jongjong

It all started going downhill once they started calling corporations "legal persons" and gave them all the same right as people, but without all of the constraints associated with having a body.

A corporation has more legal rights than a person if you think about it; no wonder they took over. A corporation can't go to jail, doesn't need a visa to work in a foreign country.

Imagine if the legal system took jail off the table and people could move to any country without a visa; only fines... We'd be overrun by criminals. They'd be able to monetize these two facts easily; they'd make more than enough money from criminal activity to cover all the fines.

What we have is not capitalism. There is way too much gate-keeping regulation for it to be called capitalism.

Just like China does not have communism. They have too much free-market activity and private ownership for it to classify as such.

If you look at China and think "That's not communism", just be aware that they may also look at the west and think "That's not capitalism." Both are right. In fact both systems have converged.

China took the better aspects of communism and capitalism. We took some of the worst aspects of both... Except maybe free speech; but people are working hard on removing that one!

It makes sense; China came out of an era of extreme government control which has been reduced over time (compared to what it was before). In the west, government control has been on the way up and we can't even imagine how bad it can get.

> When you put your name on an article that's written by ChatGPT or Claude, you're basically putting a provenance mark on it saying: I take responsibility for this. I haven't written it, but I'm putting my name on it. That doesn't give you copyright, but it does give you liability for the content

If the AI is treated as an agent during inference, distinct from its user and not merely as a tool, then this should also apply during training too.

Based on this, it seems that AI agents are consuming people's code without permission. The MIT license only gives rights to "any person obtaining a copy of this software".

So the rights are given to a 'person', and the rights pertain specifically to a person who performed the act of 'obtaining a copy of this software'.

MIT license says 'obtaining a copy' and uses the word 'software', not 'code'. 'Software' is to 'code' what 'shop' is to 'building'; if you bought the shop, it doesn't necessarily mean you own the building. These are two different things and require different clauses. MIT explicitly separates the two and emphasizes that the author of the software retains copyrights (presumably over the code as this is the only thing over which they could claim copyright).

The code is different from the software; you can write the exact same software which behaves in the exact same way using completely different code; can be poorly written or well written. The difference is extremely meaningful to the person who invested effort to write the code in a clean way.

If we say that the agent is a separate entity from the person who ran it during inference, surely the same distinction can be made concerning the person who ran the agent during training. So the term 'person' from the MIT clause doesn't seem to apply here since the person running the agent has been factored out (just as they were during inference). Also, the agent is not obtaining a copy of the software; it's obtaining copies of the code which is copyright and independent of the software (as the MIT license clearly asserts).

You understand that everything which you value about yourself is not what other people value you for. Also, what you value in others also probably isn't an accurate reflection of who they are either. Everything feels hollow, performative, lacking in substance.

Worst of all, your understanding of reality works against you because it's demotivating. It's harder to perform your role when you don't carry any illusions about its true purpose.

It's difficult to be happy when you believe that everything good which happens to you is at someone else's expense and can see a lot of the tricks.

They don't produce anything that people like. Everyone using their products is forced to use it.

Their Office suite is foisted on governments. For example, several governments in the EU moved off of it starting a few years back... Many are still on it.

Oracle database is essentially never used outside of government nowadays. This has been the case for decades. Government is just locked into expensive contracts with them and they can charge whatever.

I was a bit hesitant to put SpaceX on the list as it has provided the government better value for money than NASA when launching rockets and the innovation is real. I also believe Elon is one of the better billionaires. His businesses haven't taken away opportunities from any small business. His chosen industries have irreducibly high capital requirements; small businesses cannot compete in those industries anyway and they are too risky.

My concern for SpaceX is more about post-IPO situation. Once new shareholders have their way, I don't see how it can avoid becoming corrupted. It looks like the perfect money pipe to obtain large amounts of money from the government.

The legal system and government these days are mostly tools for oppression of the working class. Most people would be better off without any laws at all and just talking things out with their neighbours, employers and employees; refusing any form of unearned trust or authority.

Democracy doesn't make sense anymore, unfortunately; nobody has any ideas who they're voting for... Photos can be faked, videos can be faked, candidates can be pre-selected during primaries. Worse than that; each individual is shown different facts so nobody even agrees on the basic facts. So literally people are voting for constructed, imaginary individuals who don't exist outside of their heads. Also, there are too many people voting for too few representatives. This guarantees poor representation.

The legal system should be replaced by some kind of vigilante justice hybrid like in the old Roman system were people would be exonerated for taking revenge on just grounds. It's impossible for the legal system to be fair anyway. The closest to justice we can achieve is through decentralization.

IMO, the role of government is to dismantle any large organisations which compete against it for power and control and shield individuals against those organisations.

It's pretty obvious now that governments have been propping up the tech sector in many countries. It's a weird industry. Look at the biggest tech companies. They all start out with apparently legitimate products but slowly lose market share over decades, end up doing nothing and yet more profitable and more valuable than ever. Look at companies like Oracle, Microsoft, Palantir, SpaceX, Salesforce... Their main business is extracting money from the government.

The outcome I'm clinging to is freedom. Until I have that, nothing else matters.

Yes. Before I became a dad, I didn't realize how physically and financially constrained people are with regards to number of kids that they have. For me and my wife, our first child took all our time, resources and energy.

This was a shock to me because my parents had 4 children and never seemed to struggle. Also, they were much better off than me financially and could afford to hire help.

My biggest achievement as a dad is the fact that my son speaks 3 languages. In spite of the fact that I was working full time, he still learned to speak my native language fluently from me. And I still did house chores too. And still worked on side hustles at night and weekends.

But yes, I had no life at all for 5 years straight. No life at all. I'm a shell of my former self. My main concern is that I'm showing my son a bad example of how a dad should be. I literally can't afford to raise him with the same values as I was raised with... Because he would grow up into someone like me and that would be my greatest fear.

Now I understand why some parents spoil their children and focus 100% on boosting the child's confidence. This is what you have to do if you're poor. Because you need your child to believe they are above you. That way when they see how shit your life is, they don't have to believe theirs will also end up like this.

Yes, for coding, they aren't overthinking enough. I want much more thinking and less code in the PR! Even with the best frontier models, I still have to guide them towards the right solution. The more thinking they do, the less code they write.

I have quite a complex codebase where I made a lot of nuanced decisions with regards to keeping the processes embarrassingly parallel, DB indexing, caching, async/await, backpressure monitoring, spam prevention, schema validation, etc... and now the agents are really good at adding features on top and prompting is minimal.

If you have a relatively large codebase and never even once cut a corner, then the AI agents tend to follow through with that style and the ratio of reasoning-to-code increases. Worth it.

For coding, this is very interesting because the same incentives were present for humans before AI. Tech companies which had a culture of rewarding complexity would see huge Pull Requests and a lot of unnecessary complexity. I've worked in companies which would require a thousand lines of code to implement a feature which would require only a hundred or so lines at a different company. The shorter one was more reliable too. Code begets more code. The incentives created by the company culture had a massive impact... And the culture was heavily determined by whether or not the company had a market monopoly. More monopoly power -> more unnecessary complexity (presumably so that employees could achieve better lock-in/job security through the increased need to manage that complexity; in any case, the company could comfortably afford and it did not present an existential risk as it would in a startup environment).

So it's not surprising that the same dynamics are at play with AI. Now, because code is being churned out so rapidly, the effects have become much more obvious (it took me years to figure this out, but now managers can observe this same effect play out in months); many senior engineers and CTOs will echo my point; but I suspect most engineers and crucially, most managers, still don't get it...

Something tells me that the AI companies supplying the models are well aware of the tradeoff. When you can dial up the complexity of the LLM's output by 5% (I.e. 5% more tokens to solve the same problem) and see a 5% immediate increase in your revenue from a large segment of your users, that's a very tempting knob to dial up! Now when you learn that this complexity compounds and next year's revenue will be 10% higher (purely as a result of your users now having to maintain that additional complexity); this is extremely tempting! Especially in the context of users who are largely ignorant as to the true cost of the unnecessary complexity they are adding... The insider's term for this is 'technical debt' for multiple reasons; including the fact that it compounds like normal debt. Now factor in the monopolistic tendencies of those tech markets... It's a real bottomless gold mine.

Revenue from downstream corporate users comes in regardless of compounding code complexity and slower pace of delivery; those companies just keep hiring more people, spending more on tokens. Swallowing up these massive diminishing returns like an appetizer. Enshittification takes place but the downstream end user has nowhere else to go.

It's trivial for a lab to advertise themselves as being token-efficient and almost impossible for its corporate users to actually verify it.

The only real issue with that business model (possibly a fatal issue) are these open weights models which the big tech companies could use to move off the AI service platforms if the problem becomes bad enough.

Although I will admit that formal verification is looking better now than it ever did, and it's probably easier to generate proofs for those few simple safety-critical systems which benefit from it, I'm still bearish about it for the mainstream case.

I agree with "Even if fully automatic verification were within reach, it would be detrimental". Formal verification proofs make the same trade-off as overly fine-grained unit tests; they lock-down the current implementation and thus significantly reduce operational agility; because, if you make a modification to the code, you may have to re-generate the entire proof again. Proofs thus lock down sub-optimal abstractions and implementations. For many kinds of software, requirement changes are a daily occurrence and proofs would get in the way of making the required changes.

Even with complete, zero-cost automation of proof-generation, with no prompting or user-intervention (which would require the AI to have internalized a complete, perfect world-model), there would still be an incentive problem; when engineers see a lot of proofs and/or fine-grained unit tests, they are often reluctant to make the necessary refactoring to meet new requirements. Existing (counter-productive, flawed) abstractions become part of the lingo of the team and it becomes literally impossible to move off of them; yet they create a lasting barrier for new team members and when implementing new features.

The biggest problem though is that many modern software issues are flaws in the requirements (the spec itself), not in the implementation. The requirements are often produced by business people who often have a vague idea about what they want; requirements usually contain subtle contradictions or conflicts which have to be resolved.

Having worked on projects with clean, well architected code, requirements issues are by far the most common issue. On my last project, I kept coming back to my business/product co-founder with questions like: "You said that this checkbox should be on this page; but for a different onboarding flow, you said that it should be impossible for that specific user role to see/select this checkbox and the backend processing relies on this fact for reasons X, Y and Z..." or "You said to apply a filter to the collection and keep narrowing down the set as the user moves through the stages in the flow, but now you want to add a step which expands the set again with data from a different source; so now we can't just update the filter against a single collection; we need to make a separate table to hold the data from different sources; that will require some refactoring and it adds overhead since now we have to keep a lot of data per-user and we need to account for malicious spam-scenarios, etc... We can't just hold all the state in the URL (for bookmark) anymore... Users can't just share filters with each other anymore to restore the same app state across account boundaries."

In my last project, most of the work was trying to figure out what my co-founder wanted and it turns out that the idea he had in his head about the system was not logically consistent across all of its parts; a fact we only discovered after months of implementation. Also, he did not understand some of the technical limitations in terms of what kinds of data a free public API would give us access to; and that turned out to have been fundamentally incompatible with business objectives and the target market. His refusal to pivot to a premium market (where the user may have been able and willing to pay to cover the additional downstream API costs) marked the end of the project. The project was logically impossible from the start given the hard constraints of what platform to rely on, what our costs would be and what the target audience was. If we need formal verification, it would have to be for the requirements themselves, evaluated against the technical constraints. We don't need formal verification of the code.

Correct code is a mostly solved problem if you break down the typical software system into its sub-parts and identify the right platforms (e.g. CRUD, edge functions, data ingestion, data processing...) Correct requirements are a far bigger problem.

People still need a platform to host their app. AWS is far too complicated for the average person. Even with AI to help. The user still needs to get past the registration screen and put their credit card. Most will panic and run as soon as they see the letters "AWS", "IAM" and "EC2"

Then as soon as their agent hits the first bug "port already in use" they will run... Any brave soul who remains, after spending a whole day solving similar issues back to back, will be like "The port has been freed up, my nginx proxy has been configured, my nginx has been launched with systemd and my app server too! The socket ulimit has been increased, I still can't see my app!!!" Then Claude be like "Well of course, silly human! To achieve this simple task, you would have to update your security group inbound firewall rules, duh!"

Then the user be like "Claude, we fixed everything! I cannot see my app still!" Then Claude be like "You goofball, we didn't even get round to purchasing your domain name, moving it to CloudFlare DNS and configuring the A record, CNAME and AAAA record for IPv6 in your Zone file... Not to mention the MX record. You also need to obtain a signed SSL/TLS certificate from Let's Encrypt, then modify your nginx config to install the SSL certificate... Oh then we'll need to configure your nginx to act as a terminating reverse proxy for your WebSocket connection... Oh actually before we do any of that we should set up your schema validation logic and access control logic on your app server with rate limiting and backpressure monitoring... Then we'll set you up on Git version control with a CI/CD pipeline with precommit and PR hooks to run your test suite and code linting... Then we want to keep the registration and login flow of your app real simple, like stupid simple, so we'll get you a Keycloak instance acting as an OAuth2/OIDC identity provider for authentication with an email verification step which will involve us configuring the realm and creating an OAuth client and setting up the SMTP email server... Too easy!

I was open minded about it until I did a job search for "Lovable" on the biggest career website in my country and couldn't find a single job related to it in the entire country. For any other tech stack, any major town or city would give you many pages of results. Surprising.

Yes it is, how else do you propose those in power should get rid of dissidents without alerting people?

Wow. How much does this job pay? Google says $550k to $1.3 million.

For that kind of money, my AI paymaster could do no wrong.

Imagine going to work and getting paid $2300+ USD per day to sit on your ass all day thinking about ethics without having to produce any deliverables. This would be a dream job for me.

If anyone questioned my morals, I'd be like; "They're paying me $550k per year" and the other person would be like "Ahhh ok, my bad! Your logic is starting to make sense now!"

This is not the reason though.

The reason is that now there are fewer different kinds of jobs (many roles like manufacturing have moved offshore) and also fewer jobs in terms of numbers and more regulations which prevent competition and shrink the number of competing company (again leading to fewer roles).

It wasn't nepotism before, it was efficiency. I didn't have any connections and I got better opportunities back then than I do now and the people were overall better at their jobs. Nepotism is much stronger now and jobs are increasingly theatrical/bureaucratic and have little to do value creation.

It's possible that people's lifestyle in the west was propped up by the exploitation of foreign labour, that much is possible, but note that the people who did the exploitation back then are the same doing it now (or their heirs). The difference is that now the money is increasingly circulating just between each other, bypassing normal people. Soon it will bypass all humans; the entire economy will be between machine owners. Social mobility is mostly about social connections and politics nowadays. In the west, it has never been less meritocratic.

I don't trust the legal system. They could cover up the evidence, get me blocked on HackerOne, claim that my screenshots are AI-generated, hire top lawyers then make the judge to charge me for the lawyers' bill.

The big company always wins. The legal system is pure fiction at this point. What lawyer would stand against the big companies? Permanently destroying all their future career prospects.

Erin Brockovich? That's a corporate propaganda movie.

Reality is more like what happened to Julian Assange or Steven Donziger. And they had support from some powerful groups. If they didn't, we wouldn't even have heard of them. That would have been my situation. Not worth the $200 bounty.

Last time I reported a DoS bug to HackerOne, the company behind the bounty tried incite me to commit a crime against them by DoS'ing their servers using the hack I had reported in detail!

I literally showed them their server taking over a minute to respond to my request. I even showed how the delay increased proportionally to the message size... Clearly doing more processing; classic DoS vulnerability... Doesn't leave much to the imagination! But they said they would not pay me anything unless I actually proved that it scaled and caused disruption of their service!

It seemed like they were baiting me into incriminating myself for a crime that they wanted me to commit against them. It's not even the first time that I've been baited by a software company into committing a crime against themselves. I never took the bait though.

Software industry is the same. Main difference is that it is not new in that case; it has always been that way since the dawn of that industry. The software industry is worth trillions but nobody knows anything about what's going on except for hacks. Literally, getting hacked is the only way your software project can get any attention, and it should already be a popular project or else nobody will care. Attention is more scarce than it was before.

Yes, it's a spectrum. The best test to figure out the extent of your own bias is to make a list of ways which in which a particular ideology benefits you.

For most of my contrarian ideologies, they are harmful to me in most ways except "Protects my ego by redefining personal failings as system failings." Most of my contrarian ideologies have this characteristic but this point has been getting weaker over time; I don't have much pride left. I'm literally working for a bank now.

Or maybe I work for a bank just to prove to myself that I don't care about pride? Getting kind of elaborate.

I do this already with all of my beliefs.

That's not to say my beliefs are exhaustively battle-tested and correct; maybe other people have information that I'm missing but in my role as a software engineer working across multiple industries and some politically sensitive projects, I've had access to information and first-hand experiences that most other people haven't.

The thing is; if I told the average person what I witnessed without providing any interpretation, they would doubt the basic facts because some of what I experienced doesn't make sense when measured against mainstream narratives and most people's day-to-day experiences.

For example, about the crypto/blockchain sector; when I tell people about my experiences, outsiders often say stuff like "The founders behaved this way because of hype and irrationality of the markets" but when you've been through it every day for 2 years and experienced the taboos inside the company, seen the suppression of projects and community members, following a clear, predictable, repeated pattern, it's impossible to believe that it is random or irrational.

Rational people will pick the ideology which has the fewest contradictions. Irrational people will choose that which serves them best.

People who benefit from a system are unlikely to look for problems with it and unlikely to notice any contradictions. Anything which appears as a contradiction gets shrugged off as "Well, there must be a good reason, I'm not an expert. I will trust the experts." These people corrupt Occam's Razor that the simplest explanation is most likely the correct one. "The experts say ..." becomes the ultimate simplest explanation to everything... And the experts that they choose to listen to are those who are aligned with their existing beliefs.

Then you have people who notice the contradictions but they see so many from first-hand experience that they don't trust the experts and they don't have access to enough different information to be able to choose a good one; so they just adopt the least-bad one as their own... The one which best protects their self-image and whose (comparatively) few contradictions can be plausibly explained away by corruption, conspiracy or dishonesty.

But I think the latter group is in fact one step closer to the truth. The truth is there is no conspiracy (at least no grand, all-encompassing one), it's worse than that; systemic incentives create alignment and remove the need for intentional coordination. The coordination, the illusion of conspiracy is just a byproduct of incentives. The second group of people are wrong about the mechanism but their view of reality leads to the same outcome... So with each passing year, their views are being reinforced by more "proof."

The system which manufactures the incentives which create alignment is itself probably not the product of conspiracy; it's just a system which evolved slowly over time to subtly serve its operators and their supporters. They don't want to change it, not out of nefariousness, but out of justified fear of what havoc any change might wreck... Bureaucracy also plays a part.

It's very difficult to co-ordinate a lot of people to agree on anything, let alone act on it. This is one thing that the mainstream anti-conspiracy folks are completely right about.

I'm a global top-50-university-credentialed expert in Information Technology who majored in System Design and has worked for a wide range of companies of different sizes, in multiple industries (education, remote work platforms, social media, banking/finance), in multiple countries including in highly political areas like crypto... So the "trust the experts" folk can trust me when it comes to information flows and system design...

This article reflects my values and philosophy exactly... However I think the author is underestimating the cost of not following incentives. It's torture.

In this society, I'm a loser precisely because I ignored incentives. I think we live in a failed society so I don't mind that much being a loser in such society. I don't think I would feel any pride being a winner... However, I would very much appreciate the financial comfort of not having to participate in it! So that would be the advantage of being a winner.

It's the great irony; the winners who have the option to not participate, want to participate... And the losers who ignore incentives find that they are forced to participate sooner or later!

Right now I think my current job is probably unethical. Though most people would probably not think so. They just didn't look for the devil in the details.

Most people are lucky to have been following incentives subconsciously, believing that incentives align with economic efficiency and, therefore, moral virtue. I was not so lucky to believe that. Now I have to do stuff that I know to be unethical to free myself. Is this the definition of corruption? Isn't it a worse form of corruption if you don't even bother trying to understand the impact of what you're doing? Or worse, you deceive yourself?

Yes, the full extractive structure is much more visible now. Used to be discussed as "conspiracy theory" but now it's obvious that such structure exists and it doesn't require any conspiracy; shared incentives suffice. So yes, I would say that this structure is emergent.

I think the statement "code was never the hard part" is true but it's relative.

Programming is not easy and it takes years to master. Some people became really good at solving complex programming puzzles and 'Code Jams' and focused on it. Unfortunately, the same aspect which made this skill highly visible and highly praised, is what made it easiest to automate.

All those medals, trophies and certificates... not the mention advantages at big tech software job interviews... Came at a cost.

Meanwhile there is a whole group of people who have been honing their skills in software design, architecture, distributed systems, security and other less visible, less rewarded skills who have been ignored by the markets. These people still can't be automated.

It's a large problem space so after a decade or two, the coding aspect feels small relative to all the theory and experience surrounding it.

I met many senior people who didn't take programming seriously as a skill, long before LLMs.

One time, when I was at university, one of my math lecturers was boasting about the superiority of math as a discipline and said to the class "Software engineers... There are no software engineers; they're programmers."

That statement was never true but it's much more obvious now. The fact that a lot of people shared this belief highlights the fact that these other skills were invisible.

There is probably as much engineering (if not more engineering) involved in delivering a complex, reliable software project as there is delivering a complex skyscraper project in civil engineering... It's the same kind of activity; lots of interdependent parts, each with their own constraints and requiring many decisions to be made with lots of tradeoffs. At least with a skyscraper, the customer requirements are relatively very stable.

Whenever I hear some big AI CEO say that AGI/ASI might end humanity, I imagine it doing so by conditioning people to commit suicide or some kind of massive-scale manipulation. Probably setting people up into complicated bad situations and then blackmailing them.

I don't know if this is it here. These sorts of jobs can have tough moral implications. But it seems like a hierarchy of blackmail has been set up in our society and it seems like a dangerous vulnerability.

I think the part that's hard to absorb is that software engineering was a brutal path. You had to restructure your entire identity, ego, personality, social relationships around that career choice. People didn't expect their career to last forever, but they thought it was a ticket out to gain autonomy.

Back in 2010-ish, many developers thought that, in the future, they'd all be running their own systems and servicing their own customers and making a good living like that (there would be enough customers for every dev). They didn't expect that the entire industry would be monopolized by a tiny number of companies, everyone using the exact same product, no alternatives or competition, and that the value of our labor would be artificially suppressed down to $0... This happened long before AI. AI is just the latest assault.

It's impossible for a solo developer to find their own customers nowadays. You could literally offer the same product or a better product for cheaper, but you won't be able to find a single customer because it's too centralized. The trust barriers are massive. Nobody has money to try anything except large corporations and they never try any product based on what it offers; it's all business networking.

It's like; the same conditioning which caused business people to pretend that Linux didn't exist for the first few decades has now spread to all non-VC-funded software.

AI can uncover certain issues and resolve certain problems but it also has its blind spots.

I've been using Claude to create a software architecture diagram. It came up with a lot of useful functionality that we had neglected to show but upon further examination a lot of steps didn't make any sense. It added a box to do validation on some data then it put another box downstream to do validation again on the exact same piece of data (the second box had a different name but upon questioning, Claude conceded that it was the exact same validation step). When doing inference on user input, it put a box to extract specific insights from the prompt and then it put another box to do inference again on the inferred data to categorize it... And this would add latency and unnecessarily strip out useful context when doing the second inference... It really only needed a single inference step...

I re-uploaded the diagram to Claude and confronted it and it conceded to all of my points and it even noticed a theme between them and suggested that the diagram exhibited "a pattern of redundant steps." But it was incapable of suggesting viable solutions besides name changes. All of its solutions made the design more complicated. It was incapable of simplifying the design. I've noticed this with AI code; it can make things more complicated, add more features but struggles to simplify things.

Also, AI is horrible at rating things... It's just too superficial. For example, if you write a huge amount of intentionally over-engineered, tightly coupled code with poor separation of concerns but you do proper linting, define all the types, add a lot of comments, it will give it a higher score than a shorter, more readable and maintainable snippet which exhibits loose coupling and high cohesion because it lacks the superficial aspects.

It's horrible because now business people who understand nothing about coding may ask AI about code quality and it will consistently rank low quality over-engineered projects that are full of bugs, unmaintable and less secure, with a higher score.

Frustratingly, if you point out problems in its judgement, it will concede to your points without reservation, even building on top of your argument... but it will never actually tell you this stuff up-front, unprompted if you don't already know it! It never seems to reveal new insights, at best it can only expand on existing insight which you've already had.

It tells you what you want to know but not what you need to know.