Sabitlenmiş Tweet
Marcin J. Suskiewicz
1.3K posts
Marcin J. Suskiewicz
@MSuskiewicz
Structural biologist and biochemist. CNRS researcher at CBM Orléans. Interested in protein modifications & interactions. Also husband, dad of 3, friend, ☧.
Orléans, France Katılım Ağustos 2020
4.2K Takip Edilen1.2K Takipçiler
@ramez Well, it is presented in a rather confusing way, with the '100' bar striped, and the scale going exactly from 0 to 100.
English
@scottnarmstrong Apologies, it wasn't a direct reference to anything you wrote, but a reminder that there're more thn these two attitudes, and among the various positions, the artisanal one might not be unreasonable in the long term. At least it resists what it considers undesirable and dystopian
English
Daniel Litt contra Hugo Duminil-Copin and other artisanal mathematicians:
"If you believe, as I do, that producing high quality science should be one of our fundamental goals, I think you’re obligated to embrace new tools that help one do so. Refusing to is a declaration that these outputs are not important."
Daniel Litt@littmath
STEM academia serves two closely intertwined purposes: the production of high quality science and the production of human capital. These two purposes feed into each other. The obvious direction is that we develop human capital by paying people to produce science. What is perhaps less obvious is that the very fact that human labor is used to produce science has historically been an important input to its quality. The goal of science is not simply to produce papers, but rather to produce good work--that a person is willing to spend months working on a paper is a (weak) witness to the fact that it has some minimum quality. If someone has a record of producing high quality work, that they wrote a paper is a stronger witness, since it was worth the opportunity cost to write it. If many people engage with it substantially, that is even stronger evidence. This is not to say that there isn't lots of low-quality work--there is, in fact a huge amount--but we have strong sorting mechanisms, admittedly using imperfect proxies (all depending on costly human labor!), to find high-quality stuff. Arguably the paper itself is not the primary product here; in many cases the primary product is actually the expertise developed over the course of producing it, which can then be applied to other questions. If you believe, as I do, that producing high quality science should be one of our fundamental goals, I think you’re obligated to embrace new tools that help one do so. Refusing to is a declaration that these outputs are not important. But I worry that we are not on track to automate the production of good work; rather, we are on track to automate the production of papers. We need new mechanisms to ensure that we are also producing good work, and to ensure that we are developing the human capital to engage with it.
English
Marcin J. Suskiewicz retweetledi
My prediction from last summer was that the number of frontier AI models getting a gold medal at this summer’s IMO will be… zero! The reason is that they won’t bother to compete, it’ll simply be beneath them. If anyone can now push a button on Codex / Claude Code and get a perfect score, what’s the point? No, they’ll just leave the 17 year olds to take the test on their own. (The open source models will still compete for another year or so. That’s my guess!)
Similarly, I think the labs pushing “research math” is also a fad that will expire soon enough. Think about it. GPT solved a major problem (Erdos unit distance); what they’re not reporting is the 1000 other problems they attacked and failed to make progress. [That’s not exactly deception; I also don’t report the dozens of things I tried to prove and failed…] They’re also not reporting the millions of dollars all of this cost them, and for what? Right now the “for what” is advertising: they’re signaling that they’re the best model for math, so you should use them for whatever your reasoning task is. Math departments also spend millions of dollars and produce theorems, but that is their actual end goal. A tech company is happy with a million-dollar theorem only if it predicts a billion-dollar application somewhere else. Once the bubble bursts, investors will want “real” applications from AI, new drugs, self driving / flying cars, etc etc. Nobody will care that the systems are also useful at proving theorems. Nobody but us mathematicians. So like the IMO, I think the frontier labs will get bored of theorems, and will leave us humans alone to keep doing math (and they’ll give us an amazing tool with which to do it!).
Does that make sense? What do you think?
English
@OlegUrminsky Not sure I agree about this, at least not as an exclusive answer. I think society pays our salaries also so that there are human communities within the society that nurture understanding & dialogue centred on a specific topic. And in this case, using AI might be counterproductive
English
To put it another way, we the researchers are the means not the ends. Society pays our salaries to figure stuff out just like carpenters are paid to build houses. When we can do the job better using AI, we should; when AI does not result in a better job, we shouldn't use it.
English
(unpopular?) opinion: If AI can accurately do some of the research we do "by hand" now, we should do that research much more quickly & at larger scale using AI, get those questions answered efficiently, and move on to the kinds of research questions that are still unanswered.
Brendan Nyhan (@BrendanNyhan on 🟦☁️)@BrendanNyhan
I fear research presentations turning into AI witch-hunts. The ultimate question for science is whether the claim is true, not how the answer was produced.
English
@wyrdweir @benglickenhaus human-like consciousness, it got hacked.
English
@wyrdweir @benglickenhaus for consciousness. And proxies only work as indicators if one does not deliberately uncouple them from the underlying cause ('If a measure becomes a target, it stops being a good measure'). If 'speaking like a human' is the target of design, it is no longer a reliable proxy for
English
okay last one i promise - this is just rage baiting me too much
no one is arguing *the words the LLM produces* are conscious - it’s the process generating the words people are interested in
much like no one believes this article is conscious, but ted chiang presumably was when he typed it into his text editor of choice
something about choosing an example where you very easily can imagine - in fact there routinely is - a conscious entity involved really cracks me up
English
@m_shalia @benglickenhaus But I respect you have a different point of view, and apparently study it at some level -- I wish you good luck!
English
@m_shalia @benglickenhaus various complex emergent phenomena when training on large bodies of intentional text.
English
@m_shalia @benglickenhaus whose effects are there in the training set -- but without any real interiority. I think Chiang has all that covered by saying: 'something completely unforeseen about the statistical properties of large corpuses of text, which is a topic worthy of investigation'.
English
@m_shalia @benglickenhaus I tend to trust reviewed articles that pass through real academic filters, but ok, there might be interesting phenomena when training models on large sets of texts written by conscious beings (humans) - the outputs that will emerge will mimic the intentions and reactions...
English
@benglickenhaus at least 10% open to be convinced. And I do think you are a kind person that one seldom meets on the internet like that ;) -- thanks for this experience!
English
@benglickenhaus related to real chemistry/biology of neurons. If it appeared by chance and wasn't created in order to mimic, I would believe it might be conscious. But knowing people try hard to make it behave in a specific way, it is difficult for me to believe. But I also remain...
English