HNHacker News
TopNewBestAskShowJobs

ta_1138

356 karma · joined December 29, 2023

submissionscomments
ta_1138··on We need visual programming. No, not like that
As projects get bigger, things might get sadder. I worked at a certain large SF company that uses a lot of ruby, so most development was repl-based too. But this wasn't a boon but a curse, as the total lack of data format guarantees on the very large, critical monorepo meant a lot of uncertainty. What does this method really do? I guess we have to run it! It worked for this specific input... but will it work for any and all inputs that get to this data path? Let's hope so, because we aren't sure! The company spent massive amounts of money on servers for parallel testing, just so that the suite could run in less than a few weeks. And when you need a large test suite to have a chance, most of the advantages of the REPL vs a compiler have been lost.

Eventually they did the same thing you can do in common lisp: Add so much metaprogramming that doublechecks invariants, it might as well be a compiled language.

ta_1138··on Color Wheels Are Wrong (2011)
yeah, anyone that paints will care the most about other things, like which colors they actually have available, and the quality of the pigments. If you take, say, most single pigment reds that aren't very toxic, you'll see that they have relatively low coverage: Lay it over a dark color, and you aren't really getting red, but filtering the darker color with some red on top. Pure pigments however kind of mix into reasonable colors that more or less resemble what a color wheel will tell you, but many paints are reached through mixes, and mixing mixes is often not getting you the colors the wheel would predict: Chances are you are getting something pretty brown, with very low intensity. You can make fun tricks for students, like showing them two pairs of almost indistinguishable blobs of paint, but that mix to something different, because two of the colors were pure pigments, and the others a carefully made mess.

And then there's not just color, but finishes, and the body of the paint.

So yes, once actually talking colors in a real piece, while the color wheel is a very useful tool for determining pleasant schemes, just like in a website, your real primary colors you think about are never a bright yellow, a bright red, and a bright blue.

ta_1138··on Supreme Court rules ex-presidents have immunity for official acts
Let's imagine that a president decides to, say, stop aiding Ukraine, and in exchange, Putin promises said president a couple billion dollars.

Stopping aid to Ukraine is definitely an official act, which, on its own, is within their powers. Trading said official action for the billion dollars is what is fraudulent, but if I am understanding the Roberts opinion correctly, it cannot be prosecuted, because the motives of the president should not be used as part of the ruling of immunity. So... free bribery, as long as said president pays their taxes.

ta_1138··on From Infocom to 80 Days: An oral history of text games and interactive fiction
This nonsense still happens in the early sierra graphic adventures: Administer sobriety test!
ta_1138··on Read the wild email Tesla is sending to suppliers amid Supercharger chaos
Bicycles win when the level of density is still relatively low. If we are serious about efficiency, bikes don't make that more sense. Sufficient density so that most activity can be done on foot, and the rest is mass transit.
ta_1138··on Asian American women are getting lung cancer despite never smoking
Individualism in western countries varies quite a bit. In cross-country scores, Spain has numbers far closer to the middle east than to the US, and then Peru is closer to China than to Spain!

So your invdidualism argument is only strong if by western, you mean the anglosphere, the Netherlands and Belgium.

ta_1138··on The darker side of being a doctor (2017)
Pilot failures are really visible, and kill hundreds atll at once. Bad medical decisions lead to more deaths overall, but it's far less likely the public will hear it, and even if they do, it's one person at a time.

Far more people die from car crashes than airplane crashes too: It's not even close. And yet, people who have no business behind a wheel drive, and the penalties for being a poor driver are typically minimal. Same difference.

ta_1138··on Some colleges will soon charge $100k a year – how did this happen?
Not at all.

Go to one of those almost-free european colleges, and then compare the experience to a US college. The US colleges expect most students to come from far away, live on a big, expensive, beautiful campus and have massive efforts in student lifestyle. Sport facilities, museums, academic support, actual office hours... it's closer to expensive, private EU universities. Most EU universities don't do that, and they are ran to cut costs. The expense per student is low, on purpose.

They might both be called universities, but it's completely different goods. The transformations of European schools to be available to the masses just didn't happen in the US, and therefore the prices are through the roof. Give any American university the budget per student than a German or Spanish university has, and they'd have to just close, because they aren't built for it.

ta_1138··on Some colleges will soon charge $100k a year – how did this happen?
Yet many people leave college with few connections and little networking ever done. So by your standards anyone without those social skills should consider college as a total waste of time and money. College: only worthwhile for extroverts.
ta_1138··on A lavish lifestyle strains credibility (1985)
It's a common problem in all kinds of low-perforing governments: You get extremely tough regulation that makes it very difficult to do any economic activity legally, and then you can bypass said regulation with corruption. The high barriers to entrepreneurship, or just constructing a building or buying materials are a feature, as it means more advantages to those that can ignore the regulations.

We can see this even in relatively well functioning states: Say, when you make infill development expensive, but provide tax abatements to compensate, but said abatements must be approved by a board of aldermen or something like that. Being able to control who builds more cheaply is valuable, an then you just make sure that you only vote yes to good friends that patronize your cousin's consulting business.

The less discretion the bureaucrat has, the harder it is for them to be corrupt. And in the soviet union, they had a lof of discretion.

ta_1138··on Stable Diffusion 3
Unfortunately production capacity for that is limited, and with sufficient demand, all pricing is an auction. Therefore, we aren't going to be seeing that card in years
ta_1138··on Disney to take $1.5B stake in Epic Games
Disney had a group looking at NFT and metaverse initiatives, but as far as I have seen, they all got canned in last year's layoffs. manager and everything. The company is huge, so it's not impossible that other group like that might remain employed there, somewhere, but at the very least the interest have waned.
ta_1138··on Netflix: Piracy is difficult to compete against and growing rapidly
It's far less profitable than you'd think. Many of the competitors in the industry lose money.

Licensing fees/production costs/royalties eat a bunch of the money, but the infrastructure to run a streaming service isn't cheap. I know at least one of the competitors that people would call the most successful is paying a 9 figure AWS bill, plus whatever the costs are for all the caching setup that makes most popular content live in ISPs. Add the typical army of developers building apps for the mobile devices, the billing team, people doing recommendation engines, fraud detection/security team, tagging all content, and translating every blurb in a bunch of languages, possibly pay for creating subtitles for all of those languages... it's not cheap.

What is so frustrating about this is that so many of the costs would go down with consolidation. Every company has to handle the fact that age verification and privacy legislation in South Korea has to work in a just-so way, but every streaming service has to duplicate the logic. While the Netflix recommendation system probably can use some tweaks to take on all of D+'s content, I bet it'd be easier to tweak their system than to have every company have their own ML team handling recommendations. A lot of relatively low quality content wouldn't have to be made if we didn't need to make 'filler' to keep people spending sufficient time on a given streaming service while the next top quality release comes out.

In a happy world, competition brings more better content at a cheaper price for consumers, but when I look at the state of streaming, what we have is many competitors that aren't breaking even, in exchange for a streaming experience that keeps getting worse. And there's little chance it's going to get better until at least half of the worst competitors give up and go back to licensing their content to whatever the big 3 end up being.

And yes, we'd all be better off if we moved to the music model, where any subscription has 90% of the content, but do you really see, say, Apple, Netflix and Disney doing worldwide, full catalog cross-licensing deals?

ta_1138··on Why new hires often get paid more than existing employees
I've seen that attempted: I worked at a very well known SV company as it was growing. When we looked at what we were asking to even get an interview, we saw that most of our best workers could never have a prayer of passing our resume requirements when we hired them. But we had so many people applying, we could really ask for the moon and still get enough resumes.

But the fact that on paper we were getting better candidates didn't ever lead to better performance: Past a certain point, not hard to reach if you pay relatively well, the correlation between interview performance, resume and actual performance on the job is basically random. Hell, I've seen a team change managers, and seen the stack rank be basically opposite for both: One guy wanted to give a large bonus to the guy the other one wanted to put on a PiP.

Thinking that, on the interview, your candidate is magically better than your average coworker takes a lot of optimism, or a lot of pessimism about your coworkers.

ta_1138··on Why new hires often get paid more than existing employees
The reason is pretty easy: You understand the value of your workers better than a random employer, because they work here. You know when someone is a bust, and when someone is a star player. If you don't give raises to your best people, they are the ones that quit first, and when people quit, they don't quit one at a time: If I get a major raise by changing youbs, you can bet all the coworkers that I respect will hear about it, and I'll want to take them with me.

I suspect that losing a high percentage of your best employees all at once because you are paying them quite a bit less than what you'd pay a worse replacement is a bad thing you want to avoid.

ta_1138··on Why new hires often get paid more than existing employees
In my very long career, I have yet to hear a good story of someone taking said counter offer and staying very long, for one reason or another. I think it's a bad idea to offer it at all, as it sends a very clear message of what behavior is rewarded. When other employees learn that you will offer raises to counter offers, expect everyone to start interviewing.

Every manager knows which team members are key to the team's performance. They have to get raises or bonuses to match.

ta_1138··on Teach Yourself Programming in Ten Years (1998)
It's often hard to tell what information is truly fundamental, and what will age badly. I look at many books of software engineering that were popular in the early 2000s, like The Pragmatic Programmer, Agile Software Development or the Design Patterns book by the gang of 4. Those books aren't taking about technologies that we expected to be superceded in 5 years. And yet, a lot of the text regarding object orientation and design patterns is outdated at best, and outright harmful at worst.

One can also look at, say, concurrency. A topic often seen as unimportant in the 90s, but ultimately taught at the lowest level, with mutexes, threads, semaphores and such. Those aren't going away, but how often is concurrent programming all about manually setting mutexes? We use higher order abstractions, but those change too. Many languages are encapsulating this in monads, whether people know their promise implementation is a monad or not. But that's not the only way, and the most popular way can change: Maybe languages in 10 years will be all about continuations and direct styles. Maybe it'll be something else. The fact that in the end, the same basic features from the 80s and 90s will exist at the bottom somewhere isn't that useful for most programmers.

And html generation... I was writing UIs before web apps were there. A lot of things that seemed fundamental went away when the browser was embraced. But will the browser live forever, or be superceded? I suspect that it will all get replaced, or be utter legacy, eventually. Will that eventually be 10 years? 30? It's very hard to say what will remain fundamental, and what will not.

ta_1138··on Twitch to Cut 500 Employees, About 35% of Staff
What nobody tells you is that the cost of hosting all that content, and streaming it live with little lag, is far more expensive than many would guess. Netflix can have boxes with cached content very near you, and your ISP even likes it, as the most streamed content tends to be the same for many users. On Twitch, it's not so easy: You really want everyone to be seeing things with minimal lag, so they can spend all that money buying bits.

Other streaming services, which can use all of the tricks to save bandwidth and compute, have eye-watering AWS costs. Even with their high take rates, I bet there's a lot of streamers that, once accounting for their share of infra costs, end up losing twitch money.

ta_1138··on U.S. cities opt to ditch their off-street parking minimums
Of course infrastructure is subsidizing those who use it. Even more so when it comes to infrastructure that takes a lot of space, as every square foot of surface dedicated to one thing cannot be dedicated to another. And infrastructure induces demand by lowering some prices relative to others.

In an extreme case, imagine that your typical, currently congested highway gained another 6 lanes in each direction, and we raise speed limits to 100 mph. First, the right of way is now way bigger, so we need to wreck a lot of buildings for this infrastructure. It makes being able to travel near that highway really nice for a while: Minimal congestion! A straight out windfall by people owning land near it. But that also means that the area redevelops, as farmland that was in the middle of nowhere is now a reasonable commute downtown! But as the development continues, the highway will fill up. And since downtown streets were not changed, it's the area downtown near the highway that is now a congested mess, as the limitation is getting from there to buildings.

And again, a subsidy for you can be a harm for me: If every street is 8 lanes wide, it's very hard to cross. Every pedestrian crossing added is also infrastructure, but it slows down traffic. Every bit of lawn you use is another bit of distance I have to wade through to get to my destination, vs an area with the same number of houses, but without lawns.

One can like one tradeoff more than the other, but the fact that infrastructure means subsidy, and that urbanism choices are picking some people's interests over others is just factual.

ta_1138··on Disney, Warner, Comcast, and Paramount are contemplating cuts, possible mergers
Netflix spends a lot more on per developer, and has been working on the UI for longer: See, for instance, how a big percentage of manpower at Disney has been spent going from two completely different UIs between D+ and Hulu into a single one, so every feature didn't have to be made multiple times. Those kind of projects, with a quantity-over-quality approach to hiring, tend to take forever.

Netflix also has an advantage in cost management, also due to better engineers that are also better paid, and have been managing their system for far longer. The AWS bill for D+ is not the kind you get when you have good chances to be profitable. I bet a lot of engineering in the next year or two will be used just more efficient use of compute and storage. I bet many teams there are spending far more on servers than on people.

And if D+ could be doing better, and is still unprofitable, imagine the situation in other competitors that are even newer. The total cost of the infrastructure, without even counting the licensing or the marketing, might cost the more than people are paying.

ta_1138··on Disney, Warner, Comcast, and Paramount are contemplating cuts, possible mergers
You can be mean to people too! However, if you want to model the corporations correctly, you'll not go very far by thinking they are soulless machines. Most of the reasons corporations fail is people making awful decisions than help themselves personally, and harm long term corporate value.

As for low-quality content, I work with people that have made very bad entertainment. Nobody starts out thinking they'll make something bad. There might be some lowering of standards when, say, the voice acting manager and the voice actor decide that it's not coming out right, and there's no budget to try another 3 days.CGI gets planned for some amount, and when time and a half that is spent, Leia still looks like she is made of plastic. Maybe the writing sounded good to a decision maker, but they had terrible taste. Nobody tries to make it bad, but sometimes we get to choose between bad, and probably still mediocre, and massively unprofitable.

That was Disney's problem this year. A lot of content was super expensive: Go look at the budget for Dial of Destiny, or the per-episode cost of she-hulk. So high that they'd have to be the most viewed things ever to have a prayer at being worthwhile. People were trying their best, and the output wasn't a total embarrassment... but they were all money losers.

ta_1138··on Disney, Warner, Comcast, and Paramount are contemplating cuts, possible mergers
If we didn't have at least 3 major services, ultimately said services would have made contracts that would have taken most of the value of the deal: That's basically what Netflix was doing when they were the only game in town. We have too many streaming services now, but a monopolist with sensible prices was never the way it was going to end.
← PreviousPage 3 of 3