alphazard

Product management might be the worst meme in the industry. Hire people who have never used the product and don't think like or accurately represent our users, then let them allocate engineering resources and gate what ships. What could go wrong?

It should be a fad gone by at this point, but people never learn. Here's what to do instead: Find your most socially competent engineer, and have them talk to users a couple times a month. Just saved you thousands or millions in salaries, and you have a better chance of making things that your users actually want.

The ability to reverse is very dependent on the transformation being well known, in this case it is deterministic and known with certainty. Any algorithm to reverse motion blur will depend on the translation and rotation of the camera in physical space, and the best the algorithm could do will be limited by the uncertainty in estimating those values.

If you apply a fake motion blur like in photoshop or after effects then that could probably be reversed pretty well.

Naive self-interest is a characteristic of children. You didn't stumble upon some ancient wisdom of another culture, you just matured and established a family, which your instincts tell you to put before yourself.

The West is individualistic, but all that means is that individuals choose their circle of concern. Without any further guidance, that means that innate instincts about who is important (family, close friends) dominate, rather than a forced narrative about the collective.

What they call "fluid intelligence" is just intelligence and the rest are skills/aptitudes. "Crystallized intelligence" is more plainly: efficacy/productivity and it's common knowledge that people are most productive during the middle of their lives. When they have the best balance of knowledge accumulated and raw intelligence.

In humans, intelligence manifests as memory, spatial and verbal reasoning, pattern recognition, etc. What is so interesting about IQ and g (the general factor) is that all of these abilities trend together. A score in one area is a good prediction of the score in another area. There is no reason why that must be the case a priori, and LLMs are a great example of an intelligent system which is much better at recalling information than it is at reasoning.

Human aging doesn't seem to affect all of these abilities uniformly. e.g. Everyone seems to complain about memory the most (and that matches my experience), but I've been pleasantly surprised how well neuroplasticity and pattern recognition have held.

The series this is from (Free to Choose) is a great introduction to economics for people of any age. I highly recommend it.

This particular example can be misinterpreted though. It's true that no single person knows how to make that exact pencil that he is holding. But it's not true that no single individual exists who can make a pencil by themselves. If the criteria is just that it works as a pencil, then many people could make or find something that fills that criteria.

This is an important distinction because there are things like microprocessors, which no single person knows how to make. But also: no single person could alone build something that has anywhere near the same capability. It's conceivable that a civilization could forget how to do something like that because it requires so many people with non-overlapping knowledge to create anything close. We aren't going to forget how to make pencils because it is such a simple problem, that many individuals are capable of figuring out workable solutions alone.

That's only true in relative terms. In reality tax rates go up or down on everyone at the same time, because that's how the negotiations shake out.

If taxes go up on everyone, the rich are still the ones that manipulate the government, but now they have control over more tax revenue. If taxes go down for everyone, the rich are still the ones that can manipulate the government, but now the government has less revenue and can't cause as much damage.

> That's not true at all. I make a good salary as a software engineer, I absolutely think I ought to be taxed a little more than I am, and would gladly pay that money to live in the better society I believe that would create.

This confusion is precisely why the middle class has less power than ever before. You and many others have been sold a meme that your tax dollars are in service to a greater good, and you are a bad person if you recognize this to be a scam.

At an individual level, for each person in the middle class, 90% of the social programs they pay for are negative EV for them personally. It would be better for each of them if they just kept what they earned, and didn't expect to get it back "later" or "when they really need it" whenever that is. This is an empirical, testable claim, and the math will be slightly different for each person. You should check for yourself.

If everyone turned off the news, and totally ignored the messaging around taxes and government programs and just looked at their own cash inflows and outflows to/from the government, the middle class would retain far more power than they do.

> Politics has it's fair share of corruption too. But at least in my country (the UK) it's the lesser evil.

Is this a widespread view where you live? As an outsider watching the fall of Britain in slow motion, this explains so much.

No they did not. It's easier for a small number of people to coordinate, than a large number. The wealthy have about as much power as the entire middle class, but can wield it better because they are more nimble.

That doesn't change the state of the negotiation, which is that cutting taxes for the middle class will also require cutting them for the wealthy. If you optimize for your own personal notion of fairness, or retribution, you may very well fail to coordinate in your own self-interest.

> This is unwanted, first because it produces individuals powerful enough to topple the people’s will

Making the government resistant to manipulation is a distinct problem. It's a game theory/mechanism design problem, and its solution doesn't require taking in lots of money. Giving the government more power/money causes people to spend more effort to manipulate it, so any weaknesses are exploited to the fullest extent.

> and second because it is not in the interest of society for wealth to be accumulated rather than moving.

This reveals a significant misunderstanding of how capital works in an economy. None of the billionaires that come up when you type in "billionaires" into Google have access to liquid cash anywhere near the number that shows up next to their face. Their money is invested in productive projects, it's paying salaries and invested in equipment. Concentrating capital is what allows a civilization to take on big projects. As a society we want big projects to be paid for by individuals bearing the risk (skin in the game). In a free-market, capital concentrates in individuals who, empirically, know how to use it well. Spending other people's money is a great way to make sure that money is spent frivolously. You can criticize luxury spending all you want, and taxing that is something most people consider "fair", but you aren't speaking for anyone economically literate when you say that you don't want capital to concentrate. I want it to concentrate as much as it does naturally.

> I think they mean "taxation of the too rich" in that case.

Everyone wants taxes to go up on everyone making more than them, and for their own taxes to go down. The problem is this is a collective negotiation, not a discussion about what to ask the genie for when we rub the lamp. If the middle class wants to decrease their own taxes (which is the political issue that objectively affects them the most, and how they lose their power), then they are going to have to meet the wealthy half way. Idealism is the enemy of the the common sense, rational, self-interested move.

> And needless to say, a non-functioning democracy is not a proof that the concept of democracy doesn't work.

Yes, democracy is a good idea precisely because imperfect implementations of it work well. If it worked in theory and not in practice, then it wouldn't be a good idea. Contrast it to communism, which is literally an info-hazard. If you try to bring it in to existence, you won't achieve your goal, and the system you do create will be much worse for you. Even if it works in theory, it's a bad idea because it doesn't work in practice.

Taxation is the mechanism that moves power from the people to the government, and increasingly politicians and their specific interests. Do you actually believe that if your taxes went up, power would be less concentrated, or that you or your countrymen would have more power? Every government goon doing authoritarian dirty work collects a paycheck and wouldn't do their job without it.

This statement essentially boils down to "The only right way to fight me is in an environment where I expect to win"

That's how you know the DeFlock strategy is effective. They aren't playing the game that the CEO wants to play, they are playing the actual game. The actual game is minimizing the impact of cameras that are now everywhere.

Some individuals may take it upon themselves to vandalize the cameras, which can't be planned via conspiracy (that would be illegal), but those radical individuals can be "set up for success" through information. This strategy of creating an environment where effective vandalism is easy, is also part of the actual game.

> This looks a lot like a manager-once-removed undermining a reporting manager by supporting an engineer to go rogue.

Yes, and this is unfortunately the best move in many situations given how hard it is to fire people. The ideal thing to do is fire the incompetent manager who is a tax on their team rather than a multiplier. That unburdens the highly competent engineers on their team to do good work.

However, that might not be possible or even a good idea. A failed hire might reflect badly on the upper manager, or prevent them from hiring a replacement. It might require a lengthy performance improvement plan, which means you have a burdened team for that whole duration. It's often easier to de-claw the bad manager, and take back on their responsibilities yourself. Not ideal, but the best of a bad situation.

> This looks a lot like a manager-once-removed undermining a reporting manager by supporting an engineer to go rogue.

The second red flag in all of this is that an engineer doing good work on their own is labeled "rogue". Management's job is to make people productive, if they can't realize that the best configuration for their team involves some individuals working largely separate from the managers own mediocre day-to-day ceremonies/projects, then they aren't doing their job. People that don't need to be managed are the best kind of people, if you can't spot them and leverage them, get out of management.

I'm picturing a 10 second clip showing a child with a green box drawn around them, and position of gas and brake, updating with superhuman reactions. That would be the best possible marketing that any of these self driving companies could hope for, and Waymo probably now has such a video sitting somewhere.

I once took a philosophy class where an essay assignment had a minimum citation count.

Obviously ridiculous, since a philosophical argument should follow a chain of reasoning starting at stated axioms. Citing a paper to defend your position is just an appeal to authority (a fallacy that they teach you about in the same class).

The citation requirement allowed the class to fulfill a curricular requirement that students needed to graduate, and therefore made the class more popular.

It's not up to me. So I'm not "letting" or not "letting" anyone do anything.

I was stating what I believe to be a true counter-factual. If every western country had 1 nuke, the world would be safer than if a single country has all the nukes.

The west is also not "my side". I have no stake in most western countries, and their success or failure is not something I feel as part of my day-to-day. I'm glad there is more than one, so if something goes wrong I can go to another one.

The west gets special treatment because it is filled with prosperous democracies. Democracies are relatively stable, and rarely do things outside their Overton windows, like launching a nuclear weapon unprovoked. Prosperity is what makes people peaceful. Prosperous people have more to lose. No one in the west wants to backslide towards a state of nature because an invasion or unprovoked conflict went the wrong way.

Okay now factor in the probability of divorce, and the amount you get to keep afterwards, and discount it to present value, vs. paying more taxes and keeping it all. Also remember that you typically lose half of income forever, not just wealth in a divorce.

Yes this is the proverb. I often hear it quoted as A hires B hires C, as a remark on organizational decay. But the original (and the way you phrased it) is a statement about what kind of person each wants to hire out of their own self interest. That's the more insightful version IMO.

> I don't think there's any question at this point that it's in Nordic self interest to develop a nuclear deterrent.

Yes, it definitely is.

> The worst change is of course the fact that the odds of a complete societal collapse have increased dramatically.

A nuke means that anyone who wants to invade you needs to price in a total loss of their largest city as a possible outcome. That is a great disincentive, one that Ukraine probably wishes it had against Russia.

> and the risk of major accidents will increase.

I don't think that's reasonable to say about a bunch of countries getting their first nuke. The concern should be more with countries like the US and Russia that have so many nukes, which they can't possibly use effectively, and don't have the ability to properly maintain.

If every western country had exactly one nuke, the world would probably be much safer than if the US has all of them.

This completely misses the reason why you need to hire the best initially. It has nothing to do with the hardness of your own company's problems. It has everything to do with the distribution of productivity among any kind of engineer.

Engineers follow a pareto distribution. In a normal sized team, with a typical hiring funnel, you will have a few high performers, who are responsible for most of the team's productivity. If you can only hire one person from that team, then it is more likely than not that you will hire someone with productivity below the team's mean. At an early startup, this could be a death sentence. Especially since we typically reason and plan in terms of means, so it may come as a surprise that your single engineer is less productive than the mean of most teams that you have worked with.

The other reason (also not mentioned) is that you eventually want to scale hiring. That means that you need to have people, that you have hired yourself, hire more people on your behalf. The best people (A players in the metaphor) don't have imposter syndrome, they know how good they are, and how good they aren't. They want to work with other talent, that makes their lives easier, more interesting, and less stressful than covering for/babysitting other people. It's also the only way they can grow from where they are at. So they can be trusted to hire more A players, out of self interest.

The median engineer (let's call them a B player) often knows about where they stand as well, and often they will have started to diversify their skillset into organizational politics. They intuit: hiring people more competent than them gives them less leverage, and they are pretty good at zero-sum status games, that's their edge. They don't want competition, so they hire C players.

So the reason you want to start with the best is because it's the only way to ensure you can move fast when you need to, and the best way to keep the organization effective long enough to exit. All organizations decay into incompetence, but hopefully you can get yours and get out before that happens.

AI will replace humans in performing every cognitive task, unless you believe that there is something about biology that makes it categorically better for certain kinds of computation. There's no reason to believe that's the case.

LLMs and specifically auto-regressive chat bots with transformers for prediction will probably not replace engineers any time soon. They probably won't ever replace humans for the most cognitively demanding engineering tasks like design, planning, or creative problem solving. We will need a different architecture for that, transformers don't look like they get smarter in that way even with scale.

The cost of iteration here is so high that we will likely remain in a bioengineering winter until there is a way for individuals to iterate on these compounds in their own self-directed research. We need a ham radio equivalent for synthetic molecules.

Every time I read something like this, it strikes me as an attempt to convince people that various people-management memes are still going to be relevant moving forward. Or even that they currently work when used on humans today. The reality is these roles don't even work in human organizations today. Classic "job_description == bottom_of_funnel_competency" fallacy.

If they make the LLMs more productive, it is probably explained by a less complicated phenomenon that has nothing to do with the names of the roles, or their descriptions. Adversarial techniques work well for ensuring quality, parallelism is obviously useful, important decisions should be made by stronger models, and using the weakest model for the job helps keep costs down.

> there is no entity or force on this planet that can decrypt them.

At this point I think all of the modern, widely used symmetric cryptography that humans have invented will never be broken in practice, even by another more technologically advanced civilization.

On the asymmetric side, it's a different story. It seems like we were in a huge rush to standardize because we really needed to start PQ encrypting data in transit. All the lattice stuff still seems very green to me. I put P(catastrophic attack) at about 10% over the next decade.

This was a fascinating read, thanks for posting.

I'm not totally convinced that the threat model is realistic. The watermarker has to embed the watermark, the only place to do that is in the least significant bits of whatever the message is. If it's an audio file then the least significant bits of each sample would work. If it's a video file then the LSBs in a DCT bin may also be unnoticeable. It can really only go in certain places, without it affecting the content in a meaningful way. If it's in a header, or separate known location, then the pirate can just delete those bits.

The threat model presented says the pirates have to go with one of the copies, or only correct errors that are different between 2 copies. That's the part that I don't think is realistic. If the pirates knew that the file was marked, and the scheme used to mark it, but didn't know the key (a standard threat model for things like encryption), then they could inject their own noise into wherever the watermark could be hiding, and now the problem is the watermarker trying to send a message on a noisy channel, where the pirates have a jammer. I don't even think you have to sacrifice quality, since the copy you have already has noise, and you just need to inject the same amount (or more).

I seem to remember Yahoo finance (I think it was them, maybe someone else) introducing benign errors into their market data feeds, to prevent scraping. This lead to people doing 3 requests instead of just 1, to correct the errors, which was very expensive for them, so they turned it off.

I don't think watermarking is a winning game for the watermarker, with enough copies any errors can be cancelled.

It's not AI. Human institutions rely on humans acting in good faith. Almost every institution is held together by some kind of priesthood, where everyone assumes the priests know what they are doing, and the priests create a kind of seriousness around the topic to signal legitimacy, and enforce norms on the other priests. This is true of Government, Law, Science, Finance, Medicine, etc.

But most of these institutions predate the existence of game theory, and it didn't occur to anyone how much they could be manipulated since they were not rigorously designed to be resistant to manipulation. Slowly, people stopped treating them like a child's tower of blocks that they didn't want to knock over. They started treating them like a load bearing structure, and they are crumbling.

Just as an example, the recent ICE deportation campaign is a direct reaction to a political party Sybil[0] attacking the US democracy. No one who worked on the constitution was thinking about that as a possibility, but most software engineers in 2026 have at least heard the term.

[0] https://en.wikipedia.org/wiki/Sybil_attack

> Believing something is mispriced simply means the perceived risk is very low (or even negative)

That's not what risk means. Risk is the amount you stand to lose, and expected value is the mean outcome (calculated by integrating across all possible outcomes). The EV can be positive or negative, and that corresponds to a mispriced asset. Risk can only go to zero, but it can also be arbitrarily high.

Managing risk is its own thing, but to give an intuition: How much should you bet on repeated fair coin tosses that pay out 3:1, if you have $100? More than 0, because it is positive EV, but definitely not the whole 100. Risk management is about playing vs. not playing and bet sizing.

Gambling is taking on risk without positive EV. A 1:1 cointoss is the quintessential example. EV is 0, you won't make or lose anything in the long run playing this game. But it adds uncertainty, you could have twice as much or nothing. Gambling is risk-seeking, the variance of outcomes has value (to a gambler) on its own.

CS degree is starting to become a bimodal distribution. There are enough people who thought they could buy a high-salary job by getting a CS degree, and universities advertising to that effect, that the market is now flooded with candidates who have degrees on paper, but don't have the mojo.

Their brain doesn't work like a hacker's, and they would have to work very hard to compensate, but they got into this for the easy high paying job, they don't want to work hard.

Somehow other degrees seem to be better predictors of competency. A lot of physics/math folks, and various non-software engineers realized that they have a hacker's brain, and programming pays more than what they were doing, so they got into software.