ngriffiths

When I found myself using a corp-speak term, I questioned whether there was another word that felt more natural and communicated the same thing.

In other scenarios the non "corp-speak" term feels kind of wrong and the "corp-speak" one is actually better though. I feel like this is all just a realization about what the goal of the business actually is and a realization of just how many different goals you could have. In this case it sounds like it's just to do some good writing, not to grow for the sake of making more money or whatever else. And that makes a huge difference in how you talk about it of course.

Also if someone used the word "lead" I'd conclude that they... work in sales? How can that mean anything about whether they'll be helpful to you or not?

Ok one more thought: sometimes the corp-speak feels way less stuck up than "oh actually, I'm a writer. I measure progress not from other people being willing to pay me some money, but in terms of my own highly refined taste. I have many leather-bound books, and..."

By canonization, I mean the process of taking a local, one-off formalization and turning it into library mathematics: general, reusable, coherent, efficient, and compatible with the rest

I think this kind of work is constantly misunderstood and undervalued. I don't really see it as a binary thing, more like a complex skill that most people are terrible at, some are good at, and a handful of giants use to be just ridiculously productive in their field.

It reminds me of hedgehogs and foxes - foxes tend to be bad at making one off progress on their own, but are critical for accretive work.

Also I was reading a textbook the other day and thinking wow, it is absurd how much more valuable these things can be than other resources, and it's exactly because they canonize. It would be a massive loss if they stop getting written.

It's complicated. If you go put some dollars down on a prediction, the median person you're in it with doesn't have much evidence to support their position and is basically gambling. The median dollar you're up against is coming from some rich guy with knowledge somewhere between "airmchair expert in this topic" and "outright insider trading." And so it's a situation where for some people it isn't gambling at all; it even generates some useful info accessible to everyone; and it's easy not to notice that only a handful of rich people really benefit, because it feels like anyone "can win." The truth is kind of counterintuitive.

Also, having expert knowledge and then making money off of other people thinking they might have it and being wrong... is a choice.

It's also super easy to apply it wrong because going above X% in one area normally means sinking below X% in another. I think a clearer way to say it is that sometimes, you have to be almost perfect, and 98% could sound like almost perfect but it's way too low. But definitely the things you don't need to be perfect far outnumber the ones you do.

Fair, but it's hard because falling behind on the models themselves is a killer. Maybe they're saying they can't spend resources on anything except staying ahead.

One way of reading this is an article about how good Anthropic's product is. "Look at how many serious flaws users have been willing to accept in order to keep using this thing"

But they should just stop reading. It's actually not ok that it's unfamiliar, because makes you reread and get confused and distracted, all for some silly reference that doesn't make a big difference. Life is short! You can read the hard stuff when it's worth it, and just skip the rest. Surely that's the most common thing to do.

The answer is definitely still a big no, but for me the reasoning is because it will make it worse. And you apparently aren't the target audience anyway, so why should I care if you stick around.

(Whereas in the case of harry potter, the goal was to sell books, not just to produce something good).

At my job I switch between writing analysis code for research projects and writing code for apps. The difference in mindset is so dramatic. In the same way that good software has consistent names and interfaces that are ~useless when you just need the code to run once, research code has its own requirements that are ~useless in software. It's honestly a big challenge to switch back and forth. So I think it just reflects the main skillset of the people who use it (caring is not enough).

It's more exciting to think about evil people plotting how they will control the labor force and steal all its knowledge, but I think both the AI use and the alleged extraction of knowledge of design and craft are better explained by, like, "the job got crazy popular, the labor force multiplied, a lot of less passionate people got involved, and then some solutions were found"

Yeah, because it covers all the doctors. It is literally the same as saying you will pay any price. It is a luxury good.

It’s not hard to understand why people hate health insurers. When you interact with the U.S. health care system, the providers — the hospital staff, the doctor, the nurses, the technicians — all just take care of you.

Your interaction with the health insurer, on the other hand, feels like a struggle against an enemy who wants to destroy you.

Exactly, the big factor driving healthcare costs is that people like expensive providers. They like the fancy plan that covers their doctor, that doesn't limit them to a community hospital. Any choice that involves retaining even the slightest market leverage is a deal breaker for many people. I think all the clever stuff we do to optimize costs and increase transparency is worth doing and helpful, but like, it will still be expensive. We want that!

Meanwhile it is not really better to go to an expensive doctor, in terms of health outcomes. So the public health problem is to stop people (and employers) from voluntarily wasting their money on healthcare, so we can use it for better stuff.

The LinkedIn post: https://www.linkedin.com/pulse/incorruptible-spotify-brendan...

I'm not just talking about where the initial pressure comes from, but the effects of the changes. I guess the question becomes about how easily these layers get disentangled. E.g. things investors like that then result in lower sales? Things that result in greater sales but that the customers don't actually like?

Naively of course it seems like investors don't like it when sales go down, so there'd be an extremely tight link between financial market and product market feedback. But I imagine you disagree, that this breaks down easily and creates problems?

My biggest struggle with this question is that "going bad" sometimes coincides with not just financial incentives, but also more people getting value out of it. For example Spotify gradually shifting from "we make it easy to curate and share playlists" to "we make them for you to use as background music constantly." Sometimes what's bad for the early power user is great for the late adopter, and it's difficult to make any kind of broad judgment about whether the change is better or worse.

What do you say to this interpretation? In particular do you think most cases could be framed as "the key audience/customer/market has shifted"? Is it possible to find greater financial success while doing things the primary audience doesn't like?

Yes, and there are sometimes many layers to it, which is why you can think "cool, I get that" while still missing something important that would be obvious to an expert.

So be it? Everyone under 30 being permanently worse off due to a decline in education is an extremely depressing outcome, that seems like the whole argument for fixing it

It is not written though

The students who cannot read a 20-page article today are the voters who will not be able to read a bill, or the jurors who cannot follow a closing argument, tomorrow.

Obviously literacy is super important but these are examples of things where literacy plays very little role, because ~nobody can read a bill, or follow a written legal argument. I mean a very literate person can get something out of reading it, which is nice until they then completely misinterpret it, or hear what their friends say about it and get onboard purely based on vibes.

I feel like it matters more for the economy and the future of knowledge work which, uh, is a little uncertain these days.

I don't know. I used to feel this way about IDE autocompletes/suggestions. Now they are widely used, and it doesn't necessarily seem hostile. It's not that hard to imagine the same thing could happen here.

The key to surviving in such an environment is to let go of your ideas of the truth. The customer doesn't want to hear it, and doesn't want to know it.

This is exactly it!

Like you might think "the promised features are not feasible." No, the features you will soon deliver are feasible, on account of you're about to go build them! If you fail, that is still very bad. But the point of rule 1 is you don't have to act like you signed up to deliver exactly X feature on exactly Y date. Instead you can think a little bit, and then you calmly set off on a process that should reasonably end up with the customer being happy. To many people this strategy feels like lying.

In addition to covering the IPO in general last week, Matt Levine also wrote about this specific question Tuesday[1]:

Historically index providers were in the business of making these sorts of quality decisions, so that index funds were not forced to buy stocks they didn’t like.

These rules create some tension between the idea that an index is a list of all the stocks and the idea that an index is a list of all the good stocks. Historically, it didn’t matter all that much: The point of the stock market is to tell you which stocks are good, so a company with a high stock valuation should be a very good company, so it should get a high weighting in both the Index of Good Companies and the Index of All the Companies.

But SpaceX — and also maybe OpenAI and Anthropic in their coming IPOs — will probably break that link. SpaceX will probably (1) do all sorts of stuff that index funds hate and that index providers have specifically tried to exclude and also (2) be gigantic, because the market loves it.

[1] https://www.bloomberg.com/opinion/newsletters/2026-05-26/ind...

That's interesting, thanks. I only read the abstract so far but was immediately reminded of this recent HN submission[1] and the whole thing that certain ideas go together, and so they are adopted together, but the resulting bundle of ideas might be poorly suited to certain problems.

[1] https://news.ycombinator.com/item?id=48237163

The original decision is interesting because at first it seems very stupid (as acknowledged right there in the article). It's a more expensive way to do the same thing. But man, what a sales pitch, not only for their own customers but also to employees. The feeling is that the company values its people and is willing to really depend on them, and look, it actually paid off when they did that.

I think it's common to claim to care about the people without really depending on them for much (like with perks) or to depend on the work but treat people badly, and doing both is hard.

Yeah, I mean it's obviously meant to be a marketing pitch but it's not a very good one.

The hardest computational problems are not waiting for faster chips – they are waiting for machines that compute in a fundamentally different way.

Surely they don't actually believe that, right? Like you say the benefits must be limited to specific shapes of problems (not all of "the hardest" ones), and the whole history of computing is about how faster chips is an excellent answer to difficult computational problems.

Exactly, and correlated errors, where a polling error in one state predicts similar errors across the board.

I disagree that it's all pointless though. Most basically it's smart for campaigns to have a good model and let that inform strategy where appropriate. Since the president is a big deal other people's decisions are also impacted, and in the long run it pays to have good predictions of those chances. Also, the outcome sometimes is fairly certain and that isn't always easy to see.

Discussion of stats models is always complicated by the fact that a lot of people will read "30%" as a "no" prediction and claim your model is wrong if the thing happens. On the one hand, one strategy is to "hide" the numbers a bit behind a blaring headline that says "we are not sure!!" It's a bit of an art to decide when to be "sure" or not. On the other hand, in research for example you can just say screw it, I care if the correct people are correct, not if a bunch of wrong people are wrong.

I feel like the correct strategy for 538 when it was actually niche was to be precise, but then it went viral and maybe should've hit the IDK button much harder and more often after that.

The thing jumping out at me is these really are mini businesses (even though they are bad). Combine it with the main idea in "Emacsification of Software" (from recent HN front page [1]) and I guess you end up with lots of nerds running their own customized mini businesses?

It's sorta wild to think about. Am I the owner of the custom radio station my AI agent made, and does that mean I get paid for listening to the ads?

Maybe the cost of computing and running the station means it still needs a decent following to break even, not sure how the numbers work out.

[1] https://news.ycombinator.com/item?id=48118727

I would have been all over this if we had one of these when I was in school. Very cool project.

I think there are many ways someone with his lack of expertise can still be valuable, including:

- Making connections to other subjects that an expert would miss. The hall of fame of sigmoid predictions is just excellent, I already know I'm going to be reminded of it some time in the future. Very entertaining way to get the point across.

- Writing about tricky concepts in a very accessible and elegant way, which experts are notoriously bad at doing themselves - they are often optimizing for other specialists.

- Being able to write with an air of speculation and experimentation with ideas that experts and institutions often can't afford. Experts have to maintain their track record; Scott Alexander can say "lol just double the timeline"

As a fun exercise replace AI with "junior" and "junior" with "mid-level." It holds up pretty well, as a manager you have responsibility for the work your team does and "make everyone put in more hours for no reason" is dumb. Maybe it comes across a bit neglecting of the "juniors" (in particular, it doesn't show any desire for figuring out ways for AI/"the juniors" to grow their responsibilities in a sustainable way).

Imagine reading that version as someone who doesn't know how big companies work. "But then they'll just fire all the mid-level managers, since they don't do any of the actual work!" Haha, boy would you be wrong.