Its a reasonable view to take that "human math" [ math residing in human minds ] is the only math that counts.
Math that only resides in the weights of models, or arcane forms such as a long lean proof or even an unread textbook .. is not the math that we should be striving for.
Likewise all other technology [ and culture ].
LLMs and AI / AGI / ASI could lead to a new renaissance of math discussion and expansion of human math and science. Or the opposite, where we outsource all our thinking to the AI, and no new generation of artisans is trained by doing hard problems, and in a generation we have killed off human math.
Likewise all of the fields of human intellect. We need to make sure we protect future generations of doctors, biologists, software developers, architects, engineers, librarians, musicians, artists ...
A moratorium on AI development might be the only way to achieve this preservation of human culture.
> one of the obvious applications of novel mathematical results is in building stronger AI models.
This gets repeated a lot and seems to be one of the primary stated goals of making AI solve math problems, but I still have no idea by what mechanism this is even supposed to happen. I guess they could make some minor improvements to matrix multiplication algorithms or whatever but I don't see what groundbreaking theorem could possibly significantly improve LLMs.
It's the kind of thing where it's sort of expected that you wouldn't know, right?
I think we don't really understand why deep learning works as well as it does, the thinking around that is, as far as I can tell, mostly a collection of empirical observations.
A fundamental theory of learning that can be used to predict optimal network architectures might enable smaller models that consume less energy.
> Likewise all of the fields of human intellect. We need to make sure we protect future generations of doctors, biologists, software developers, architects, engineers, librarians, musicians, artists ...
So many thoughts come to mind at once, they're a jumble in my head rather than a single coherent narrative.
John Henry comes to mind. As does Agent Smith's "I say your civilization because as soon as we started thinking for you, it really became our civilization, which is, of course, what this is all about" monologue in The Matrix. I've not read (or listened to) "With Folded Hands ..." or "The Machine Stops", but I have read the Wikipedia plot summary of both.
Do we want to have comfortable lives, or do we want to serve each other?
"Computer" used to be a profession; I grew up around adults bemoaning that "kids these days can't do mental arithmetic", the Pi Zero I've not switched on for probably a year now could beat all humans simultaneously at that (even if everyone was as good as the current world record holder) and yet we still teach arithmetic in schools.
Nobody needs to knit, and yet we do so for fun. Youtube's "Primitive Technology" channel, which has spent around a decade speechlessly making iron from bacterial slime found in a creek, using only clay and sticks and leaves and vines naturally found next to that creek.
Like I said, no coherent narrative. It's been a while since my stream of consciousness became a river delta; usually at worst it only meanders a bit.
Understanding the world and math, and discovering beautiful explanations is a worthy pursuit for humans in and of itself.
However, math and science and engineering and biology/medicine have profound implications - those useful parts of that 'culture' - which we call 'technology' - extend human lifespan, healthspan, prosperity, safety.
The main practical value of investing in maintaining a population of math artisans is that they are/were needed as part of a well-trained science community needed to discover and develop new useful technology.
I think its a rational view to say LLM generated "AI-math" is a net positive if and only if, it results in _more_ high quality "human-math". By human math I mean well digested math residing in human minds, being discussed between humans and being actively re-discovered by humans, including a pool of new human student devotees.
A very bad outcome is where we lose the next generation of scientists/lawyers/mathematicians/engineers/doctors/researchers/authors because we let the AI do it all. In this dark future we outsource all our 'thinking', and avoid the years long training of grappling with hard-to-understand aspects of reality. All our culture is sucked into the event horizon of an AI black box.
AI could be a new renaissance of human math- and science- culture, or it could be the death of it.
AI isnt going away, the financial incentives and high current economic inequality, geopolitics guarantee that it will proceed as fast as possible.
AI has given us the structure of nearly every protein, which may well solve Alzheimers and cancers. AI may well find a solution to stable plasma for nuclear fusion, unlocking vast cheap energy and help us humans halt global warming.
So, how do we make sure that the economic windfalls of AI are reinvested back into human-culture, funding more human-math and human science .. resulting in a deeper pool of well educated scientists, engineers and researchers ?
I don't think that will happen by default - we are likely to have an AI-assisted dumbing-down rather than an AI enabled age of enlightenment.
Most of the general public are happy enough not to learn any math - when it doesn't make you much money and university is such a debt burden anyway. School and university students can just get the LLM to do their assignments. Why should they pay tax to fund science at universities when the cost of living is so high, and the LLMs can do all the research anyway ?
I think part of the solution is a policy to tax the windfalls of AI to mitigate the downsides of AI.
Governments should have taxed the carbon polluters and used those funds to mitigate climate change effects and research new forms of clean energy, but didnt.
We should tax LLM/AI profits and use that to fund science research - not just research to mitigate the effects of runaway AI, but to fund general science and math research and teaching, to guarantee that there is a net increase in human-math and human-science.
It is a cautionary tale that fewer students know their times-tables by heart, because a cheap calculator can do that ... but it gives me some hope that there are still people enjoying playing chess, even though chess programs are super-human.
This is true of almost any field - if we have black-box technology that only the AI can understand / fix / augment / extend, then we will lose the ability for humans to learn and understand that tech.
Its similar to outsourcing manufacturing, you lose the ability to manufacture. If you don't exercise a muscle, it atrophies. If some people dont learn to write code [ or do math or write music or books ] the hard slow way, we will lose that deep skill.
If young people don't grapple with hard problems in science, math, economics, creative writing, drawing, music making .. they will not be trained, their cognitive skills will not develop with use. We will have no doctors or engineers or musicians or lawyers in a decade or two.
We don't want to become a race of fatties who never exercise, or dummies who just scroll social media feeds without exercising their minds.
It needs a more coherent cogent discussion, leading to dare I say it sane _policy_ ... which the frontier model companies and hyper-scalars are not incentivized to finance, and governments too slow and weakly funded to finance.
Other topics we seem to have been unable to have deep and informed discussion on, leading to sane policy :
- inequality / taxation
- demographics / aging population
- climate change mitigations
- affect of social media esp on young people
- safety of self-driving cars
- economic impacts of LLMs
- educational "" ""
- security "" ""
- surveillance of in-home devices
- data collection and sharing
- right to repair
Failing to react rationally to a rapidly emerging global pandemic a few years ago should have been a wake up call but wasnt.
If the technorati are woefully unprepared for the age of AI, which is being rapidly thrust upon us by the big money, then the wider public dont stand a chance of being informed / prepared.
Enjoyed this. More importantly its an emerging problem we need to urgently solve.
Apart from my own desire to read the internet to find out what actual people think .. Im worried about a kind of bit-rot, where any decade now, gen-pop reader has no idea what content came from humans and what came from AI.
Im assuming the people who _train_ AI [ LLMs ] actually need to separate the two, and avoid the feedback loop of training the next LLM on the output of the previous LLM.
Presumably, this would lead to a kind of reversion to the mean, akin to making photocopies of photocopies back in the day, resulting in copy degradation.
We already have a kind of bit-rot from moving away from physical media - many documentaries on TV, and audio / video / movies / games are being lost as they are not moved to permanent archive storage.
Old out of print books should be scanned and made public and permanently available online, as they are part of our cultural heritage [ not least for the purposes of training current and future AI as the defacto archives / oracles ]
> Im assuming the people who _train_ AI [ LLMs ] actually need to separate the two, and avoid the feedback loop of training the next LLM on the output of the previous LLM.
Oh they don’t have this data degradation problem at all. They have large data curation teams that scrutinize the “data mix”, ensures the model is always performing better.
I think the current gold-rush of nearly all money into GPU Datacenter and frontier LLMs is essentially starving the economy of innovation.
Academics and founders who might work on developing practical products using NN / ML / RL techniques to solve a realworld problem in engineering/logistics/medicine are not getting investment money. VCs and most people are blind to the fact there is AI outside of LLMs, despite the fact that we have seen AlphaGo and AlphaFold as evidence of non-LLM AI progress in hard domains.
This is perhaps a sub-problem of a larger issue - hyper-inequality means that capital is not allocated to talent [ capital is localized, talent is more widely spread throughout the population ].
We are not getting money to things that will grow our future such as :
- small innovative startups
- university science research
- people who are young enough to have kids, being able to afford them
- new garage bands / authors / musicians / photographers
- public works / infrastructure / libraries
- local retail : bookshop, artisanal bakery, cafe
My thesis is that during the 70s-90s we had higher tax, lower inequality, lower median income to median house price ratio, higher levels of innovation and more original art, literature and music being made.
AI could be a golden age of human flourishing - but thats not where we are heading, what we are seeing is a territory rush by the megacorps.
The fact that RAM and GPU prices have risen so fast, is evidence of supply and demand effect where inequality steals resources from the commons [ middle of the economy ].
Can a talented garage inventor / math or arts student afford a Ryzen AI dev platform, let alone a DGX spark on which to create the next important technology innovation ?
I see results like the OP as speeding up progress. Tests like ptau make it faster and cheaper to screen for participants in clinical trials. This allows more resources to be spent on pre-screening, to identify candidates that might be healthy at time of trial start, but who are likely to develop cognitive impairment during the trial, which can be helpful for better measuring efficacy etc.
There was a good, accessible guidelines paper in the Lancet that described helpful interventions at all stages of life to decrease the odds of developing cognitive impairment later in life.[0]
In terms of late-stage treatments to halt or reverse progression, that still seems very far away. The causes are so diverse... My father died of vascular dementia not too long ago, so I'm familiar with the caregiver side in addition to the business.
While AlphaFold is great, to my knowledge, these proteins have several non-significant IDRs (intrinsically disordered regions), which have been and continue to be hard to model with AlphaFold. I work on computational modeling, so if I had to guess, I'm very unsure if people are using it for whole protein modeling, rather than looking at more specific domains where it probably does much better.
Math that only resides in the weights of models, or arcane forms such as a long lean proof or even an unread textbook .. is not the math that we should be striving for.
Likewise all other technology [ and culture ].
LLMs and AI / AGI / ASI could lead to a new renaissance of math discussion and expansion of human math and science. Or the opposite, where we outsource all our thinking to the AI, and no new generation of artisans is trained by doing hard problems, and in a generation we have killed off human math.
Likewise all of the fields of human intellect. We need to make sure we protect future generations of doctors, biologists, software developers, architects, engineers, librarians, musicians, artists ...
A moratorium on AI development might be the only way to achieve this preservation of human culture.
reply