HNHacker News
TopNewBestAskShowJobs

currymj

3,822 karma · joined March 27, 2016

submissionscomments
currymj··on Lush: My favorite small programming language
i would compare base R to basically a shell. meant to be used interactively. okay for small scripts. you can write big programs but it will get weird.
currymj··on Recovering from a kidney donation
If you need to donate to your wife and aren't compatible, separate from the NKR vouchers, there are "kidney exchange" programs where pairs of people in that situation are matched with each other. So you'd donate to someone else, and at the same time someone would donate to your wife. Actually calculating the match among all those pairs is a fun NP-hard optimization problem.
currymj··on Something weird is happening with LLMs and chess
VW was breaking the law in a way that harmed society but arguably helped the individual driver of the VW car, who gets better performance yet still passes the emissions test.
currymj··on Genetic Discrimination Is Coming for Us All
An important fact (illustrated wonderfully in “Gattaca” which everyone rightly loves) is that there is not naturally any such thing as genetic privacy. If someone wants your DNA, you are constantly leaving it everywhere you go, and sequencing and checking SNPs just keeps getting cheaper and easier.

So genetic privacy, or at least nondiscrimination, is something we need to actively create with laws and social norms, if we want it.

currymj··on The online sports gambling experiment
the article proposes that placing bets should be done in-person at specific locations. no mobile apps or websites.
currymj··on The death and life of prediction markets at Google
thanks for the response! those calibration charts don't look too bad. that's reassuring.
currymj··on The death and life of prediction markets at Google
the market price gives you a number between 0 and 1, which should move higher as the event becomes more likely, so it's pretty useful.

however interpreting it as a probability, or an average of agent beliefs, or anything like that, seems tricky. i assume these internal markets are not deep and liquid enough that you can just throw up your hands and say "EMH". it works if you assume risk-neutral traders who will just trade up to their correct price but as I understand it, breaks down with realistic traders who may limited capital and are usually somewhat risk-averse.

i wonder how these prediction markets dealt with that. was there any postprocessing of market prices to get final probabilities? based on interviews with traders or observed trading behavior, did the traders behave in such a way that the market price could be interpreted as "pretty much" just a probability?

currymj··on The Death and Life of Prediction Markets at Google
the same dynamic is true of insurance, credit default swaps, and lots of other financial products. this causes trouble from time to time but can be managed if the products serve a useful purpose.
currymj··on FDA proposes ending use of oral phenylephrine as OTC nasal decongestant
phenylephrine is still pretty effective as a topical nasal spray. so don't write it off if you see it in that form.
currymj··on Model Predictive Control in the Browser with WebAssembly
there's a lot of work in the broad area. most of it doesn't engage with the classical control theory literature (arguably it should).

some keywords to search for recent hot research would be "world model", "decision transformer", "active inference", "control as inference", "model-based RL".

currymj··on Trump wins presidency for second time
I was in Europe for a while recently (Switzerland). Thought about staying but when I realized that even if I naturalized, I could never really be Swiss, and furthermore future children would not really be Swiss (even if they too naturalized), and at best perhaps my grandchildren could be considered somewhat provisionally Swiss... not appealing. Too much old, too much history, if you aren't embedded in it you are permanent outsider.
currymj··on Grindr Illegally Used RTO to Thwart Union, Forced Out 1/2 of Staff, NLRB Alleges
there's not much that will generalize here. the NLRB has a lot of procedural powers around union campaigns and elections. there are a lot of things that, if a company does them in response to a union campaign, are considered to violate the employees' rights -- not their rights to good working conditions, just their right to decide whether or not to form a union in a fair environment.

usually this is doing things like making hours worse, enforcing rules more strictly, etc. in response to a union campaign. but they can also get in trouble for positive things like relaxing rules, giving raises, etc. and in a few funny cases they can even get in trouble for supporting the union.

in other words, if Grindr had, in response to unionization efforts, decided to placate employees by expanding WFH, they could have gotten in trouble in the exact same way.

currymj··on Is the Q source the origin of the Gospels?
Another point: while there are some unique religious aspects to your experience, it also sounds pretty similar to complaints about secular universities.

Very common to hear about students who don't care about the material at all, they just want a job. So they are bored and frustrated by anything, especially theoretical material, that isn't going to be obviously useful for interview prep or their future careers.

currymj··on Is the Q source the origin of the Gospels?
there are other Christian communities than American-style Evangelical Christianity who tend to be more intellectual and do like arguing about this kind of thing.
currymj··on A Serious Man: On Bruno Latour
I think "undeformed travel" means that when people quote it or refer to it, they tend not to change his original words.
currymj··on Our First Generalist Policy
interestingly chess has made a huge comeback recently even though this was a widely-held attitude for a while after Deep Blue.
currymj··on ChatGPT Search
Generative Pretrained Transformer is also a terrible brand name but it doesn’t seem to matter.
currymj··on ChatGPT Search
A good language model is one with low perplexity.

https://en.wikipedia.org/wiki/Perplexity

Reasonable name for a language model startup.

currymj··on Why are close elections so common?
i would like a better voting system in the US.

but e.g. Australia has a great voting system (on all levels from the voting rules to the election day administration), yet they still seem to have fairly screwed up partisan politics.

i think it might improve things but wouldn't help make politics fundamentally better.

currymj··on Dropbox announces 20% global workforce reduction
probably whoever makes the loan will hedge, maybe by shorting the stock. or buying puts, but then whoever sells them the puts will hedge by shorting the stock.

if they decide not to hedge, then it must be because they don't think they're exposed to much risk, which basically means they like the stock and would be willing to own it. it feels like it mostly works out.

currymj··on A return to hand-written notes by learning to read and write
apple notes on new iPadOS does this right now -- just cleans up your handwriting to be slightly neater but still look like you.
currymj··on Artemis moon suit designed by Axiom Space and Prada revealed in Milan
I think that the SpaceX EVA suits are just a little too sleek and modern yet bland. Like default Unreal Engine spacesuit assets.

Pleasantly surprised to see these spacesuits look classic yet beautiful. I guess I shouldn't be surprised that Italian fashion designers did a good job.

I have no view on the technical merits of the suits or the program.

currymj··on Germany's 49-euro ticket resulted in significant shift from road to rail
that is true, but tons of highly frequent on-time commuter trains doesn't help anyone who wants to travel long distances.

right now if you want to go between two cities you always have to plan on missing your connection.

currymj··on Research in psychology: are we learning anything?
i think we kind of agree with each other. doctors based on experience found things that worked better at treating patients. and they also had completely wrong theoretical models to rationalize these things.

i don't really see monotonic progress: e.g. going from Avicenna to Paracelsus seems like a lateral move at best as far as scientific knowledge (and both theoretical systems stuck around competing for hundreds more years), even though Paracelsus and his followers did make some important contributions to medicine in practice and medicine improved over that period.

so you had practical progress for a long time but scientific knowledge, meaning theories of nature, stayed equally bad.

currymj··on Germany's 49-euro ticket resulted in significant shift from road to rail
among intercity trains run by Deutsche Bahn, the on-time rate hovers around 2/3. it's technically true that "most were on time", but one third being late is really terrible, especially since you often need to make a tight connection.

even if your family has no justifiable basis for believing what they do about DB, they happen to be correct by luck.

https://ibir.deutschebahn.com/2023/en/combined-management-re...

currymj··on Research in psychology: are we learning anything?
i think that astronomy/physics/gravitation is actually a pretty weird case in the history of science and most things don’t go that cleanly.

better examples might be medicine, where people just bounced around from one insane wrong theoretical system to another (humors to “empiricism” to bad air to Paracelsus to Avicenna and round and round and round). but somehow progress happened anyway. actual steady scientific progress only took off in the 19th century.

or chemistry, where alchemical theories were also completely bizarre, mostly mysticism and poetry. but despite being “not even wrong”, people following them became pretty good at laboratory skills

After many centuries, this laid the ground for empirical chemistry, and after a few more centuries, a theoretical system emerged that is close to right. But there was a lot of progress even under the “not even wrong” theories.

currymj··on Research in psychology: are we learning anything?
partially disagree with this, every proto-science historically had a bunch of wrong but highly sophisticated theories. medicine, alchemy (as mentioned in the article), physics, biology (Aristotle), astronomy. for some reason it seems you need the wrong theories to organize the empirical data.

I actually think Freud’s elaborate mental structures have some of this feeling to them.

currymj··on Megalopolis is baffling and plainly nuts but worth it
i can understand walking out due to disgust or confusion but “bored” is hard for me to comprehend.
currymj··on Megalopolis is baffling and plainly nuts but worth it
in my theater the audience was laughing uproariously at a lot of it (some but not all intended to be funny). then everyone applauded at the end.
currymj··on Legalizing sports gambling was a mistake
i think the number of random gamblers who get so consistently lucky that their bet size gets reduced, is probably quite small. this is because you usually lose money betting on sports, because sports betting is bad. it's mainly going to be people doing obvious arbitrage, and secondarily people who truly are professional gamblers.

this can also be spun in a positive way: if that does ever happen, the bookies are literally forcing someone to quit when they are ahead! isn't that considerate of them.

unfortunately, i think sports betting platforms just have many strong arguments that controlling bet sizes in this way is fine.

← PreviousPage 6 of 34Next →