Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Fraud, so much fraud

science.org

Post reply on HN
Print view View original

441–450 of 868 posts

Re: Fraud, so much fraud

#441

Post by eichin »

jknoepfler wrote: post #422
wredue wrote: post #322

Earlier quoted context omitted.

Hacker news is completely flooded with “AI learns just like humans do” and “AI models the human brain” despite neither of these things having any concrete evidence at all. Unfortunately it isn’t just bosses being fooled by this. Scores of people push this crap. I am not saying AI has no value. I am saying that these idiots are idiots.

The reality on the ground (for me) has been refreshingly sane. I work at a company with a substantial BI/ML footprint. Our head of research was tasked with evaluating the applicability of LLMs to either our product or our daily workflows. To date the consensus is that there isn't much there for our product, that integrating LLMs into our models would introduce more problems than it would solve, and that we should cau…

Refreshing but rare; usually this kind of eval gets done by someone excited to do it because they've already been "intellectually captured" by the hype.

Re: Fraud, so much fraud

#442

Post by echoangle »

benlivengood wrote: post #419
echoangle wrote: post #392

Earlier quoted context omitted.

What evidence for „AI models the human brain“ do you want? Isn’t a neural network pretty clearly a simplified model of the working of the human brain? What is there to prove?

One clear piece of evidence would be ruling out "AI models the corvid brain" or "AI models the cephalopod brain" which might narrow it toward the human brain. That it's functionally impossible to do either leads me to believe that "it models some form of intelligence" is about the best we can prove.

I don’t understand the standard of modeling you seem to assume.

Modeling a human brain, a cephalopod brain and a corvid brain aren’t even mutually exclusive if your model is abstract enough.

When I say „a neural network models a human brain“, I’m talking about the high level concept, not the specific structure of a human brain compared to other brains. You could also say that it models a dogs brain if you wanted to. It’s just the general working principle that is kind of similar. Does that not count as a model to you?

Edit: Here’s a simple example: I would say that a simple SIR model „models COVID infection in humans“. But that doesn’t mean it can’t also model Pig Flu in Pigs. It’s a very abstract model so it applies to a lot of situations, just like a neural network basically models the brain of every reasonably advanced animal.

Re: Fraud, so much fraud

#443

Post by garrettgarcia »

fluoridation wrote: post #428
klenwell wrote: post #341

Earlier quoted context omitted.

I think my favorite Simpsons gag is the episode where Lisa enlists a scientist (voiced by Stephen Jay Gould) to run tests to debunk some angel bones that were found at a construction site. In the middle of the episode, the scientist bicycles up to report, dramatically, that the tests "were inconclusive". In the end, it's revealed that the bones were a fraud concocted by some mall developers to promote their new mall.…

Well, the incentive is that he didn't want to run the tests out of laziness (i.e. he lacked an incentive to run them). He ran to Lisa to give his anticlimactic report not to be deceptive, but rather he just happened to be cycling through that part of town and just needed to use the bathroom really badly.

I thought his incentive was to defend the idea of miracles/faith/angels/God.

Re: Fraud, so much fraud

#444

Post by jltsiren »

hansonkd wrote: post #416
jltsiren wrote: post #373

Earlier quoted context omitted.

> The research might be novel, but the procedures for research and publication are very similar. This is wrong. Some time ago, I completed the checklists for publishing a paper in a somewhat prestigious multidisciplinary journal. Large parts of the lists were about complying with various best practices and formal requirements in different fields. I often didn't even understand the questions outside my field. And the…

I think you are desperately trying to fit something black and white rather than thinking critically that there is a spectrum of research, some of which is similar to others which can easily have procedures for insuring and others that are more complex that require more diligence from the insurance company. Just like nearly every single thing an insurance company does. Yes there is novel research that has never been d…

The cases that would require more diligence from the insurance company are the kind of research that should be encouraged. Breakthroughs are more likely to happen when people take risks and try something fundamentally new, instead of adhering to the established forms. Your insurance model would discourage such research by making it more expensive.

Additionally, even if we assume that the insurance model is a good idea, it should be tied to individual researchers, not universities. The entire model of university research is based on loose networks of independent professionals nominally employed by various organizations. Universities don't do research, they don't own or control the projects, and they don't have the expertise to evaluate research. They are just teaching / administrative organizations that provide services in exchange for grant overheads.

Re: Fraud, so much fraud

#445

Post by ricksunny »

Der_Einzige wrote: post #354
tensor wrote: post #148

Earlier quoted context omitted.

I'd say the public needs to develop some rational impulse, it already has plenty of skepitism to the point where people no longer trust science the methodology. Instead, they genuinely believe there is some alternative to finding the truth, and now simply believe the same old superstitions and bunk that people have prior to the scientific revolution. Speaking of Orwell, I don't think science comes into it. Rather, wh…

Critical thinking = the ability to be skeptical, literally it is the ability to criticize. Great critical thinkers become lawyers, post modernist intellectuals, and other parts of the "talking" class of intellectuals. Unfortunately, it's far easier to talk shit than it is to build things. We've massively over-valued critical thinking over constructive thinking. Most people want to dunk on science. Few people want to…

>Many people act like submitting papers is impossible for non-Ph.D's.

I agree. But academia reinforces this perception. I feel like PhD's only give serious consideration to the utterances of other PhD's. The rest of the public consists of the unwashed masses, and at best gets the smiling-nod treatment from teh PhD.

PS I (a non-PhD) managed to publish a paper during the pandemic (doi: 10.3389/fphar.2022.909945 ). One of the biggest barriers was the item you mentioned quoted above, and the bogeyman of "epistemic trespass" in general, as operating in my own psychology. I've since become noisy in advocating for the #DeSci movement.

Re: Fraud, so much fraud

#446

Post by mfld »

dekhn wrote: post #62

These sorts of articles raise so many thoughts and emotions in me. I was trained as a computational biologist with a little lab work and ran gels from time to time. Personally, I hated gels- they're finicky, messy, ugly, and don't really tell you very much. But molecular biology as a field runs on gels- it's the priimary source of results for almost everything in molbio. I have seen more talks and papers that rested…

Re: the role of (gel) images as the key aspect of a publication. To me this is very understandable, as they convey the information in the most succinct way and also constitute the main data & evidence. Faking this is so bold that it seemed unlikely.

The good news IMO: more recent MolBio methods produce data that can be checked more rigorously than a gel image. A recent example where the evidence in form of DNA sequencing data is contested: https://doi.org/10.1128/mbio.01607-23

Re: Fraud, so much fraud

#447

Post by benrapscallion »

DevX101 wrote: post #207
clpm4j wrote: post #166

I'm not a researcher or academic, but when I think of roughly how long it takes me to do meaningful deep work and produce a project of any significance, I'm struck by the fact that his 800 papers isn't a red flag? Even if you allocate ~3 months per paper, that's over 200 years of work. Is it common for academics to produce research papers in a matter of days? From the article: Masliah appeared an ideal selection. The…

It's kind of like, when reporters say a CEO built [insert ridiculously complex product here], ex: ascribing the success of OpenAI to Sam Altman, or Apple to Steve Jobs. Sure there were important in setting the direction, and allocating resources but they didn't actually do the work. Similarly, the heads of famous science labs have lots of talented scientists who want to work with them. The involvement of a lab direct…

800 papers is still a red flag, even if throughout those years he had 30 postdocs at any given time (which is only true for large labs).

Re: Fraud, so much fraud

#448

Post by benrapscallion »

stephenbez wrote: post #189
clpm4j wrote: post #166

I'm not a researcher or academic, but when I think of roughly how long it takes me to do meaningful deep work and produce a project of any significance, I'm struck by the fact that his 800 papers isn't a red flag? Even if you allocate ~3 months per paper, that's over 200 years of work. Is it common for academics to produce research papers in a matter of days? From the article: Masliah appeared an ideal selection. The…

People are listed as authors if they advised or contributed to the papers of their grad students or other people in their lab.

They still need to read the science and agree with it.

Re: Fraud, so much fraud

#449

Post by vladms »

nonrandomstring wrote: post #312
inglor_cz wrote: post #277

Earlier quoted context omitted.

Was it ever? Ancient texts are full of complaints that people have become immoral, unlike in the Golden Age, which was conveniently long ago.

Of course. We've always had criminals. And as you say, throughout history people have complained about it. All I'm saying is that today people positively celebrate it, and - according to the parent poster - we now demand criminality as a necessity for "success".

Just looking at capitals of many countries, lots of monuments (statues, etc.) are dedicated to conquerors that were basically celebrated for successfully killing, subjugating and stealling from the others.

There are still people that have success without anything criminal.

Re: Fraud, so much fraud

#450

Post by fluoridation »

mistermann wrote: post #245
yunwal wrote: post #143

Earlier quoted context omitted.

How does this work for things like COVID vaccines, where waiting for a reproduction study would leave hundreds of thousands dead? Ultimately there needs to be some level of trust in scientific institutions as well. I do think placing higher value on reproducibility studies might help the issue somewhat, but I think there also needs to be a larger culture shift of accountability and a higher purpose than profit.

I believe if we taught philosophy in school to a non-trivial level we wouldn't have to rely on trust/faith. I wonder if it's possible to get people to wonder why the one discipline that has the tools to deal with all of these epistemic, logical, etc issues isn't taught in school. You'd think it would be something that people would naturally wonder about, but maybe our fundamentalist (and false) focus on science as th…

Suppose you need to make a decision on a topic that's contingent on P being true, which someone has already tested. How would you go about making the decision without testing P yourself (because that would mean that you would have to do the same for every decision in your life)?
Post reply on HN