HNHacker News
TopNewBestAskShowJobs

stanford_labrat

395 karma · joined November 10, 2020

submissionscomments
stanford_labrat··on Electron band structure in germanium, my ass
> Right on up to professorships, this is how science really works.

Why I am making my exit from academia and research entirely as soon as I finish my PhD. The system is filled with wonderful, intelligent people but sadly simultaneously rotten to the core. It in fact, did not get better as I moved from undergrad to grad school.

stanford_labrat··on Accelerating scientific breakthroughs with an AI co-scientist
So I'm a biomedical scientist (in training I suppose...I'm in my 3rd year of a Genetics PhD) and I have seen this trend a couple of times now where AI developers tout that AI will accelerate biomedical discovery through a very specific argument that AI will be smarter and generate better hypotheses than humans.

For example in this Google essay they make the claim that CRISPR was a transdisciplinary endeavor, "which combined expertise ranging from microbiology to genetics to molecular biology" and this is the basis of their argument that an AI co-scientist will be better able to integrate multiple fields at once to generate novel and better hypothesis. For one, what they fail to understand as computer scientists (I suspect due to not being intimately familiar with biomedical research) is that microbio/genetics/mol bio are closer linked than you may expect as a lay person. There is no large leap between microbiology and genetics that would slow down someone like Doudna or even myself - I use techniques from multiple domains in my daily work. These all fall under the general broad domain of what I'll call "cellular/micro biology". As another example, Dario Amodei from Claude also wrote something similar in his essay Machines of Loving Grace that the limiting factor in biomedical is a lack of "talented, creative researchers" for which AI could fill the gap[1].

The problem with both of these ideas is that they misunderstand the rate-limiting factor in biomedical research. Which to them is a lack of good ideas. And this is very much not the case. Biologists have tons of good ideas. The rate limiting step is testing all these good ideas with sufficient rigor to either continue exploring that particular hypothesis or whether to abandon the project for something else. From my own work, the hypothesis driving my thesis I came up with over the course of a month or two. The actual amount of work prescribed by my thesis committee to fully explore whether or not it was correct? 3 years or so worth of work. Good ideas are cheap in this field.

Overall I think these views stem from field specific nuances that don't necessarily translate. I'm not a computer scientist, but I imagine that in computer science the rate limiting factor is not actually testing out hypothesis but generating good ones. It's not like the code you write will take multiple months to run before you get an answer to your question (maybe it will? I'm not educated enough about this to make a hard claim. In biology, it is very common for one experiment to take multiple months before you know the answer to your question or even if the experiment failed and you have to do it again). But happy to hear from a CS PhD or researcher about this.

All this being said I am a big fan of AI. I try and use ChatGPT all the time, I ask it research questions, ask it to search the literature and summarize findings, etc. I even used it literally yesterday to make a deep dive into a somewhat unfamiliar branch of developmental biology more easy (and I was very satisfied with the result). But for scientific design, hypothesis generation? At the moment, useless. AI and other LLMs at this point are a very powerful version of google and code writer. And it's not even correct 30% of the time to boot so you have to be extremely careful when using it. I do think that wasting less time exploring hypotheses that are incorrect or bad is a good thing. But the problem here is that we can pretty easily identify good and bad hypotheses already. We don't need AI for that, what takes time is the actual amount of testing of these hypotheses that slows down research. Oh and politics, which I doubt AI can magic away for us.

[1] https://darioamodei.com/machines-of-loving-grace#1-biology-a...

stanford_labrat··on DA, sheriff, who shared woman's nude photos on phone are covered by QI
yes by text they might have missed nothing...but what's missing here is that Chapelle is doing his "white-man voice" which is kind of a big detail.
stanford_labrat··on What really happens inside a dating app
And at some places like Hinge, they'll identify your type so well that your "standouts" page (the one you have to pay to message people on) will be composed of 100% people you would probably like to speak to while your free feed will just be composed of people you have no interest in talking to!
stanford_labrat··on What really happens inside a dating app
My guess is that time would be the first little work-break from the post-lunch session. About an hour after lunch and you take a little break pause so you swipe on Hinge to see if you got any matches.
stanford_labrat··on Why is homeschooling becoming fashionable?
My family has lived in 4 states and 3 countries and the only time we ever homeschooled…was in NJ back in the early 2000s.
stanford_labrat··on ChatGPT Pro
sure, here's a link of a conversation from today 12/9/24 which has multiple incorrect: references, links, papers, journal titles, DOIs, and authors.

https://chatgpt.com/share/6757804f-3a6c-800b-b48c-ffbf144d73...

as just another example, chatgpt said in the Okita paper that they switched media on day 3, when if you read the paper they switched the media on day 8. so not only did it fail to generate the correct reference, it also failed to accurately interpret the contents of a specific paper.

stanford_labrat··on ChatGPT Pro
I don’t recall that it ever mentioned if it did or not. I don’t have the search on hand but from my browser history I did the prompt engineering on 11/18 (which perhaps there is a new model since then?).

I actually repeated the prompt just now and it actually gave me the correct, opposite response. For those curious, I asked ChatGPT what turned on a gene, and it said Protein X turns on Gene Y as per -fake citation-. Asking today if Protein X turns on Gene Y ChatGPT said there is no evidence, and showed 2 real citations of factors that may turn on Gene Y.

Pretty impressed!

stanford_labrat··on ChatGPT Pro
I recently (in the last month) asked ChatGPT to cite its sources for some scientific data. It gave me completely made up, entirely fabricated citations for academic papers that did not exist.
stanford_labrat··on California's most neglected group of students: the gifted ones
When I was going through school the gifted program allowed for kids to skip 1 grade in math and 1 grade in science. I think this was reasonable and didn’t lead to much bullying. Also helped that we had a large gifted program. A math class might’ve been 20-30% gifted kids at any given time.
stanford_labrat··on Oxford accused of relying on young academics employed on gig-economy terms
this is all my opinion as a 3rd year phd candidate who is thoroughly disillusioned. in STEM science there are two types of labor: thinking of experiments to do, and actually doing experiments. doing experiments is high-skill, manual labor. you often work with dangerous materials for crazy hours (some experiments take 10 hours to just do and you have no choice but to work from start to finish) and do tasks that are physically and technically challenging. as such, the fair market wage for this labor is actually pretty high. when i was an employee at Stanford, I was making nearly 80k in salary alone and i had fantastic health, retirement, and vacation benefits. oh, and i had overtime as a union employee to boot.

higher institutions realized that they could conveniently circumvent the rightfully high cost of labor to do experiments by hiring "graduate students" who they can pay poverty wages to do the manual scientific labor in exchange for a degree. as a student i am classified as a 20 hour per week worker, 20 hour per week student, and paid 37k a year. to be honest i wouldn't mind this if i actually were allowed to work 20 hours per week but this is not the case. the expectation, is that you will work 50-60 MINIMUM. and don't even get me started on postdocs. the PhD market has become so oversaturated due to the over-training of PhDs that to be competitive you must do an additional 5-8 years of "training" as a postdoc (again at poverty wages around 45-50k) before you get the opportunity to APPLY faculty positions.

this has had some predictable but unfortunate consequences. for one, institutions no longer select for the best and brightest. their number 1 criteria for offers of admission is "ability to finish the phd" aka willingness to be subject to these labor conditions for 4-7 years. of course you do need a minimum level of intellectual ability to get there, but trust me when i say this, the bar is low. and academia has suffered because of it imo. prestigious institutions get by as you mentioned by having their pick of the smartest AND hard working.

stanford_labrat··on Scientific American's departing editor and the politicization of science
the irony for me at least, is that even at hyper liberal institutions all my colleagues (students, post doc, faculty even) entertained the idea that a lab leak was possible. just when it came to the media this hypothesis was labeled as a conspiracy theory.
stanford_labrat··on Lessons from my first exit
my very first exposure to this was paying for parking. i worked at a large private university which had a beautiful but very big campus. ofc for some reason the institution decided that employee's had to pay for their own parking, which i disagreed with on moral grounds and the fact that parking was so far away it was like a 20 minute walk just to get to my work space...so i did an experiment where i basically never paid for parking in my entire 4 years there.

it would have cost me about 2 grand to diligently pay for parking. the actual expense from tickets? about 100 bucks maybe 200. and some of them got automatically thrown out just by challenging them. now towards the end of my time there i did see a guy get towed as he was walking up to his car. i started paying for parking then.

stanford_labrat··on Did scientists revive an extinct animal or just breed a less stripey zebra?
you'd be absolutely right that the global warming justification is a retro-active "find a problem" situation. the true motivator lies in the technology that Colossal is attempting to accomplish the mammoth cloning with.

as it stands right now, cloning of any organism requires at the very least living fibroblast cells. Of course, you cannot get these from deceased mammals. Colossal's strategy is to essentially blast the host genome with fragments of the subjects DNA (which you can get from deceased/fossilized tissue) and hope that the end result is an elephant cell that's been reprogrammed into a mammoth-like cell from which to do cloning.

imo if they really wanted a mammoth, they could just make a hairy elephant. that would be 10x easier. but the whole idea is being able to resurrect...mammals.

stanford_labrat··on Do AI detectors work? Students face false cheating accusations
Seeing my fellow grad students cheat, then brag in the public student lounge to multiple people about having the highest score on the exam by “oh I didn’t even understand that question I just copied the answer key and I still got the highest score”, broke something in me.

Our institutions are failing us, and I have never been more disillusioned.

stanford_labrat··on Do AI detectors work? Students face false cheating accusations
Ironically enough, our lecture halls were simply not big enough. The space capped out at around 300-600 people and for popular topics such as programming 101 every semester would easily have 1500+ enrolled.
stanford_labrat··on Chemistry Nobel: Computational protein design and protein structure prediction
username...checks out
stanford_labrat··on GLP-1 pills are coming, and they could revolutionize weight-loss treatment
AFAIK Fractyl's "drug" is not gene therapy but a surgery where they remove part of the lining of your colon. It is not currently possible (nor will it likely be) to use gene therapy for obesity treatments. The technology is not suitable for this kind of medical application.

slide 5: https://www.fractyl.com/wp-content/uploads/2024/10/DGVS-2024...

stanford_labrat··on GLP-1 pills are coming, and they could revolutionize weight-loss treatment
For those not aware: Metformin is a popular (unapproved?) nootropic.
stanford_labrat··on California bans legacy admissions at private universities
apologize for the tangent but my cousin and her husband (both asian american, both harvard alumni) chose to move back to our hometown in Seattle so they could send their kids to lakeside, the same school bill gates went to.

i assume to increase their chances of getting into harvard. back of the envelope math works out to be $715k for their 5-12 education which may be more bang for your buck but also the college entrance strategy is not so much based on donor contribution and more on target school selection.

stanford_labrat··on California bans legacy admissions at private universities
break us colleges into tiers based on academic rigor. assuming you met XXXX standardized test score you are automatically sorted into one of the appropriate schools at random on your 18th birthday.

no take backs, safety schools permitted.

stanford_labrat··on AlphaProteo generates novel proteins for biology and health research
> We'll be able to print and grow organs in mass.

I'm a PhD candidate doing my thesis work on stem cell models and tissue engineering for organ transplant...I think this technology is certainly a large leap forward but I think you are a little overzealous with this claim.

stanford_labrat··on Cell-Free Bioactive Scaffold Supports Cartilage Regeneration-Large Animal Joints
i'm a 20 something former semi-pro athlete with grade 4 osteoarthritis in my knee. pain was massively bad but over the past 2 years has improved significantly. here's what i did: 1) mindfully progressively overload the leg muscles to build strength 2) quit playing rugby (so stop playing hard contacts sports with a lot of abrupt, sudden movements) 3) do my best to lose weight.

our situations are different but i wish you luck.

stanford_labrat··on The workers have spoken: They're staying home
the number of bad bathroom hygiene at my current workplace makes me very glad to be living right across the street where i can go use the facilities whenever i please in the peace and quiet of my own bathroom.

underrated perk for sure!

stanford_labrat··on Taking away iPhone made daughter a better person
yeah i agree, if it wasn't clear what i meant to say was that i functionally didn't have a smartphone until high school was over, and as a result i think i was socially excluded from a lot of things - the social groups from hs seem to have survived college (which may be a quirk of my specific hs) whereas i only talk to two people who i was deeply close with.

and i think it's even worse now.

stanford_labrat··on Taking away iPhone made daughter a better person
n=1 but i didn't have a smartphone until halfway through senior year of hs (2013-2014) whereas most of my peers had one from 2010 onwards. i wouldn't say socially stunted but popularity wise i would say not a full member of the social groups i was tangentially in, not in any group chats, etc.

deep 1-1 connections for the most part because yeah i was texting and messaging people but other than that...i think it's probably only worse now. there are group chats/discords for each social group probably and not having a smartphone would mean you can't participate.

stanford_labrat··on Armor from Mycenaean Greece turns out to have been effective
imagine being a medieval peasant conscript and seeing that guy from the first few seconds of the first video charging at you, knowing full well he's from the warrior class and trained his entire life for the job of fighting.

terrifying...

stanford_labrat··on 300k airplanes in five years
Fighting static positions is what you do when your opponent is literally your neighbor occupying slightly different dirt that touches yours. Which is not the case with China and the US.

When you switch from land to water tooling your military doctrine to things like air superiority, missiles, island hoping make much more sense to me at least.

stanford_labrat··on Jim Simons has died
even small funds, this was the investment thesis at the hedgefund that my dad worked for circa 2006-2008. they promised super dampened volatility but, as you might guess, they went belly up during the great recession.

The cause? Someone, somewhere in their financial product chain was not being faithful about the volatility of the asset they were basing their whole model on so when the market went tits up they did to.

stanford_labrat··on Senate passes reauthorization of key US surveillance program after midnight
Because in reality the two party system is not accurate. It’s rich versus poor, those with power versus those without. Nobility versus peasants. That’s just how it works.
← PreviousPage 2 of 5Next →