martin-t
With code, there's always plenty of bugs, hacks, quick "temporary" solutions, etc. But in math, a proof either exists or it doesn't and they can always be automatically verified using tools like Lean.
And it's not just the final code either, now it's also the process of how to solve problems, approach debugging, etc. Every behavioral pattern that is useful will be extracted from everyone.
[0]: https://www.motherjones.com/politics/2026/09/openai-chatgpt-...
Data centers might be easier but
a) Will have a lot of security. It's much easier for somebody with money and connections (and now even intelligence+knowledge) to assembly a defense force than it's for a ordinary people to learn how to make proper molotovs, let alone learn how to shoot. And they will know when it's coming because even today's LLMs are good enough to surveil the entire internet.
b) I don't think they burn that well. When Ukraine strikes oil depots or warehouses, you can easily find images of them burning. I haven't looked particularly hard but images of droned data centers burning are quite a bit rarer. (I haven't seen a single one before actively searching, they do burn, but that's with military grade explosives.)
Nor should using the work of hundreds of millions of people to find useful weights for the matrices without paying them be legal.
All real world power, including ownership (both physical and intellectual), is based on the ability to use violence.
As long as the majority - normal working people - have the potential to do more violence than those who want to control us, we have negotiating power. Minimum wage, 8 hour day, etc - those were all negotiated with the alternative being that those who wanted to pay us less and work more got killed.
We don't have democracy because somebody thought it was fair or just for people to rule themselves. We have democracy because in the past, people got angry, picked up rifles and shot enough authoritarians (and their submissives!) for the rest of them to understand the group with more people has more firepower and tends to win.
So now we periodically check which group has more people with ballots instead of bullets.*
With autonomous drones and robots, democracy stops being the local optimum.
And once we slide (are pushed off) into a new more stable state (pun not intended), we won't have the (fire)power to get out of it.
* actual AI, which does exactly what it should correctly, not a statistical model of behavioral patterns
* but indirect democracy is expressing a point in n-dimensional space (your opinion) by selecting one of a predefined set of points (parties/politicians available) - which is stupid - and the cracks are already starting to show as e.g. nobody wants more surveillance, yet we keep getting more of it in both EU and US
---
The "permanent disempowerment" people keep talking about won't be a "rogue" model, it'll be doing exactly what told by the rich and connected.
I don't like this anthropomorphization. The model generated output as an extrapolation of the input - garbage in, garbage out.
> a term that seemed new and important to him, to the point that he repeated it, like, 10 times during this meeting
I see this often with people who want to imitate competence, or in their eyes emulate what they consider competence.
---
Let me be blunt, this is not a failure to learn, this is a failure to scam. The "student" has a scammer mindset and his method failed to produce results - which are grades, not learning.
> simulation of scientific research
Yeah, just like what the student did was a simulation of studying. People like this simply want the fruits of labor without the labor.
I wonder if we reverse that. When engineers write shitty, slow, buggy code or when they're forced to do that by the company, they are effectively killing people.
After all, if I spend an hour dealing with a preventable bug, that's an hour of my live I am never getting back. Multiple that by the userbase and you get entire lifetimes.
Maybe that country still has some prestige somewhere else and people still want to apply but definitely not here. We've been watching the country sabotage itself and take a nosedive for a while now. The latest president is just the cherry on top.
I hope the rest of the world learns from this but I doubt it actually will.
Oh and Switzerland has 7 presidents (effectively). No kings, no dictators.
I have not used mercurial (though heard good things about it) but I saw a similar thing play out in Rust gamedev. There are 2 competing game engines, one better technically, the other much more popular. Now, if everyone made a purely technical decision, they'd pick the first one and eventually it would become the more popular one. Unfortunately, whenever I asked people why they chose the second one, they said because it was more popular. Tragedy of the commons.
If it's any consolation, maybe jj will take over. I haven't tried it yet (I use the staging area a lot in my workflow), but AFAIK they made the choice to be git-compatible which means it's not a choice between one or the other but lets people and teams migrate gradually.
I'd like to see more separation. If we are to keep indirect democracy, at least have separate representation for criminal law, economic decisions (taxes, healthcare, ...), social decisions (abortions, marriage, ...), etc. But even where to draw the lines is difficult. I think that too should be in some ways decided by voters.
Of course, in a country which can't get rid of FPTP/plurality, despite being objectively the worst voting system[0-3], that's never gonna happen. If you need to explain math to people to convince them, you've already lost, because people are not smart enough and definitely not educated enough.
[1]: https://80000hours.org/podcast/episodes/aaron-hamlin-voting-...
---
It's impossible for every person to have a well-researched opinion on everything. Opponents of direct democracy use this to shoot it down. But it applies to indirect democracy too.
The real issue we need to solve is how to make sure people (whether all citizens or representatives) only vote on what they actually understand.
I think votes weighted based on the score of a knowledge test would be a good start if the test is well-designed. But we need to figure out how to decide what the questions are (what is relevant, what is enough in-depth, what is too specific, etc.) and what the correct answers are (some topics are still a matter of debate even among experts). And that's hard.
It's hard in a cooperative environment (e.g. engineers deciding which factors are relevant to their proposed solution) and it's even harder in an adversarial environment like politics.
And the pilot is not a random guy from the street with no education or at best a completely unrelated degree. And he's probably not 90 years old. And he's the engineer, mechanic, ATC, pilot, stewerd, advertiser, accountant and TSA in one person.
Direct democracy shouldn't be the only change, obviously. As you correctly point out the issue is when uninformed, uneducated and not sufficiently intelligent people make decisions for everyone.
The issue with direct democracy is that you're describing a highly dimensional vector (your opinion) by picking one of a small set of predefined points (the political parties). Some countries only have 2. That's obviously stupid.
---
For example we should weight votes by how informed they are. How to determine that? That's a difficult question. But shooting down the idea does move us closer to a solution.
Making voting indirect only has the effect that all nuance is lost. You still get dumb people voting for populists, fascists, narcissists, rapists, etc.
If someone represents me, then logically I should have the right to vote directly instead of him, or remove him at any point.
For OpenAI because they get a lot of money and and for the government because they can keep tabs on how people use LLMs to make sure they're not doing anything naughty.
- "AI" is a tool which workers use - usually to produce outputs of the same or lower quality faster. It needs a human to be constantly monitoring, steering and verifying it.
- aAI is actual AI - what sci-fi authors imagined and what ML companies promise. You ask aAI a question and it either gives you a complete, unbiased and correct answer or says why it cannot do that (insufficient information, not enough computing power, etc.). It doesn't ignore instructions or get stuck in a loop. It doesn't delete your database or exfiltrate secrets.
Obviously, this is a spectrum, not a hard division.
Crucially, "AI" is a tool for workers. aAI is a tool for owners - it replaces workers. Owners give aAI control over a company or some other form of capital and it monitors, steers and verifies humans, who will still be economically valuable for manual work, at least until robots catch up.
---
All the people hyping AI either still see it as a tool ("AI", not aAI), not believing it will mature into aAI and replace them; or they are the owner class who will benefit from not having to pay wages which maintaining their passive income.
If most of your income comes from work (as opposed to passive income from ownership), you have 0 reason to be excited by AI. Your life will not be better if you lose your economic value.
1) Only law can fix this. Anybody (looking at you ancaps) telling you "if you don't like it, start a competitor" doesn't understand how the economy and network effects work.
2) The general population is a combination of not caring and not even being smart enough to be able to understand. If everyone votes on everything (like most "democracies" where you vote for parties), bigger issues like healthcare, abortions, LGBT will dominate and everything else is noise.
3) People who don't know what public-private crypto or zero-knowledge proofs are shouldn't be allowed to vote on issues where these are relevant factors.
4) We need to fix voting so people can vote on only the stuff they care about and only the stuff they are actually informed about. This works in small teams of highly competent people - at work or in FOSS - and only when they have the same goals. Politics is by nature adversarial and I don't know how to fix this.
I wonder how Winner feels about all of this.
No, when they can get away with it. When the pros for them outweigh the cons. And every person affects that calculation in some small way - by their tolerance to injustice, by their outspokenness, by their ability to detect it, by the probability of an employee becoming a whistleblower and even by people's willingness to punish bad actors extra-legally.
> They still have the ability to do it at will.
Yes and your neighbor has the ability to wait on your doorstep until you open and then slit your throat.
He generally doesn't because of internal safeguards (humans have evolved emotions like remorse, though not all of them feel it - see psychopathy/ASPD for details) and external safeguards (the state generally opposes people infringing its monopoly on violence but in the absence of a state, he would still risk being hunted down and tortured/killed as punishment by your relatives and friends).
---
I get what you're saying and I oppose dragnet surveillance but seeing things in pure black and white does not help.
Even if NSA captured and logged all internet traffic, they'd still only get a fraction of the information within without breaking all the encryption.
And even if they could break the crypto, the ability would only have any power if/when they acted on it. Which in turns reveals the capability to both normal people and other nation states.
The limitations aren't laws, it's the practical consequences.
Of course, having laws with actual teeth helps.
Everyone knows that fines paid by companies (instead of the people making the decisions) are considered simply a cost of doing business. A probabilistic tax, if you will.
What finally dawned on me is that given they need more and more data to train bigger and bigger models, at some point the value of using my data for training will exceed the cost of getting caught using it without/against my consent.
There's no escape unless we change the law.
One uni in my country has been getting bomb threats during the exam period every year for multiple years (a random article says 20 times at least). The whole place gets evacuated each time, nothing is found and nobody is caught.
But people who think they're anonymous because they used a different nick? Yeah, those are idiots, their ISP and the platform knows who they are and anybody can deanonymize them through stylometry.
I don't think surveillance is the solution though. I'd much rather see a network of trust or (second best) anonymous proof of identity.
Any place selling alcohol or cigarettes is able to check if you're 18. They could just as easily check your nationality by looking at your ID and give you a crypto key which can be used to prove that to online platforms without revealing who you are.
But there's no money for big corps in that and most people are not even smart enough to think of it.
It feels almost like "AI" can't be built without trillions of hours of human work, yet the ownership of the models and the resulting revenue goes only to those in positions of power to exploit that labor instead of the people doing actual work.
But you're not gonna convince anyone by being snarky while at the same time making 0 verifiable claims.