ineptech

Developing this a bit; I get that founders might object to being lumped in with the asshole-rich-guy demographic associated with the term, but "capital" doesn't mean people who never had to work hard, it means capital. When we talk about laws and regulations that advantage capital over labor or v.v., we're talking about money rather than people. The fact that founders get rich via stock options rather than by paying themselves huge salaries is an example! If labor had won, wouldn't pg be trying to pick a startup to work for rather than one to invest in?

> You would think, after having been on the side of labor in its fight with capital for almost two centuries, that the far left would be happy that labor has finally prevailed. But none of them seem to be. You can almost hear them saying "No, no, not that way."

This one really threw me for a loop. Do startup founders making or seeking a big exit really think of themselves as "labor"?

It's still pretty deranged. "When Mark shares on Facebook, a part of the narrative is a generational narrative..." Like millennials are sitting around reading whatever Mark Zuckerberg posts on facebook and discussing it with each other.

Agreed, we need a postscript from the author explaining that he set out to avoid drawing a comparison with chaotic systems, but after months of editing and tearing up handwritten drafts found himself arriving at one anyway.

Well I hope so, but it would take something more exotic than the directory service described in this article, right?

... thinking it through, maybe not. I guess if a client initiates AddVerification for some twitter handle, and the directory service sends that handle a private twitter message with some guid, and then the client sends AddVerificationStepTwo with that guid, the directory service could append both of those messages to attest that that user is associated with that handle. A malicious directory service (or a real directory service that's calling a malicious twitter clone) could fail to correctly verify someone, but it wouldn't be able to add a verification for the wrong client or for a client that didn't initiate it, which is probably good enough?

edit: just realized you're the author, thanks for working on this stuff!

This is a directory of public keys, which every secure messaging app including Keybase must also have, but federated rather than centralized. That's a necessity for e2ee and it's very cool to implement it as an append-only version of the pub-sub messaging the fediverse already has.

It has nothing to do with the identity verification a la Keybase, unfortunately, as that would be a really good and useful feature for the fediverse, but I think it would be a lot harder to do.

In the early 80s, the local pirate scene in my hometown revolved around a guy who I'll call George. George's entire basement was devoted to boxes full of diskettes (5 1/4" in those days) of pirated software and photocopied documentation. All kinds of software - games, office apps, scientific stuff, you name it. Some was downloaded from BBS's but the majority of it was shipped USPS from god knows who.

Thing is, he used virtually none of it. He collected software for the sake of collecting it. He didn't even play video games, just loaded them up once to make sure they ran. He was a hoarder basically, who had stumbled into a niche hobby, and like most hobbyists he would happily share it with anyone who asked.

I'm not saying every pirate is like that, or even most, but I am saying, I don't think the pirate scene works without people like that. The music fans are spokes, but people who do it for the sake of doing it are the hubs.

For any allegedly-random distribution, it's possible to statistically determine an upper-limit on the size of non-random-appearing structures. The upper limit for such structures in our universe is thought to be about 370 MPc, about 1/3rd of the size of this ring.

A lot of these questions are much more clearly addressed in the previous paper by the same authors, which is much more layperson-friendly: https://academic.oup.com/mnras/article/516/2/1557/6657809?lo...

> Why is this thread acting as though teaching our children self-control is somehow a different lesson?

Because some things require more self-control than others.

Imagine if the drinking fountains in public parks had non-alcoholic beer instead of water. Installed and maintained by Budweiser, totally free of charge, guaranteed ice-cold. Responsible parents would avoid the beer fountains and bring a thermos of water, but a lot of parents would just let their kids drink from them, right? And even very responsible parents would probably let their kids drink from them occasionally, because we forgot the thermos, and it's right there and it's free.

Raising non-drinking kids in that world would be harder than in this one. Everything you said about "teaching self-control" and the importance of parental effort would still be true, but it would also be true that when kids in that world start drinking earlier than in ours, it's not solely the parents' fault.

That probably sounds like a ridiculous example, but I think it's pretty analogous to schools showing Youtube videos. My kids had never heard of youtube before they saw it in their second grade classroom, and I'm not thrilled about it, but I get it, because it's right there and it's free. Raising kids is hard no matter what, but the fact that the world's biggest and best source of good educational content also happens to be riddled with crap makes it harder.

Not everything is like that. The library doesn't have Stephen King mixed in with the picture books. The Disney channel isn't also Pornhub. The parks don't have beer fountains. That's the difference.

> "set reasonable restrictions on phone usage"?

It's not so much that people are too dumb to do this or don't understand the danger, it's that doing this well is tedious and time-consuming and expensive. Setting up parental controls and monitoring them is a chore, you have to sign up for a bunch of apps you don't want and read outdated docs and troubleshoot random error messages and crap. Most of the good stuff costs money, all of the shitty stuff is free. Finding educational apps that aren't terrible requires research, whereas the worst garbage imaginable is heavily advertised.

So, I'm not saying you're wrong, but that this is a problem with incentives as opposed to a moral failing. I think there's a direct parallel to the rise of highly processed foods. The people 2-3 generations ago understood that junk food was unhealthy, and very few people fed their kids 100% soda and hot pockets, but junk food was so much cheaper and easier that people gradually allowed more and more into their diets, and now here we are with an adult obesity rate 3x that of 1960.

> "teach your child about the issues with being on a phone all day"

Edit to add, it seems most of your other responders are responding to this bit, which I agree is not realistic (children aren't rational to reliably refrain from something just because you can explain why it's bad) but I can see why you're frustrated that they're zeroing in on that part and not the part I responded to.

I think you're reading that comment rather uncharitably, I interpreted "description of reality" to be a synonym of "model".

Comparing someone's wage to the value they produce is a fine way for a company to decide whether to hire someone, but I didn't think that was the question you were trying to answer, was it? Perhaps I misunderstood, but I thought you were asking something like, what policy would help people who are at the margins, which is an economy-wide question that can't be answered from one employer's perspective. Workers may only be "worth it" at some wage, at some point in time, but that wage is subject to supply and demand just like everything else. A policy intervention like raising the minimum wage will alter that supply/demand curve.

For example, suppose janitors all make the minimum wage. If we increase it, there might be some company at the margin that will go without janitorial services, but most companies will pay their janitors the new wage, which (from the "a worker costs $X and produces $Y" model's perspective) will look a lot like the nation's janitors suddenly started producing more value. Ergo, it's not to say that that model is wrong, just that it's not useful in answering a question like should we increase the minimum wage.

> someone who only produces $7 of value is no longer employable

This is the wrong model. You're using a worker's wage to describe their productivity, and a big reason for the mess we're in is that wages stopped increasing with productivity fifty years ago. (search "wages productivity graph")

Any source for this at all? I've seen it before and it's a top-rated comment here, but it doesn't sound plausible and it's exactly the opposite of what game theory would predict.

No, I'm 90% certain both Dijkstra and OP meant that what's critical is mastery of any tongue. In context, he was talking about what makes someone good at programming, not about the need to contribute to a team or communicate ideas.

https://www.cs.virginia.edu/~evans/cs655/readings/ewd498.htm...

It's a complex topic but a popular answer is requiring the owner of the last-mile infrastructure to make it available to competitors at cost. I think that is (roughly) what places with good/cheap broadband do, e.g. South Korea.

This is a kludge because we don't have the actual solution to keeping prices low, which is market competition. No one wants the government to tell big companies what to do, but when you let a monopolist take over you're stuck with either consumers getting screwed or burdensome regulation, which are both bad.

Same deal with net neutrality. In a healthy market we'd have it naturally, without laws prescribing it. We don't have a healthy market, so here we are. And there's no greater gift to the monopolists than the position you're taking.

I'm not sure? Mine still works but I've had to manually upgrade it a few times. For a scheme like this we'd probably need to reimplement it (just the public keyring and challenge proofs on social media platforms, not the crypto cruft). Helpfully I think the client is FOSS.

I don't have enough time left on this Earth to explain the concept in a way that politicians could implement, I'm in my 40s. In my preferred alternate universe, Keybase was sold to a benevolent billionaire. Or more realistically, a normal billionaire who intended to run it at a loss until he could leverage it to effect world domination, but managed to mess it up somehow and get it nationalized. Or something. I can dream...

Question that sounds idiotic but is quite serious: how do I make it illegal to lend money to me without confirmation via Keybase? (edit: or some similar cryptographic identity proof)

The only reason to keep my name/address/SSN secret is that companies will lend money to a person who has that info, and then try to make me liable for it regardless of whether that person was me. That's a problem, but the solution isn't for me to keep my identity secret, it's for companies to stop doing that.

I should be able to march into some government office, prove my identity to their satisfaction, and give them a private key. Then, if Wells Fargo lends money to someone who can't prove ownership of that key, that's Wells Fargo's problem. Keybase does this fairly well, and is essentially abandonware since the founders were (if I remember right) acquihired by Signal. So, can we just nationalize it or build something similar, declare it to be SSNv2, and move on with our lives?

I don't think choice is involved here, public and capital-financed companies that aren't maximally greedy get eaten by those that are. When your competitors raise enough capital to undercut you in the short term, you don't survive to the long term unless you raise more capital as well, and you can't both offer the highest returns, one of you is going to end up buying the other, and only after all the corners have been cut and every customer that can be screwed has been screwed. It's a race to the bottom that doesn't necessarily have a solution. Isn't this essentially how Maytag went tits up? To compete with the cheap foreign appliances they took on a bunch of debt, and then cut costs to service the debt, resulting in the "amanatags" that destroyed the reputation for quality they'd built over 3-4 generations?

Eh, when something goes extinct I'd say that's evidence that it isn't viable. That's the paradox of capitalism: anything that can be ruined eventually will be. I'd pay 2x or 3x for 30yr appliances for environmental reasons, but I don't think the company that makes such a thing could raise the capital to build them in the first place.

Is it even realistic for a company to sell $1000 appliances that last 20 years? I know we can build such an appliance, because they did in the 80s, but can a public company survive selling them?

> a server farm aboard a ship that only sails in international waters.

How would this help? If the server software is actually trustworthy and the connection to it is e2ee then it really doesn't matter where it runs. And if it isn't, this kind of setup protects whoever owns it more than whoever rents space on it.

I agree that copying music is not all that similar to an AI model using training data, but they're both examples of a superficially similar activity being transformed by automation and scale. The difference between manually copying a song on cassette for a friend and uploading that song to napster resembles the difference between a human learning to draw in an established artist's style to an AI model learning to draw in an established artist's style. In both cases, the offender's defense is a variation on, "It was okay when humans did this slowly..."

This is a little oblique, but isn't there a parallel to the music piracy debates of the 90s here? The defense of those who steal training data, "It's not stealing because I'm only taking a little tiny piece from a whole bunch of people", seems like a neat inversion of the defense offered by people torrenting mp3s in their dorm rooms.

It's worth noting that, back when it suddenly became technically feasible to copy all songs for free, the response from the rightsholders of music was, essentially, to demand that we all collectively pretend it isn't. It might make sense for the rightsholders of training data to try to do the same, if they could speak with one voice and if they had lobbyists, but I guess the fact that they can't and don't settles that.

And from the following day, their paraphrasing of https://www.theverge.com/2023/11/4/23946353/generative-ai-co... is succinct:

> Meta: We stole so much stuff that, actually, we didn't steal any stuff

> Google: If things were different, things would be totally different

> Andreeson Horowitz: But we already spent so much money

> Microsoft: Think about how this would hurt the little guys, like us

> Anthropic: Fucking shut up about it

> Hugging Face: It's super legal but honestly it might not be idk

> StabilityAI: It's legal in places that aren't here

> I'm always reticent to fully engage in "The Dialogue," regardless of its momentary configuration. It's a smoothie made from shibboleths; you have to be able to detect them at only a few parts per million because once these things metastasize, they stop being about whatever they were about and instead become increasingly loud recitations of various catechisms and loyalty oaths.

Boy does that ring true.

Sprints make/made sense in the world of slow deploy cycles, especially for lengthy manual regression testing. If it takes you three weeks to get from a story being done to a story being in production, deciding which sprint a thing goes in is the difference between customers getting it on day n or day n+21. Once that's not a problem you can kind of forget sprints exist.

Two thoughts:

> "Way of work should be defined at a team level by its people, not by the company."

OP delivers that like a surprise, but isn't that the single core defining tenet of Agile? I thought it was.

> Sprint Planning — as simple as it seems, a long session (up to 4 hours)..

My big insight is, painful sprint planning is a result of painful deploys. Once you accomplish real CICD, you can just dispense with sprints and sprint planning entirely.

I'm not sure if that's something most people would agree with or not so I won't belabor it unless someone argues, but I can go on at length about that...