currymj

OTC decongestants that actually work, some useful info for those of us with bad sinuses.

- pseudoephedrine taken orally.

- phenylephrine, but only as nasal spray, not if you take it orally.

- Oxymetazoline (Afrin) nasal spray and others in this broad family

- propylhexedrine, sold OTC as Benzedrex as a vapor inhaler. Unfortunately people crack open the inhaler and swallow the whole thing as a drug of abuse, so often they are out of stock seemingly because of shoplifting, or not sold at all because the pharmacies don't want to deal with the hassle.

Anything that goes directly in your nose has the potential to cause rebound congestion after a couple days which can be pretty bad.

unfortunately everyone is now doing this with LLMs, it actually feels like it might be on its way to being a negative signal, as strange as that sounds.

do you know why residencies value the number of research items? Why would having a large number of garbage papers be seen as a positive signal at all?

Attendings and existing residents are consulted in the ranking process, they are picking people they will have to work pretty closely with for 4 years, they have skin in the game. Why does anyone put any weight on such a clearly bogus metric?

residencies have decided to outsource part of their hiring decisions to journal peer-review processes. so now for some submissions, editors and reviewers are not actually doing scientific peer review, but rather screening job candidates for hospitals.

peer review is built to assume good faith work by people who are all part of a community of scholarship, it can partially hold up to people within the community gaming metrics. if people are just going to appear, game the system to publish some papers, and then disappear into their real careers, there's no hope of this working.

i don't understand why residencies want med students to publish papers anyway. it's very difficult to do good scientific research, it requires training, time, and almost always apprenticeship. none of this is part of the medical school curriculum, which is why we need special MD-PhD programs for people who want to do both. nobody expects that doing a PhD in biology or epidemiology would give you any clinical know-how, why is it reasonable to expect the reverse?

the ideas have been worked out by people more sophisticated than me, good encyclopedia article on the topic:

https://plato.stanford.edu/entries/computation-physicalsyste...

this paper makes a lot of modest, carefully hedged, and reasonable claims.

in its tone however it's written as if it's a brutal takedown of... somebody's perspective. It's hard to tell whose or what perspective exactly. Maybe I'm just misreading the writing style.

(Personally, I think the general case here is one of the better objections to computationalism about consciousness. You can make it even more absurd.

There exists some isomorphism between the velocities of the molecules in a glass of water, and the states of a Turing machine simulating a human mind. So is the glass of water conscious? Actually there are many such isomorphisms to many possible conscious minds, so is every glass of water simultaneously having every possible conscious experience?)

for a truly profound and powerful buzzword, you can even honestly and accurately call what you've done "neurosymbolic AI"

I believe you, but if you are ever in a position of authority, please don't expect anyone else to function well on 4 hours of sleep.

it has been known to happen.

For example, spearheaded by Knuth, the community effectively abandoned the Journal of Algorithms and replaced with with ACM Transactions on Algorithms.

however it's difficult. a big factor is that professors feel obligated towards their students, who need to get jobs. even if the subfield can shift to everybody publishing in a new journal, non-specialists making hiring decisions may not update for a few years which hurts students in the job market.

On the whole you should rarely read papers, you want to read a whole literature in an area. Academics embedded in the field can do this easily. Academics outside of an area know to do this, and to bounce things off an expert to make sure you have the context and aren't over-indexing on a flashy result. Everybody learns the painful lesson in grad school to not just read a paper and believe everything will work as it says.

Somehow the general public and policymakers got the idea that if a paper gets published in any non-fake journal, this is an official endorsement that it's 100% correct, everything in it can be read in isolation, and it's safe to use all claims in the paper to direct policy immediately.

I think academia is partially to blame for encouraging people to believe this rather than insisting on explaining the nuances of how to interpret published research. On the other hand, nobody wants to hear a message that things are nuanced, and they will have to do costly hard work to get at the truth.

I think a world where "you can take any published paper at face value...without going direct to primary sources and bouncing it off an expert in the space" would be great, but it never existed, and it's just fundamentally impossible.

it varies enormously by field.

in CS you will have intense grind weeks around conference deadlines and a more manageable but challenging pace of life otherwise.

in wet lab science you live by the schedule set by your experiments, which often involves intense hours.

if you think of it like a bond it’s pretty fantastic. coupon rate 3.5% and you got it at a giant discount to par even though it’s actually (according to this guy’s beliefs which proved correct) nearly certain to be repaid.

sometimes you can see them do this and sometimes you can see they just work through the problem in the reasoning tokens without invoking python.

sympy is good enough for typical uses. the user interface is worse but that doesn't matter to Claude. I imagine if you have some really weird symbolic or numeric integrals, Mathematica may have some highly sophisticated algorithms where it would have an edge.

however, even this advantage is eaten away somewhat because the models themselves are decent at solving hard integrals.

there is some inevitable "insider trading" in commodities markets. for example if you're a giant agricultural company, and you want to hedge the price of soybeans, you have some extremely relevant insider information about the soybean market. but you're still allowed to trade soybean futures. very different than securities.

if prediction market contracts really are regulated as commodities, then presumably a lot of insider trading must be legal, although there must be limits of one kind or another and probably if you do something really egregious you might be prosecuted under some legal theory.

I have rented an apartment in Zürich (a hotel-room sized studio as you say, though with high quality construction and amenities). it was indeed pretty frustrating to go through the apartment search, but it is possible to rent housing, as evidenced by the fact that millions of Swiss citizens and residents live indoors.

Switzerland is like this and is also a real democracy. Although the food is not as good.

Anthropic already was using "Clawd" branding as the name for the little pixelated orange Claude Code mascot. So they probably have a trademark even on that spelling.

this would be a good development. seems very far off.

this is probably a net negative as there are many very good scientists with not very strong English skills.

the early years of LLMs (when they were good enough to correct grammar but not enough to generate entire slop papers) were an equalizer. we may end up here but it would be unfortunate.

the earlier list of ICLR papers had way more egregious examples. Those were taken from the list of submissions not accepted papers however.

the harsher the punishment, the more due process required.

i don't think there are any AI detection tools that are sufficiently reliable that I would feel comfortable expelling a student or ending someone's career based on their output.

for example, we can all see what's going on with these papers (and it appears to be even worse among ICLR submissions). but it is possible to make an honest mistake with your BibTeX. Or to use AI for grammar editing, which is widely accepted, and have it accidentally modify a data point or citation. There are many innocent mistakes which also count as plausible excuses.

in some cases further investigation maybe can reveal a smoking gun like fabricated data, which is academic misconduct whether done by hand or because an AI generated the LaTeX tables. punishments should be harsher for this than they are.

Especially for your first NeurIPS paper as a PhD student, getting one published is extremely lucrative.

Most big tech PhD intern job postings have NeurIPS/ICML/ICLR/etc. first author paper as a de facto requirement to be considered. It's like getting your SAG card.

If you get one of these internships, it effectively doubles or triples your salary that year right away. You will make more in that summer than your PhD stipend. Plus you can now apply in future summers and the jobs will be easier to get. And it sets your career on a good path.

A conservative estimate of the discounted cash value of a student's first NeurIPS paper would certainly be five figures. It's potentially much higher depending on how you think about it, considering potential path dependent impacts on future career opportunities.

We should not be surprised to see cheating. Nonetheless, it's really bad for science that these attempts get through. I also expect some people did make legitimate mistakes letting AI touch their .bib.

I recommend actually clicking through and reading some of these papers.

Most of those I spot checked do not give an impression of high quality. Not just AI writing assistance but many seem to have AI-generated "ideas", often plausible nonsense. the reviewers often catch the errors and sometimes even the fake citations.

can I prove malfeasance beyond a reasonable doubt? no. but I personally feel quite confident many of the papers I checked are primarily AI-generated.

I feel really bad for any authors who submitted legitimate work but made an innocent mistake in their .bib and ended up on the same list as the rest of this stuff.

the papers themselves are publicly available online too. Most of the ones I spot-checked give the extremely strong impression of AI generation.

not just some hallucinated citations, and not just the writing. in many cases the actual purported research "ideas" seem to be plausible nonsense.

To get a feel for it, you can take some of the topics they write about and ask your favorite LLM to generate a paper. Maybe even throw "Deep Research" mode at it. Perhaps tell it to put it in ICLR latex format. It will look a lot like these.

bigram-trigram language models (with some smoothing tricks to allow for out-of-training-set generalization) were state of the art for many years. Ch. 3 of Jurafsky's textbook (which is modern and goes all the way to LLMs, embeddings etc.) is good on this topic.

https://web.stanford.edu/~jurafsky/slp3/ed3book_aug25.pdf

I don't know the history but I would guess there have been times (like the 90s) when the best neural language models were worse than the best trigram language models.

Steam Deck is Arch-based, that's most likely why.

i would like to understand what people get, or think they get, out of putting a completely AI-generated survey paper on arXiv.

Even if AI writes the paper for you, it's still kind of a pain in the ass to go through the submission process, get the LaTeX to compile on their servers, etc., there is a small cost to you. Why do this?

if you run into anyone who is serious about the "fake Nobel" complaint, just ask them why they have such a high opinion of taking money from an arms dealer, and such a low opinion of taking money from a public institution in one of the Nordic welfare states.

for many types of scientific computing, there's a case to be made it is the best language available. often this type of computing would be in scientific/engineering organizations and not in most software companies. this is its best niche, an important one, but not visible to people with SWE jobs making most software.

it can be used for deep learning but you probably shouldn't, currently, except as a small piece of a large problem where you want Julia for other reasons (e.g. scientific machine learning). They do keep improving this and it will probably be great eventually.

i don't know what the experience is like using it for traditional data science tasks. the plotting libraries are actually pretty nicely designed and no longer have horrible compilation delays.

people who like type systems tend to dislike Julia's type system.

they still have the problem of important packages being maintained by PhD students who graduate and disappear.

as a language it promises a lot and mostly delivers, but those compromises where it can't deliver can be really frustrating. this also produces a social dynamic of disillusioned former true believers.