HNHacker News
TopNewBestAskShowJobs

gjulianm

4,409 karma · joined November 21, 2011

Mathematician and computer engineer.

http://twitter.com/gjulianm

[ my public key: https://keybase.io/gjulianm; my proof: https://keybase.io/gjulianm/sigs/nfNutn3AksS22AzUBMHIkR3wqz46pUi-dOCF6GN2SbY ]

submissionscomments
gjulianm··on Rapid Covid tests miss 90% of asymptomatic cases
> Not to mention, I think people are going to look back at something with a 1% death rate and wonder why we panicked so hard.

I'm sorry, but what the hell are you talking about? In my city there were so many dead people they had to transform a skating rink into a morgue. ER rooms were completely overcrowded, COVID took so much effort to fight in hospitals that other diseases got later diagnoses and treatments, which resulted in even more dead people. Lots of elder people died in care homes because there wasn't any help available for them. There were no ambulances available at all. My mother had to go the hospital due to her infection worsening and she had to stay in the hospital gymnasium in couches, because there weren't enough beds nor space for them. They had to make an improvised hospital in a conference center to hold all the hospitalized people, thousands of them. We had a complete lockdown and it still took several weeks for cases to start decreasing.

We panicked so hard because it was fucking scary. Maybe the reason it wasn't for you is that you lived in a place where the government took precautions before it got too big, but I don't think it's ok to be so cavalier about it when so many people had such a hard time.

gjulianm··on Ask HN: When were flagged submissions/comments added to the profile page?
It does show removed flagged submissions if you have "showdead" enabled in your profile.
gjulianm··on Amateurs obsess over tools, pros over mastery
In instruments, going for the absolute cheapest one is almost a recipe for disaster. I don't know about violins, but I had piano classes a decade ago and now I'm with clarinet classes. In both cases, the cheapest option would have push me away from the instrument. In pianos, the cheap electric ones are very crappy, usually don't have the full key range, the weight of the keys is non-existent, and sometimes even the size is wrong, not to mention the sound. For clarinets, you can find ones for less than 100 dollars on Amazon, but they sound like crap, the keyboard is fairly bad and they fight you when you want to make a decent sound. If you buy one of these, the most likely outcome is that learning is so hard and unrewarding that you end up abandoning it after a few months.

Usually the best idea is to start out with the entry level range of established brands. I'm not up to date on pianos, but for clarinets that's $500 minimum. Yes, it's more expensive, but if you have at least a basic commitment to learning it's going to pay off. The other option if you aren't sure is getting a loaned or rental instrument, but please don't buy cheapo instruments because that's just throwing money down the toilet.

gjulianm··on If PEP 703 is accepted, Meta can commit three engineer-years to no-GIL CPython
> Meta, and big companies with a specialized team might be able to exploit it by making sure their entire stack is GIL-free.

Not really. Most of the popular libraries will have nogil versions ready for sure (numpy, pandas, pytorch, etc). Server applications will probably benefit from being able to serve multiple requests in a real multithreaded environment without many changes (in fact, I bet a lot of HTTP libraries will see simplifications due to not needing to deal with processes for real parallelism). Even packages without changes will probably have at least a small degree of thread safety (for example, two instances of the same object should be independent and could be used in different threads) that allows users to still leverage parallelism.

gjulianm··on If PEP 703 is accepted, Meta can commit three engineer-years to no-GIL CPython
My experience with the multiprocessing module is exactly opposite. It's been a few years since I last used it seriously (I still use Python a lot) mainly because of how bad it was. The main issue is precisely communication between processes: Python uses the pickle module to serialize/deserialize data, which not only adds quite a bit of performance cost, but it's very hard to debug when for some reason some object is not supported by pickle (and it's a common occurrence).
gjulianm··on If PEP 703 is accepted, Meta can commit three engineer-years to no-GIL CPython
Multiprocessing in Python isn't exactly like forking on other languages. There's no shared Python memory, it's just a new interpreter instance, and then you have to use sharing mechanisms (shared arrays/values, pipes, queues) to transfer data explicitly.
gjulianm··on If PEP 703 is accepted, Meta can commit three engineer-years to no-GIL CPython
As far as I know PEP703 includes provisions to make operations on containers thread-safe. That should at least avoid most crashes.

Borrowed references can be more problematic, but it seems most cases could be fixed by replacing GetItem with FetchItem calls.

Overall, as another C Python extension, I don’t really think it’s going to be that much of a pain. In fact, I could even get away with no changes (other than build fixes and such) if, for example, I guarantee that each “main object instance” is only accessed by one thread, and I’d still get a lot of benefits from the nogil.

gjulianm··on If PEP 703 is accepted, Meta can commit three engineer-years to no-GIL CPython
I’m not sure that most will require rewrites.

First, if the extension isn’t being used in a multithreaded environment, nothing should change. Yes, it isn’t thread safe, but it doesn’t really matter in that context. And given how bad GIL-Python works with threads, I doubt the majority of extensions are written for multithreaded applications.

And the ones that are written for multithreaded are probably already releasing the GIL for long running computations, so they should be written with at least a little bit of thread safety in mind. And in the worst case it shouldn’t be too difficult to hack the code to include a GIL-like lock that needs to be held by library users in order to ensure thread safety without really changing the architecture of the extension.

gjulianm··on “Greedflation” is a nonsense idea
Just a quick example: BP almost doubled their profits in 2022 even when the price oil didn't really double with respect to 2021 prices. https://www.bp.com/en/global/corporate/news-and-insights/pre...

> But if your profits increase by 5%, and inflation increases by 9%

One should actually look not at inflation but at the increase in costs of that company. Not all companies are equally exposed to inflation, they will suffer different price increases.

gjulianm··on “Greedflation” is a nonsense idea
At no point does this article argue the idea that corporate profit increases are a significant driver of inflation. It's just whataboutism: "but governments did X!" "but consumers did Y!"

The thesis that "shortages create higher profits" is also not proven at all. Yes, in a shortage prices get higher, but at the same time there's less sales. Not to mention that downstream business also suffer from higher prices.

In the end this article comes from a place of treating capitalism as a machine that should be left untouched. Governments, consumers and workers must behave and fall in line, but God forbid questioning who's taking home the money that's mobilized to compensate for certain issues (COVID, war) and what could be done to stop that. Apparently it's never the fault of companies, as if they were some natural phenomenon whose behavior can't be changed.

gjulianm··on Titanic tourist sub photos show wreckage being brought ashore
It's simplistic because "voting wrong" assumes that voters can vote right and ignores problems such as limited party systems (see the US, where a lot of people feel their only choice is between "bad" and "worse"), lobbying, limited information, media influence... Not to mention that even if you "vote right", other people might "vote wrong" and thus your efforts are void.

It also assumes that everything can be realistically changed by "voting", as if legislation wasn't a slow process that's error prone and subject to manipulation, and also sometimes limited. See climate change, for example. Shell knew about climate change in the 70s and kept that information hidden. How could someone in the 70s could have realistically voted against that?

And finally, it assumes that people can vote in the countries that are doing "the bad things". I don't live in the US so I had no way to vote against the US pulling out from the Paris agreement, even when that decision will affect me. Someone in a sweatshop can't vote in western countries to force companies to stop using forced labor overseas.

So yeah, dismissing the influence of the ultra-wealthy in how the world is run and what future is going to be left for us because "you could have voted right" is incredibly simplistic.

> They may influence who runs and donate money but even that is a matter of laws that were loosened or can be passed to fix them (e.g.:citizens vs united)

So wait, citizens can be blamed for "voting wrong" but billionaires can't be blamed for influencing who runs and with how much money? Specially when those politicians supported by billionaires will definitely stop by all means necessary those laws that would take away some of their power? That doesn't seem fair.

gjulianm··on Titanic tourist sub photos show wreckage being brought ashore
> Their parents voted badly and they vote badly and then blame anyone but themselves in this democracy. Can't even blame billionaires controlling the media because everyone has a smartphone now. They all stay in their bubbles and hate each other and everyone outside that bubble and blame anyone but themselves.

That does seem like a very simplistic take. How do you fit climate change into that? Or rising inequality? Or opaque lobbying? Or, in the US, an electoral system that just gives two options to people? Or the housing market going bonkers almost everywhere?

gjulianm··on Titanic tourist sub photos show wreckage being brought ashore
It's not as much as "celebrating someone's death" (you can't really celebrate the death of someone you didn't know even existed) as much as pointing out the idiocy of paying 250k dollars to see the Titanic in what can be only described as a MacGyvered cylinder of death, ignoring regulations, advice from experts, and even basic engineering. People being killed by their own hubris has usually been a source of comedy.

> This stuff combined with sentiment/hatred towards people who are politically hostile/disagreeable paints a picture of a bleak future for this society.

Well, something that might add context for you is that a lot of people already think the future for society is very, very bleak; and they think that the ones to blame are just like the ones in the submarine, specially the CEO. So it's no wonder there's not much empathy going around for them.

gjulianm··on It's weird that people get mocked for changing their minds
> Look at when a public culture figure changes their ideas about something in a way that breaks our image of their identity- Jim Carey starting to make paintings or Lewis Hamilton designing clothes. If someone is well known for something then it seems like there’s a general aversion to that person expressing their thoughts and opinions in a different domain.

Usually the issue is not about them doing something that breaks their identity but about using their position of influence to get an advantage with having far less skill and talent than the competitors that didn't start out famous.

gjulianm··on Clearing 500k feral cats from New York’s streets
Another, more reasonable hypothesis, is that you were wrong about your preconceptions of cat reproduction. The time to reach sexual maturity seems to be pretty well established, and the number of kittens of course will vary depending on conditions but it's not rare to be in those cases of female cats having multiple litters in a single year. Cat pregnancies last two months approximately, so it's not unreasonable at all.
gjulianm··on Clearing 500k feral cats from New York’s streets
> Who's to say cats can't live independently among humans (and rats and other animals) in cities?

I knew about this nonprofit via a cat colony caretaker, @poetssquarecats (I follow her on Tiktok and she's amazing) that talks about the struggles of feral cats and uncontrolled cat colonies. Yes, of course animals in the wild also suffer from diseases, parasites, and injuries. But these animals aren't in the wild. They're far more constrained than in the wild, with a lot more access to food, and with a lot more dangers. One of the problems is that, in those conditions, the number of cats can and does explode (female cats can become pregnant at 4-6 months old already, and they have on average 16 kittens a year in multiple litters).

And yes, of course getting rid of them is going to have unwanted consequences, but having them has other unwanted consequences. For example, as the article mentions, birdlife is severely affected, which can also have other unwanted consequences. I think that in that case it's better to have the cause under control: either way it's going to have unwanted consequences, but at least when you control it you know that it's happening.

gjulianm··on Passkeys now support external providers
AFAIK, passkeys aren't really linked to a domain. It's old-school public-key verification, the server stores your public key and uses it to verify the signature of a challenge they send to your device on login. As long as the different domains/apps can share the public key you should be able to login.

And for things like Plaid, I think banks are moving towards OAuth-style permissions, where you login to your bank and authorize the connection. Under the hood, Plaid or other app can connect to the bank with limited permissions using API authorization keys. It's a different problem, I think.

gjulianm··on Update on Sharing
This is a US rollout of a restriction that they have already deployed in Latin America and Spain among other countries [1], that they restrict your account to a single household. So, no password sharing with friends.

I am honestly surprised they keep pushing this. In Spain, Netflix is among the most expensive options, and the most expensive if you consider the features you get (the 7,99 euro option just gives you HD, not full HD, and only one concurrent device); and the content quality has been going downhill and has lost a lot of staple shows. Most people I know that had Netflix just canceled their subscription once the deadline hit, because the price for extra members is ridiculous (basically a full subscription cost) and without password sharing it's not worth it compared to the other options.

I guess that the move didn't have as much effect as people thought it did, and that's why they're continuing the rollout.

1: https://about.netflix.com/en/news/update-on-sharing-may-us

gjulianm··on Microtiming in Metallica's “Master of Puppets” (2014)
Not really the chorus, but mostly the instrumental part at the beginning (after the, let's say, acoustic-slow part). I think there's some short timings there too.
gjulianm··on Microtiming in Metallica's “Master of Puppets” (2014)
I think that another example of Metallica doing "whatever" with the timing is just in the previous song in the album, Battery. If you try to follow the beat, it mostly follows a 4/4 but sometimes they just cut it short. I feel it's more accentuated in Battery, because it's a little bit more random and contributes to the rushed feeling of the song.
gjulianm··on Add opt-in transparent telemetry to Go toolchain
Well, precisely the case for telemetry is to make improvements that aren’t obvious to the users. If they were obvious people would see them and send bug reports.

I have used telemetry for mobile apps in the past and now at work to monitor performance of certain software we deploy in client data centers, and there’s a lot of things I notice in the telemetry that users wouldn’t find and report. For example, I remember I was able to fix an issue in a mobile app because I noticed startup times increasing each time users opened the app, and I could fix the problematic cache quickly. I bet most users didn’t really notice. Same at my current job, we’re able to detect slowdowns and processing bottlenecks that for the users just show as subtly erroneous data. Do they notice those fixes? Nope, but that’s the point, I want to be able to fix things before they notice and report them.

gjulianm··on Add opt-in transparent telemetry to Go toolchain
You can manage to have things without telemetry, but having telemetry is incredibly useful. I think the example of "how much of our user base actually uses these features" is a very good one, specially in a compiler where maintaining old features could be adding a lot of complexity to the code base. And, as they also explain, a lot of bugs and undesired behaviors are things that the users won't know they have to report and just accept as part of the normal behavior. Things like cache misses, slowed down compilation times in certain situations, sporadic crashes... All of those things could be improved if the developers knew about them.
gjulianm··on Researchers craft a fully edible battery
My first question was "but why?" and apparently it has more applications than one would think at first, as it says in the article: edible batteries in toys for kids would eliminate toxicity risk and supporting electronics that go inside our body (they are working in edible transistors).

I also wanted to highlight this hilarious phrase in the article:

> While all of these components are nontoxic and edible, the researchers clarify that a user is not supposed to eat the battery on purpose, which would obviate the whole point of making it rechargeable.

"Please don't buy our batteries for eating, eat a banana or something"

gjulianm··on Why is inflation so sticky? It could be corporate profits
I don't think it's about the free market not being a thing. It's that, in a context where every company is increasing prices and customers have resigned and accepted that reality, the increased profit from extra clients is less than the increased profits from higher prices.
gjulianm··on Vacations in the Soviet Union
I don't think that people are "inexperienced" in conversations nor that the negative socialization comes that much from not sharing a common ground. A lot of negative socialization can come from people not having an interesting conversation, or being too self-centered, or repetitive, or convoluted, or not knowing when to just leave the other person alone.

I am also skeptical of the fact that "the working class is growing more isolated by the day". Looking around my environment, it seems that the advent of social networks (FB, Whatsapp, Instagram, etc) has made it far easier to keep in contact and to organize meetings with people. I know that I would be socializing far less if the only means of communication we had were the phone and letters.

gjulianm··on Vacations in the Soviet Union
Agreed. For some people, conversation takes energy. Being in a group takes energy. A vacation where you're out with complete strangers, in a group where you can't really do whatever you like and have to follow and compromise and discuss things, sounds like hell. I'd have to take a vacation from that.
gjulianm··on Bob Lee, former CTO of Square, has died after being stabbed in San Francisco
What is the ideology? I’m asking honestly because from the perspective of an European, what the US calls “left” would usually be center or even center right here in my country.
gjulianm··on APL: An Array Oriented Programming Language (2018)
> If you open up a math book and see integral signs everywhere, that immediately gives you hints for the kind of domain you're dealing with. If, on the other hand, you see mostly polynomial equations, then you're probably not worried about limiting processes. Mathematical symbols, at the graduate level and beyond, take on a communicative role in addition to the pure algebraic one you internalize during undergrad.

The title of the book also does that. Few books, papers, articles or exercises consist exclusively of symbols. I just took some random math papers from Arxiv and most of them are extremely wordy, and most of the symbols used are letters are names for values and functions defined right there, not operators that you need to reference elsewhere for the definition.

> What would you rather read? 30-ish equations or PyTorch library code? Which do you think would be easier to grok? I am, of course, referring to the 30-line self-contained APL implementation of a neural net with performance on par with PyTorch:

I mean, obviously the PyTorch library code. Not even talking about symbols and syntax, at least the PyTorch code has readable names. We could also mention the lack of spacing to match different concepts or groups of operations.

I find it really funny that the claim is "APL is like math" when precisely APL does things that are very much frowned upon in math: excessive use of symbols, long symbolic equations, lack of literal explanations...

gjulianm··on APL: An Array Oriented Programming Language (2018)
What rules? I am only saying that symbols add cognitive load, even more if they're context dependent. I don't think that's controversial at all, and APL and other array languages do have context-dependent symbols.

And yes, I tried learning an array language but I didn't get far for two reasons: one, I didn't really see where could I apply it; two, it's really hard to discover things and learn incrementally when you can't really use autocomplete or even Google (punctuation and symbols are not well supported in search engines) to help with the large set of symbols/operations with multiple meanings each.

gjulianm··on APL: An Array Oriented Programming Language (2018)
It's like a standard library except it's harder to search for the symbol you want via autocomplete, harder to infer the meaning of symbols you don't know (even when the C standard library has some terrible names) and easier to confuse (from the page, having symbols like ,. and .. or ,: and .:).

Of course, I assume that people end up knowing most of them and that not all of them are necessary. I'm not arguing that, I'm saying that symbols are an extra cognitive load and I find it funny that the argument for them is "well this is how is done in math" when precisely there's a push in math to avoid using excessive symbols.

← PreviousPage 5 of 34Next →