komali2

Proliferation of data doesn't mean proliferation of the thing itself.

There's a massive barrier between knowing how to build a nuke and actually building it.

Also, isn't proliferation the basis for MAD? Well if we believe in that, naturally it means that the world would be safest if every individual had their own nuke ;)

> AI allows anybody to enact essentially anything

No it doesn't. Where is your evidence for this?

> While "some chatbot" isn't the problem, general intelligence superior to humans absolutely is.

This doesn't exist. Why are you pretending it does?

It came and went from that for a loooong time though. They were still calling us Yankees in Europe well into the 50s.

They can't, which is why I take the average of all results, and then have a final sanity check that considers the context, so if the text contains a bunch of references to "The Disposessed," and two models record "Shevek thinks Sabul is a propertarian," and three record "Shrek thinks Sabul is a propertarian," the final sanity check records Shevek.

I don't think this is viable for critical record OCR. I think the only way to do that is one pass with a frontier model and then a mechanical turk manual review passthrough with good compensation that allows for a slow and methodical approach. Plus of course much better scanning than a phone camera.

I basically kept rabbit holing this problem and finally settled on "80% and done is better than sitting on this problem for 4 years waiting to have time and equipment for a 99.99% solution." Crank the pictures between Claude code sessions, run the local LLMs when I'm asleep, done, now I can free text search years of journals plus I have photo backups now finally of them.

Sounds kinda like exactly what I set up in the repo I linked in the original comment heh... Minus the openrouter calls.

That's the workflow in the GitHub repo I linked, I guess the person you're replying to didn't look. You could just change the calls to be to frontier models instead of local ones if you want.

> But here we are: so far the story is pretty much 100% "fortune favors the bold", and Meta is leaning into it.

Been thinking about this lately when I saw some mathematicians talking about more AI regulation after the millennial prize thing.

Move fast break things, fortune favors the bold, breaking rules doesn't matter if you win, etc silicon valley ideologies to me seem to be domination ideologies. Uber may toy with playing nice with taxi law until it's too annoying, then they shift to getting as many people as possible to have their commute, convenience, or livelihood to depend on it, then they can turn around to regulators with their finger on their chin and say "oopsies, I guess everyone needs us now, better consider that in your regulatory efforts!"

It's the brunch equivalent of changing the restaurant the night before to one you prefer by DMing everyone in the group chat with "we decided to change to antipodean because dave had a breakup with one of the waiters at french twist," and then when everyone arrives and Dave says "I've never dated a waiter in my life," you say "well we're all already here so..."

Anyway, it seems pointless to try to regulate this culture out of these kinds of companies or really this entire scene. Any law is now just an obstacle, and besides, various government actors seem to be angling for industry retirement positions, so passing useful regulation seems basically impossible.

Imo two can play at this game. When law fails to serve the interests of the people or is even weaponized by the interests of Capital against us, what's the point of it? I'm interested in things like LLM distillation and open weight models that take the winds of profit out of the sails of these "disruptors" to make their activities have no payoff. Wonder what other things we can do that would make someone ask "why would I send all my data to Facebook when I could instead just..." Pirating, free culture, leaks, hell outright theft, if we're going to disrupt things, why not DISRUPT things?

> I do wish I could figure out some sort of recommendation system to find new magic, though. Obviously not to steam new/unowned songs, just to point me in a direction next time I go to the store.

Navidrome can scrobble to https://listenbrainz.org/ which has a recommendations mechanism

Or set up "soulsync" and just use the library scanning, scrobbling, and recommendations / artist watching features if you don't want to pirate the music as well.

In Taiwan they mix "positive" and "negative" indication signs.

So for example, you'll see a sign indicating cars are OK for a freeway. That means, heavy motorcycles and scooters are NOT ok. You can only take your heavy motorcycle on it if it also has a heavy motorcycle sign. But SOMETIMES they'll use the slightly different motorcycle sign which means scooters ok. Well, if it's ok for scooters, it must be ok for heavy motorcycles as well, right? Wrong! Heavy motorcycles must then take the car lane, if there's a lane marked for scooters.

Also sometimes there's a road marked no cars! Scooters ok. Heavy motorcycles? Who knows!

As for pedestrians: lol just die

Cars ok, heavy motorcycles ok scooters not ok: https://taipeiscooterrentals.com/wp-content/uploads/2026/01/...

Cars ok, heavy motorcycles ok, scooters not ok by implication: https://blog.calebjay.com/posts/rent-motorcycle-taiwan/cars_...

Quick, google is telling you you need to cross this river, are you allowed on this road or not??? https://www.google.com/maps/@25.0756527,121.6435485,3a,75y,9...

Time to head back to Taipei proper after hanging at e7play, let's cross the river... oh wait, there's 600 cars behind you blaring their horns as you enter this intersection, which way are you going? Can you go straight?? Who knows! https://www.google.com/maps/@25.0722621,121.4934581,3a,75y,4...

I've had plenty of friends bicycle into places they apparently aren't allowed to. There's no sign that says no, but the bicycle isn't allowed by the implication of some combination of allowed-vehicle signs that are lacking a bicycle-allowed sign.

I set up an OCR flow using local models on all my many tens of journals stretching back the last 30 years.

I would say it's about 80% accurate, which means it's missing enough key words to make a lot of it uselessly unintelligible. I can easily compare the images against text I turn up in a grep which is nice if I'm looking for something.

Allegedly Claude set up a system for retraining for my handwriting, but it would require me to manually revise several hundred pages by hand so I don't think I'll ever do it.

https://github.com/508-dev/journal-ocr

Well, that's what I meant by my European coworker - perhaps his country has found a socialist balancing point that maintains a level of capitalism that somehow doesn't tip into power cascades concentrating power around Capital. IMO that wouldn't be a liberal capitalist society though, that would be a socialist-capitalist society, which perhaps the Netherlands and Nordic countries are, I'm not too familiar. With the level of privatization striking Canada, I think it'll end up like the USA.

If we're gonna take a look at places, where I'm really interested in watching is the PRC, Cuba, and Vietnam. Capitalism managed to beat the Soviet Union, will it to any other remainders from the era, or will the conversion to state capitalism create a strange equilibrium? In any case it's not like the PRC is doing much better in terms of worker's rights, you still have to justify your existence through labor there more than you do in the aforementioned Nordic countries.

Personally I'm thinking the solution will be establishing parallel forms of mutual aid and organization with the goal of making the State and capitalism obsolete and moot. Why would I buy food when my neighbors and I grow and share more than enough? A stretch goal but I don't see any other path. At least trying to reach it makes for lots of fun projects like gardening, homelabbing, setting up federated software, etc.

Further reading: Walkaway by Cory Doctorow, The Conquest of Bread by Peter Kropotkin, The Anarchist Cookbook by Keith McHenry (and look into Food not Bombs, Food not Lawns etc while you're at it), and various solarpunk copium like "A Half-Built Garden" by Ruthanna Emrys.

I wrote it, and I invite you to tell me why I'm wrong - genuinely. I'm here for the conversation. I don't know how else to write on something other than authoritatively. If I'm wrong on the internet, people will definitely tell me!

In person, I'll hedge of course, it's a natural part of conversing and inviting people to share.

Arguing whether someone is permitted to speak on a subject sounds like it would take as much effort as simply discussing the subject itself.

How did you identify how businesses are worker-owned? Curious how it compares to other verification methods e.g. what my co-op had to go through to get a .coop domain as determined by these guys: https://identity.coop/policies/ Wasn't too hard or anything, just wondering what people are looking for to verify co-op-ness

For categories, I guess software engineering would go under "other?" Did you find many services co-ops? I'm aware of a couple other software engineering co-ops/collectives in Europe but none in the USA. Seems to be a bit barren for some reason.

> I don’t want a vote on whether or not my bakery puts sesame seeds on the bread. I just want to buy a loaf of bread thanks.

You have an illusion of choice right now between sesame or non sesame bread through the low-resolution "voting" mechanism of the market, which doesn't actually consider your desires when counting votes, merely the profitability between your desires and a supplier's margin. If you like sesame, but if it's not profitable enough, too bad, no sesame for you. "They can charge more," "someone will fill the niche," yes we've been told that about capitalism for the last century, yet it's never been true, so I guess it was a lie. Think of all the products people want that we just can't get anymore because some big org decided we can't have it, or better yet, think of how nearly monopolized industries have leveraged their power to move markets towards more profitable directions. Clothes - big, baggy, unicolor blobs. Phones - black squares.

It's not only more ethical for breadmakers to he democratic, it's simply more efficient in terms of delivering what the people want and need. Markets are a low resolution, exploitable mechanism that can lead to contrary incentives. Better to just cast a vote.

By the way I would add that co-ops are more ethical because if it's set up correctly, nobody is profiting from your labor more than yourself. In a traditional org, the vast majority of your productive profit is captured and used in a way you have no say - lately mostly to line the pockets of shareholders and a captured boys club of nepo MBA types that often don't even have that much experience in your industry. In a co-op, you the worker decide what happens with the profit margin on top of your labor. You can use it to have more money for yourself (and your coworkers), or you can invest it more intelligently in the business than an exec thousands of miles away in the private jet your profit margin pays for. This also works out better for your customers and users.

I've been thinking about this recently: I think our economy is fake.

Restaurants are expensive lately, but still, as a former restaurant owner, even the expensive 10$ takeout burrito I don't think would be possible without the great many laws the supply chain breaks. Undocumented labor from the farms through to the restaurants, for starters. Then the fact that the equipment we use is built in places that don't have our labor or health and safety laws, which lets those places manufacture for much cheaper prices, which feels like cheating.

If we actually had a level playing field and didn't break laws, I don't think most of the luxuries we enjoy could actually exist at anywhere near current prices (and thus not at all).

A lot of co-op talk around this thread, if anyone else has worked in / founded co-ops, I would love to chat, email in my profile. I founded one about 5 years ago and have been kinda shooting from the hip, would love to trade ideas.

I founded a software engineering co-op. We're always interested in finding new members. https://508.dev

Right now our main revenue is through subcontracting, and admittedly we haven't pulled a fresh gig in a couple months, mostly because most of us are working a gig which limits our ability to BD - a limitation we've been kicking around how to resolve for ages.

We want to come up with some alternative revenue streams that aren't directly selling our labor. I soft launched one recently (a mock interview thing that I'm not quite ready to link completely publicly) but I'm also really interested in ideas other people have.

My goal for the co-op has been to make it a greater than sum of parts actualizer for each of our potentials. So far it's been pretty good on that front but for me the big win is gonna be if we can get a revenue generating project out.

Anyway, rambling. Feel free to reply with questions or email me (email is better since I'm trying to stay off HN due to internet comment addiction). https://508.dev/join has more info on joining.

Random fun fact: all our internal SaaS is FOSS and self hosted! Excluding discord. Matrix didn't work well for us.

You've constructed a strawman of the field of psychiatry and are arguing against it. I assume you mean the DSM5? It's basically a dictionary. Is the entirety of the English language and all things that happen in it represented by the Oxford English Dictionary? Read it and you've already read Shakespeare AND Google's transformers white paper!

You're ignoring mountains of studies with empirical evidence for the efficacy of various psychiatric treatments.

You think you're being empirical because you can't find a chain between a molecular chemical reaction and a given psychiatric disorder and its treatment, like the link between what a bacterial infection does to the body and then what the antibiotics do to the bacteria. I encounter this a lot in engineers, who are used to working in relatively simple, deterministic systems. If you want to blow your mind, go look up the data storage capacity of the DNA in a skin cell. Have trouble parsing a vibe coded PR, try parsing the human genome.

You're mistaken that what you perceive as subjective is related only to the field of psychiatry. The human body is too complicated to model accurately, so for example nutrition is also a highly "subjective" field leaning heavily on empirical data gleaned through studies, rather than a deterministic perfect link between the mechanisms of a tummy ache and the chemical composition of the food that might be causing it (actually it might be the food triggering a reaction in the gut bacteria which is... OR it could be psychosomatic OR it could be unrelated to nutrition at all and be ...)

If you need determinism and full system explanations to understand things, that's great, you chose the right field: computer science is easy to map. Down to the last bit, down to the logic gates in the CPU, you can describe everything, you can determine the exact cause of a bug or behavior. Biology and medicine aren't like that, especially not psychiatry. The human brain is widely considered the most complicated thing in the universe (that we're aware of). It can't be modeled or simulated with the degree of accuracy of a CPU. It's nondeterministic and unpredictable. The best we can do is do large aggregate studies and start picking apart the patterns and labeling them. It's definitely still an early science but that doesn't invalidate the treatments which studies show do work.

It's definitely weird the "power" of the human brain on human biology, but the evidence is overwhelming. Studies on psychosomatic symptoms, studies on placebo, studies on therapy treatments, if the effects weren't real, they wouldn't show up so consistently in populations across the world.

So again, I get it, you don't fully comprehend the field so you dismiss it. However I again remind you: your dismissals aren't convincing, they just expose your ignorance of this topic. Go on believing whatever you want, but I guess I'd recommend keeping this confident ignorance to yourself.

What I suspect is you specifically believe some subset of psychological disorders are "fake," am I right? If I had to guess, for people with your opinion, it's usually, depression, ADHD, maybe autism, and maybe gender dysphoria? Lots of other disorders listed. Surely you don't think schizophrenia is fake? Tourettes? PTSD? If you're gonna say the whole book is fake you have a lot of studies you need to dismantle.

I think the idea is that since the inventers of bopomofo were exposed to other alphabets, it's still considered a descendant alphabet. I usually think of descendant as something that visibly manifests its ancestry, so for example modern traditional characters look somewhat like the earliest Chinese characters, or, all romance languages sharing some sounds or even words. So maybe we need a different way to describe things like wheels and alphabets.

I'm not sure I understand your point, sorry. What do you mean?

> "let's all get together and build some free software collaboratively".

G0v hackathons in Taiwan are still this, at the end everyone presents what they worked on and there's no judging or anything. Some of the projects have been going for years.

There was a hackathon two weeks ago, you can see all the videos from all the demos here https://m.youtube.com/@g0vTW

They happen every two months. Some people have started g0v chapters abroad, maybe you could consider it for your region!

No, since it's a silent failure, it's not plausible. We have to assume all results we get are the actual model performance, because, it's the actual model performance as we understand it.

Someone trying to solve similar problems will have similar results if the "silent failure" applies consistently in aggregate. So, this is the model's performance.

What makes you so sure? There's been massively successful government funded and run projects before. Soviets beat the Americans to space, after all.

So, what kind of problems are you having it try to solve?

Sorry to belabor this but it's basically pointless saying you have nuts it can't crack without showing us the nuts.

> Okay, okay, that whole idea of 'communism' is just as silly

But the communists are smarter than the free world, at least the Americans, which I take it to represent the peak of capitalism and western liberal traditions. The PRC literacy rate is 96.67%, the USA is 79%. In 1937 the Soviet literacy rate was 75%, the USA appears to have been 97% literate then? [1] so somehow the Americans have become nearly as illiterate as a recently industrialized nation of peasants.

Ah, apparently late 70s literacy rate in the Soviet Union was 99.7%. [2].

> but nobody believed in communism either

I really recommend reading some Mao/Stalin era publications, not just from folks like Lenin but general notes from standing committees or national congresses. Even today the national congress of the PRC will get into all sorts of debates about communism. I don't believe their current system is socialist, but they sure do, and there's no doubt that there were a lot of true believers around Mao. I strongly doubt the cultural revolution or red guard could have happened without a lot of peasants genuinely believing in the cause.

[1] https://nces.ed.gov/naal/lit_history.asp#illiteracy

[2] https://www.degruyterbrill.com/document/doi/10.1515/sem-2016...

Edit: Cuba's literacy rate is 98% lol https://data.worldbank.org/indicator/SE.ADT.LITR.ZS?location...

The things you describe may be problems in your country, but they're not universal.

Something or set of things must specifically be going wrong wherever you live. It would probably be interesting to identify what.

> Having to work is less agency than not having to work.

Both are working. One is getting money that they get to decide what to do with, the other (the wife) doesn't get paid but gets room and board but very little autonomy - for example, they were expected to deliver sexual gratification on demand, mood or no.

It's honestly surprising to me how people seem to still not be aware of the amount of labor women at home do, especially in previous eras. Read any old book, you'll find wives cooking every single meal solo for every single person in the household, managing every aspect of the kids' lives, functioning as a secretary for the husband, cleaning the entire house, often doing the yardwork, managing the social calendar, all while keeping up appearances so they're an attractive wife. They worked way more hours than their husbands.

> I'm not sure you are aware that China has monitoring operations for its citizens outside China.

Sure, you and I know that, but most PRC people don't really believe in that or are aware. At least nobody I've met.

It's starting to feel like ethno-nationalism is the answer.

See: the PRC. Support for surveillance is allegedly high. Anecdotally, talking to PRC citizens in circumstances where they don't need to worry about said surveillance (e.g., when they're vacationing in Japan and I want to pester someone and practice my mandarin), they generally like it. Makes them feel safe.

The CPC has sold them on a vision of them as members of the state-race "Chinese" (which is not really an ethnicity any more than "American" is) and the surveillance as a thing that keeps them and their "Chinese" lifestyle safe from non-Chinese. Uighurs have to be extra surveilled until they're also Chinese, which, many are now according to the CPC.

So PRC citizens feel safe and cozy among in the country for "their people," not realizing this whole ethnonationalist concept is at best 100 years old, maybe even younger. During the Qing dynasty, there's a whole hell of a lot of people that think of themselves as "Chinese" that definitely weren't by the dynastic government.

I smell similar happening in Russia, the USA, and Israel, with State support. It looks like right wing groups are trying to pull it off in the UK and Germany as well.