virgildotcodes

Proof linked here - https://cdn.openai.com/pdf/04d1d1e4-bc75-476a-97cf-49055cd98...

Prompt used - https://cdn.openai.com/pdf/04d1d1e4-bc75-476a-97cf-49055cd98...

Lean formalization of the proof - https://github.com/openai/cdc-lean

Do you think it will be a net social benefit for people to be taught 2 + 2 = 4 and 2 + 2 = 5 with equal weight, and let them come to their own conclusions?

How about if the person who owns the educational institution puts their thumb on the 2 + 2 = 5 side of the balance for their own ends?

I'm not sure if you actually haven't checked it since, but will give you the benefit of the doubt.

Accounts pushing white supremacy, the reversion of women's rights, hatred towards other on the basis of their race or religion, climate change denial, denial of science and promotion of pseudoscience, etc etc. are heavily promoted across the platform and get millions of engagements.

If you create a new account, the majority of the accounts you are shown and suggested to follow will be those pushing the above.

They've switched to a model of paying their users for engagement, which naturally encourages users to post the most engagement bait they can, which tends to be inflammatory and utterly lacking in depth or nuance.

I think you're speeding past the word "average" in the sentence. I'd argue that current frontier models already exceed the abilities of average humans across the majority of tasks you can do on a computer, although you might be able to argue that they tend to be a bit slower?

That latter part is debatable though - have you seen a non-technical person try to figure out something new on a computer?

Mine was not that severe. I would get mildly nauseous during normal commercial airplane turbulence, although very nauseous in small planes and the worst nausea of my life after a glider session, very seasick to the point of being debilitated, and then probably a 6 or 7 out of 10 when reading in the car for more than a few minutes.

Definitely worth a shot IMO.

I had motion sickness for the first ~38 years of my life, and had it cured.

Posting here because I went the majority of my life without knowing that such a treatment existed.

About a year ago I flew to the University of North Dakota and spent an hour a day over 3 days sitting in a simulator that would spin faster and faster while the facilitator talked to me over the headset and kept me distracted.

Was nowhere near as bad as I'd feared, honestly maybe a 3/10 on the discomfort/nausea scale.

Ever since then, motion sickness completely gone. I used to despise boats, calling them mankind's cruelest invention. Taken many flights, reading car trips, hot humid boat rides in rough seas in southeast asia since and... nothing.

Would 100% recommend. I think you can find the program at the bottom of this page, it's $400:

https://www.aerospacephysiology.com/

> There is no evidence it was highly coordinated and many reasons to believe it’s emergent. Cable TV and then social media created systems competing for attention. The content originators became increasingly decentralized, increasing both diversity and ruthlessness.

Hmm, not sure how you could conclude this given the abundance of evidence regarding the activities and influence of people like the Koch Brothers, Rupert Murdoch, Peter Thiel, the fossil fuel industry in terms of the global warming discourse, the general corporate and wealthy forces shaping Republican/Democrat policy over the last ~50 years.

The media influences alone that fuel the sensationalization of these issues are transparent, as are the threads that bind these media groups and those in power over them.

Look at what's happening to CBS and will soon happen to CNN due to the Paramount merger and the Ellisons/Bari Weiss for example.

> Historically untrue.

What? In what part of history have disorganized masses shown themselves to be powerful against "the intentional propaganda efforts made by the most powerful actors in society spanning decades" that I'm referring to?

Almost by definition the successful grassroots movements of the past that have created change were organized, no? I also don't believe there's ever been as effective a media (social and conventional) apparatus in human history as we've had the last half century.

> Starving, uneducated masses are easy to repress. Distracted masses only so long as they look away. And I think we’re seeing signs the American voter isn’t looking away.

I mean, this conversation started over the culture war bullshit that seems to have about as good a grip on Americans' attention as ever, although I agree that the material economic conditions are degrading so badly that they are more and more becoming the priority consideration.

That said, channeling that anger towards scapegoats like immigrants or jews etc is an old and effective playbook and I don't see why we wouldn't call that distraction.

I mean, I also have times where I find myself blaming people for being so stupid.

That said, you have to realize that this has been a very intentional propaganda effort by the most powerful actors in society spanning decades. Disorganized masses are largely powerless against that sort of effort and the outcomes are predictable.

In terms of why bird brains would be exceptionally efficient for their volume (and I assume by extension, mass), would be that weight is at a premium for them.

Replying to a sardonic comment - believe it or not, list.

About a year back I tried to build a few latency-sensitive products on Groq, and found the response times would vary so wildly that it just about erased any benefit in their throughput. Is that still this case?

Max seems to me to be notably better than the others.

Idk, I’ve seen Deliverance, plus all those horror movies that start out on some desolate road in the woods at night. Seems likely a fair depiction of country living.

No thank you, I’ll stay in Manhattan and not get kidnapped and murdered by monsters tyvm.

Honestly, even if you join the field audit - who's to say they didn't plan everything in advance and that these aren't all crisis actors?

Who's to say they didn't drug you, and you performed the "field audit" on a psychedelic trip from within a padded cell controlled by the nefarious cabal of do-gooders?

Frankly, if you believe your own eyes and ears aren't lying to you, you may be a touch naive.

There are many ways of knowing if you actually want to find out. Third party auditors, evaluations, reputation… just… reading reports and looking at evidence of field work.

Too many people use this as an excuse to throw up their hands and say “well I guess I have no choice but to keep all my money for myself and donate nothing, damn, darn, totally not the conclusion I was secretly hoping to reach, not at all, oh man, damn.”

What, lol? You want me to write a comprehensive account of each country in Africa that had foreign interventions and enumerate them? What’s the character limit for comments on HN, how many full comments do you think I’d need to get to something approaching comprehensiveness?

Zimbabwe, we’re talking about the one that had China, USSR and SA providing weapons and training up until the end of Rhodesia around 1980? Then IMF/World Bank imposed market liberalization in the 90s, then sanctions from 2002-2024?

Truly, can’t understand why they’d be in bad shape, must be all their own fault. My brilliant white brain thinks it must be something genetic, if you know what I mean, nudge nudge wink wink.

Are you under the impression that western activity in Africa ceased with the end of colonialism? No fomented coups, conflicts, revolutions, arms and funding for rebel groups, continuous bribes and support to corrupt government officials to secure the flow of oil, minerals, etc. out of those countries into western hands? No proxy wars between the west and the USSR?

Read more about the history of the continent.

Better 20% of your money reaches a starving child than 0%.

EU debt to GDP is ~80% avg across member states, US is ~123%.

US CPI 3.8% as of April, EU HICP 3.2% as of April.

IMF projects 2.3% growth US vs 1.1% growth EU in 2026.

Not sure I would say the EU is in a borderline catastrophic economic situation and the US is not?

If both are in catastrophic situations (it seems one could fairly make that assessment in either case), then again we’re back to “all societies are unstable, would we like to be pragmatic and compassionate during our brief moment of stability or vindictive and self destructive?”

Yeah for sure, speaking purely on the American common framing of big L Liberals as akin to social liberals rather than classical/economic liberals.

While we’re at it I, too, would rather be given $xx billion a month and a pony than be in a union and only get a $350k bonus.

I think I am making a shrewd decision here if my math and equine knowledge holds.

We’ve gotta add American Liberals, majority of Democratic Party to the list. The Sanders faction is unfortunately not yet the prevailing force.

Maybe that’s still what’s happening, they’re talking about average after all :)

> 9 men hunt and 1 man eats free, so the 9 men are carrying the weight of the 1.

Updated the quote to the historically accurate 99 vs 1.

> This system is inherently unstable and unsustainable.

The countries cited are extremely stable. Arguably far more stable than the US.

That said, we can bring in the rest of Western Europe if 5 countries aren't enough of an example. They have union participation rates between 10% and 50%, median around 20%. The thing is, they have much larger proportions of their workforces covered by collective bargaining agreements - France for example is at 10% union participation yet 98% of labor covered by bargaining agreements.

Western Europe and the Nordics combined = ~400 million people, bigger population than the US, and far more diverse, so the common refrain of "small homogenous population" doesn't hold up to scrutiny.

Of course, all societies so far have eventually been unstable.

We can just choose whether our unstable society will be a vindictive one that prioritizes punishing wrongdoers over the wellbeing of the whole, or a pragmatic and (as a nice bonus) compassionate one which prioritizes the wellbeing of the whole over a puritanical urge to purge the unworthy.

This doesn’t seem to be a universal rule at all, but smells more of a boogeyman promulgated within US society.

The nordics are anywhere from 50% - 90% of all labor unionized and they absolutely destroy the US on every standard of living metric.

It seems to me a case that echoes “better to let 99 guilty men go free than to execute an innocent man”. Of course, in this case, the ratios are actually reversed. Should we execute 99 innocent men to make sure that 1 guilty guy gets punished?

There will be some free riders, just like there will be some welfare queens, just like there will be some voter fraud.

That said, these cases represent a vanishingly small minority of the whole, and the cure is far worse than the disease.

Yeah, it’s like living in an unsustainable society whose luxuries you enjoy are entirely predicated on the destruction of the natural world, the enslavement and abuse of your fellow human beings, and the death and torture of billions of other sentient beings annually.

If you’re honest, you know it’s evil, but it’s pretty undeniable that all the affordances this provides us are useful (to the beneficiaries) and that we all contribute to it daily.

Eh, there’s definitely some value in understanding for yourself via experience which models are actually good for which use cases. The benchmarks are unreliable imo, and as I’ve interviewed developers who don’t really use AI, they say things like how they don’t think the (free versions of) copilot or ChatGPT (requests routed to their cheapest models) don’t seem very good. Totally out of touch with the capabilities of the leading models and harnesses.

I think the real argument is just staying employable. Companies are expecting faster and faster turnaround, and it’s simply becoming impossible to meet these deadlines with fully handwritten code. Even before outright mandates on AI usage. If you refuse to use AI, they’ll bring on someone who will, whether or not the quality drops, high quality code is not the primary goal of the business.

Dogshit, hideous vibe coded messes are launching daily and reaching 6-7+ figure ARRs while leaking customer data. Nobody cares in this environment.

If you’re a freelancer it’s even worse, the expectations are that producing a fully functional moderately complex app shouldn’t take a single person more than a couple months, and ideally one.

Expectation for a contractor coming into an enterprise codebase that’s been running for 11 years with a dozen+ internal devs and a mishmash of legacy and new tech -> they want you to implement a totally new feature which touches half a dozen systems in the app ready to demo in 6 weeks and launch to the public in 8.

> Nobody knows where it works

A large percentage of code being written today is AI generated. If none of it worked it wouldn’t be so.

> This is where it definitely does not work.

The person said it’s clearly working for their friends’ purposes. That means it works.

Ratios > absolute numbers, of course.

If only 4 people die violent deaths out of a total population of 5, that’s an extremely violent population to be a part of.

If 8 people die violent deaths out of a total population of 100,000. That’s a much more peaceful population to be a part of, despite the greater number of absolute deaths.

Yeah I'm not sure what the exact context of the statement is.

I am absolutely certain that we have not already discovered let alone implemented the best possible learning algorithms. Humans have had more time to evolve, there's a great chance that we do learn more efficiently, and have developed specialized brains that are primed to learning things like how to navigate the physical world on planet Earth as bipeds.

That said, to say that we operate with less training data is just ignoring the reality of all the data we're training on at all times.

If we were to model in lossless fidelity what humans are capable of seeing, hearing, smelling, tasting, feeling, thinking consciously and subconsciously etc. essentially all the data flowing through our minds that we are constantly training on every moment of every day, even while we sleep/are unconscious, what sort of bitrate do you think would be required?

Modern LLMs train on datasets in the what, tens of terabytes in size? Let's call it 100 TB.

I would imagine that to losslessly reproduce the full suite of human sensory data (whatever that means for things like taste, touch, smell) would require a bitrate that hits that 100 TB total relatively quickly?