also support allocating extra resources to students from lower socio-economic backgrounds, especially those in upper-level classes.

How about a Montessori-like system, where students progress through classes at their own speed (instead of age)? They would still interact with similarly-aged students in social activities like meals, except disruptive students who are bothering others would be a separate group (who would be assigned some form of therapy).

Newcomb originally specified that Omega would leave Box B empty in the case that you tried to decide by flipping a coin; since this violates algorithm-independence, we can alternatively suppose that Omega can predict coinflips.

You’re right, I misunderstood the problem.

Why do they do that?

Why do people gamble?

Alternatively, they also misunderstand the problem. I wonder if the “practice run” method of predicting their behavior would change their mind.

Note that sometimes the "regulation" isn't from the government, but a parent organization.

For example, sometimes managers assign employees useless tasks to take credit for managing a higher number of employees, since that's the metric they get promoted on. Or to spend their yearly budget so next year's isn't reduced. Or because one of their employees is their boss's incompetent grandson.

When companies become large enough, they become pseudo-governments. A large, poorly-managed organization creates bullshit, regardless of whether it’s public or private.

Unfortunately, I frequently hear tales of managers assigning humans work they knew would ultimately be discarded, to inflate bureaucratic metrics. For example, it's common for organizations with yearly budgets to intentionally waste the entire budget if they wouldn't normally spend it all, because otherwise they'd be allocated less next year, and sometimes they do this by paying employees for unused work.

And unfortunately, I predict at least some of these organizations will replace the efficiency gains from AI with more useless emails, software, etc.

Fortunately, there are plenty of good use-cases for widely-available AI inference. OTOH example, people could create more immersive game worlds with AI NPCs, and use any extra inference for more detailed world simulation.

Computer speed has exponentially increased for decades. Developers have found plenty of bad use-cases for this extra speed (e.g. advertisements), but lots of good ones (e.g. easier programming languages, better graphics, and ML).

My interpretation:

Free will is indistinguishable from randomness, and your brain has some randomness. The alien understands your personality, but can't predict randomness. For example, maybe immediately before the experiment, they cloned you and ran a perceptually identical experiment; then, if your clone picked both boxes in the practice run, the alien didn’t fill the opaque box for the real run.

You can win $1,001,000, but only if you're lucky. For example, let's say you have a 50% chance of choosing both boxes. Then the alien has a 50% chance of filling the opaque box. You have a 25% chance of winning $1,001,000...but a 25% chance of winning $0, and 25% chance of winning $1,000.

You can't trick the alien: if you're more likely to choose both boxes, the alien is less likely to fill the opaque box. Formally, if you with probability p pick both boxes, the alien with probability 1 - p fills the opaque box. Imagine your clone, in the same perceived surroundings, with your same strategy.

Alien / You One box Both boxes
Empty opaque box $0 * (1 - p)p $1,000 * p^2
Full opaque box $1,000,000 * (1 - p)^2 $1,001,000 * p(1 - p)

If the experiment was repeated ∞ times, on average you'd win $1000p^2 + $1000000(1 - p)^2 + $1001000p(1 - p) = $1000000 - $999000p; increasing p strictly decreases your average win. The statistically optimal strategy is to always pick one box.

Chamath, Marc Andreessen, Elon Musk, are a subset of the successful population.

Other successful people have good opinions, and/or a small online presence. The bad ones get most publicity.

Steve Wozniak may be a latter example.

What do you think about cyberlibertarianism?

Cyberlibertarianism is exactly what it sounds like: the belief that the internet should be fully unrestricted and ungoverned. The idea coalesced in the 1990s when the consensus on tech was far more optimistic.

I think it's a beautiful, unattainable ideal. It symbolizes (more than libertarianism) a broader absolute freedom and physical transcendence, to realize whatever you dream. But in reality, absolute freedom is impossible, power hierarchies are inevitable, and the internet is a physical construct that can be seized (on the other end of the spectrum, individuals and companies bypass without consequence internet restrictions like copyright, even in repressive countries via complex VPN setups). Intersectionally, the internet has led to good (e.g. long-distance communication with friends/family) and bad (e.g. asociality and toxicity from social media); should it be as unregulated as today if individuals and groups won't stop themselves from negative spirals (which may anyways lead to future violence and restrictions)?

Cyberlibertarianism's Origins

The ideas of cyberlibertarianism have been described in Cyberspace and the American Dream: A Magna Carta for the Knowledge Age (Esther Dyson, George Gilder, George Keyworth, Alvin Toffler, 1994) and A Declaration of the Independence of Cyberspace (John Perry Barlow, 1996).

Basically to summarize the latter, it begins with

Governments of the Industrial World, you weary giants of flesh and steel, I come from Cyberspace, the new home of Mind. On behalf of the future, I ask you of the past to leave us alone. You are not welcome among us. You have no sovereignty where we gather.

and includes statements like

We are creating a world where anyone, anywhere may express his or her beliefs, no matter how singular, without fear of being coerced into silence or conformity.

Langdon Winner's "foresight"

The actual term "cyberlibertarianism" first appeared in Cyberlibertarian Myths And The Prospects For Community (Langdon Winner, 1997).

I think this is a good read. In a sea of cyberlibertarian idealism and optimism, Winner was realistic and pessimistic. He actually defines cyberlibertarianism in detail, then predicts how it will be adopted and warped, in practice, with outcomes.

Specifically, Winner defines cyberlibertarianism by breaking it into four sub-beliefs:

  • Technological determinism: technology will rapidly, radically, inevitably reshape society
  • Radical individualism: technology will enable freedom and self-fulfillment, unencumbered by "inherited structures" like social obligation, money, and government
  • Free market, specifically from the Chicago school of economics. The Magna Carta argues for "property rights in cyberspace" by quoting Ayn Rand's The Property Status of Airwaves
  • Abundance and liberal democracy. As technology keeps getting faster and cheaper, digital scarcity won't exist. As the internet connects people, they'll get along better. As the internet grants everyone access to vast literature and free debate, societies will become more democratic

Then he pivots to realism with this (IMO) excellent paragraph

As is generally true of ideologies, this framework of thought serves to both illuminate and obscure. It certainly illuminates the desires and intentions of those who see themselves on the cutting edge of world-historical change in Silicon Valley, Seattle and other high tech centers. More specifically, it illuminates what are ultimately power fantasies that involve radical self-tranformation and the reinvention of society in directions assumed to be entirely favorable. But this ideology obfuscates a great many basic changes that underlie the creation of new practices, relations and institutions as digital technology and social life are increasingly woven together.

In the remainder, Winner successfully predicts that cyberlibertarian dogma will lead to:

  • Anticompetitive monopolies
    • Ultimately restricting individuals' freedom in the absence of government restrictions
    • Controlling the distribution of information, therefore influencing the zeitgeist, therefore influencing democracies
  • Dissolution or mutation of existing institutions in ways that aren't entirely positive
    • Replacement of physical communities with online ones, which are inadequate
      • Because they split into echo chambers
      • Because a minority of users carry the majority of discussion
    • Replacement of local stores with depersonalized online (centralized) ones

Winner didn't have extreme foresight, just observation. These "predictions" had already began: the television industry (e.g. CNN) was already large and influencing the zeitgeist to further its interests, toxic online communities had already started forming (e.g. Usenet), and local stores were already being replaced (e.g. by Amazon). Winner also looked at historical literature on philosophy, economics, and politics.

The Intolerable Hypocrisy of Cyberlibertarianism

This blogpost showed up on Hacker News and inspired my post.

tl;dr: the author of this rambling blogpost describes the evolution of the internet under cyberlibertarianism (the dominant viewpoint in its early years), then criticizes cyberlibertarianism using the problems of today's internet.

I don't really like it: it's full of ad hominems, meaningless analogies, and overconfident claims (especially about other's thoughts). But it's somewhat informative, and I agree with the underlying ideas: cyberlibertarianism is naively optimistic, hence today's internet has failed to reach its full expectations.

Generally, a culture of distrust towards women that would mandate (or heavily incentive) paternity tests is far more liable to result in arguments and general suspicion, even in otherwise harmonious relationships once the issue is raised as it necessarily implies a fear on the husband’s behalf on the trustworthiness and fidelity of his wife.

The opposite: once the husband asks for a paternity test, there's already an argument and suspicion, and the only way it would be resolved is if the test confirms they're the father.

The child as such would benefit the most from being raised within such a home, even if in reality the child is actually not biologically related to one of its guardians.

I agree that the father should stay. But I argue that forcing him to pay child support is actually counterproductive here.

I follow A.Shipwright and one other artist, and my feed is filled with art (IMO only mediocre but I’m picky).

Scenario: A person roles into the hospital with a gunshot wound to the [organ that can be lived without]. The shooter has the same blood type as the victim.

Question: Is it ok to take the organ from the shooter to replace the organ of the wounded person?

Utilitarian: You can take the organ from the healthy person in the waiting room, they are easier to find and might have been the shooter anyways.

@stoatherd provided a good argument against this reasoning. In such a society, healthy people would avoid hospitals.

Likewise, the current situation discourages the adoptive father from supporting the child or even himself.

I actually think the adoptive father should be encouraged to raise the child as his own, key word encouraged. Coercion activates the innate human desire (common in men) to resist. No penalty for abandoning the child makes the father feel autonomous and "in control" when he raises them anyways, even if there's a (e.g. social) reward. (It also gives him control over the wife, but she still has the option of leaving him; if you think it's a bad or unfair outcome, can you think of a better or fairer one?)

But why not the biological father?

(And if he’s dead/incapable, maybe the state has to pay, but that’s the case when somebody isn’t tricked. Or that can be an exception, since the adoptive father would have less reason to envy him, although I still think it’s bad)