5 ms·

That's great, but AlphaGo used artificial and constrained training materials. It's a lot easier to optimize things when you can actually define an objective sco

That's great, but AlphaGo used artificial and constrained training materials. It's a lot easier to optimize things when you can actually define an objective score, and especially when your system is able to generate valid training materials on its own.

nopinsight 1y ago
There are quite a few relatively objective criteria in the real world: real estate holdings, money and material possessions, power to influence people and events, etc. The complexity of achieving those might result in the "Centaur Era", when humans+computers are superior to either alone, lasting longer than the Centaur chess era, which spanned only 1-2 decades before engines like Stockfish made humans superfluous. However, in well-defined domains, like medical diagnostics, it seems reasoning models alone are already superior to primary care physicians, according to at least 6 studies. Ref: When Doctors With A.I. Are Outperformed by A.I. Alone by Dr. Eric Topol https://substack.com/@erictopol/p-156304196 https://substack.com/@erictopol/p-156304196
- jimbo808 1y ago
  It makes sense. People said software engineers would be easy to replace with AI, because our work can be run on a computer and easily tested, but the disconnect is that the primary strength of LLMs is that they can draw on huge bodies of information, and that's not the primary skill programmers are paid for. It does help programmers when you're doing trivial CRUD work or writing boilerplate, but every programmer will eventually have to be able to actually truly reason about code, and LLMs fundamentally cannot do that (not even the "reasoning" models). Medical diagnosis relies heavily on knowledge, pattern recognition, a bunch of heuristics, educated guesses, luck, etc. These are all things LLMs do very well. They don't need a high degree of accuracy, because humans are already doing this work with a pretty low degree of accuracy. They just have to be a little more accurate.
  - Davidzheng 1y ago
    I think at this point it's an absurd take that they aren't reasoning. I don't think without reasoning about code (& math) you can get to such high scores on competitive coding and IMO scores. Alphazero also doesn't need training data as input--it's generated by game-play. The information fed in is just game rules. Theoretically should also be possible in research math. Less so in programming b/c we care about less rigid things like style. But if you rigorously defined the objective, training data should also be not necessary.
    - Jensson 1y ago
      > Alphazero also doesn't need training data as input--it's generated by game-play. The information fed in is just game rules This is wrong, it wasn't just fed the rules, it was also fed a harness that did test viable moves and searched for optimal ones using a depth first search method. Without that harness it would not have gained superhuman performance, such a harness is easy to make for Go but not as easy to make for more complex things. You will find the harder it is to make an effective such harness for a topic the harder it is to solve for AI models, it is relatively easy to make a good such harness for very well defined programming problems like competitive programming but much much harder for general purpose programming.
      - Davidzheng 1y ago
        the harness is obtained from the game rules? the "harness" is part of the algorithm of alphzero
        Jensson 1y ago
        > the "harness" is part of the algorithm of alphzero Then that is not a general algorithm and results from it doesn't apply to other problems.
      - GolDDranks 1y ago
        Are you talking about Monte Carlo tree search? I consider it part of the algorithm in AlphaZero's case. But agreed that RL is a lot harder in real-life setting than in a board game setting.
  - jll29 1y ago
    Having worked briefly in the medical fields in the 1990s, there is some sort of "greedy matching" being pursued, so once 1-2 well-known symptoms are recognized that can be associated with diseases, the standard interventions to cure are initiated. A more "proper" approach would be to work with sets of hypotheses and to conduct tests to exclude alternative explanations gradually - which medics call "DD" (differential diagnosis). Sadly, this is often not systematically done, and instead people jump on the first diagnosis and try if the intervention "fixes" things. So I agree there are huge gains from "low hanging fruits" to be expected in the medical domain.
  - iteria 1y ago
    Being a walking encyclopedia is not what we pay doctors for either. We pay them to account for the half truths and actual lies that people tell about their health. This is to say nothing about novel presentations that come about because of the genetic lottery. Same as an AI can assist but not replace a software engineer, an AI can assist but not replace a doctor.
- jacquesm 1y ago
  Humans are statistically speaking static. We just find out more about them but the humans themselves don't meaningfully change unless you start looking at much longer time scales. The state of the rest of the world is in constant flux and much harder to model.
  - gizajob 1y ago
    I’m not sure I agree with this - it took humans about a month to go from “wow this AI generated art is amazing” to “zzzz it’s just AI art”.
    - jacquesm 1y ago
      That's culture, not genetics.
    - MSFT_Edging 1y ago
      To be fair, it was more a "wow look what the computer did". The AI "art" was always bad. At first it was just bad because it was visually incongruous. Then they improved the finger counting kernel, and now it's bad because it's a shallow cultural average. AI producing visual art has only flooded the internet with "slop", the commonly accepted term. It's something that meets the bare criteria, but falls short in producing anything actually enjoyable or worth anyone's time.
      - Earw0rm 1y ago
        It sucks for art almost by definition, because art exists for its own reason and is in some way novel. However, even artists need supporting materials and tooling that meet bare criteria. Some care what kind of wood their brush is made from, but I'd guess most do not. I suspect it'll prove useless at the heart of almost every art form, but powerful at the periphery.
ben_w 1y ago
Sure, that does make things easier: one of the reasons Go took so long to solve is that one cannot define an objective score for Go beyond the end result being a boolean win or loose. But IRL? Lots of measures exist, from money to votes to exam scores, and a big part of the problem is Goodhart's law — that the easy-to-define measures aren't sufficiently good at capturing what we care about, so we must not optimise too hard for those scores.
- Jensson 1y ago
  > Sure, that does make things easier: one of the reasons Go took so long to solve is that one cannot define an objective score for Go beyond the end result being a boolean win or loose. Winning or losing a Go game is a much shorter term objective than making or losing money at a job. > But IRL? Lots of measures exist No, not that are shorter term than winning or losing a Go game. A game of Go is very short, much much shorter than the time it takes for a human to get fired for incompetence.
  - ben_w 1y ago
    Time horizon is a completely different question to what I'm responding to. I agree the time horizon of current SOTA models isn't particularly impressive. Doesn't matter in this point.
FrustratedMonky 1y ago
"artificial and constrained training materials" Are you simply referring to games having a defined win/loss reward function? Because pretty sure Alpha Go was ground breaking also because it was self taught, by playing itself, there were no training materials. Unless you say the rules of the game itself is the constraint. But even then, from move to move, there are huge decisions to be made that are NOT easily defined with a win/loss reward function. Especially early game, there are many moves to make that don't obviously have an objective score to optimize against. You could make the big leap and say that GO is so open ended, that it does model Life.
- neltnerb 1y ago
  That quote was intended to mean -- "artificial" maybe I should have said "synthetic"? I mean the computer can teach itself. "constrained" the game has rules that can be evaluated and as to the other -- I don't know what to tell you, I don't think anything I said is inconsistent with the below quotes. It's clearly not just a generic LLM, and it's only possible to generate a billion training examples for it to play against itself because synthetic data is valid. And synthetic data contains training examples no human has ever done, which is why it's not at all surprising it did stuff humans never would try. A LLM would just try patterns that, at best, are published in human-generated go game histories or synthesized from them. I think this inherently limits the amount of exploration it can do of the game space, and similarly would be much less likely to generate novel moves. https://en.wikipedia.org/wiki/AlphaGo https://en.wikipedia.org/wiki/AlphaGo > As of 2016, AlphaGo's algorithm uses a combination of machine learning and tree search techniques, combined with extensive training, both from human and computer play. It uses Monte Carlo tree search, guided by a "value network" and a "policy network", both implemented using deep neural network technology.[5][4] A limited amount of game-specific feature detection pre-processing (for example, to highlight whether a move matches a nakade pattern) is applied to the input before it is sent to the neural networks.[4] The networks are convolutional neural networks with 12 layers, trained by reinforcement learning.[4] > The system's neural networks were initially bootstrapped from human gameplay expertise. AlphaGo was initially trained to mimic human play by attempting to match the moves of expert players from recorded historical games, using a database of around 30 million moves.[21] Once it had reached a certain degree of proficiency, it was trained further by being set to play large numbers of games against other instances of itself, using reinforcement learning to improve its play.[5] To avoid "disrespectfully" wasting its opponent's time, the program is specifically programmed to resign if its assessment of win probability falls beneath a certain threshold; for the match against Lee, the resignation threshold was set to 20%.[64]

nopinsight 1y ago

There are quite a few relatively objective criteria in the real world: real estate holdings, money and material possessions, power to influence people and events, etc. The complexity of achieving those might result in the "Centaur Era", when humans+computers are superior to either alone, lasting longer than the Centaur chess era, which spanned only 1-2 decades before engines like Stockfish made humans superfluous. However, in well-defined domains, like medical diagnostics, it seems reasoning models alone are already superior to primary care physicians, according to at least 6 studies. Ref: When Doctors With A.I. Are Outperformed by A.I. Alone by Dr. Eric Topol https://substack.com/@erictopol/p-156304196 https://substack.com/@erictopol/p-156304196

jimbo808 1y ago
It makes sense. People said software engineers would be easy to replace with AI, because our work can be run on a computer and easily tested, but the disconnect is that the primary strength of LLMs is that they can draw on huge bodies of information, and that's not the primary skill programmers are paid for. It does help programmers when you're doing trivial CRUD work or writing boilerplate, but every programmer will eventually have to be able to actually truly reason about code, and LLMs fundamentally cannot do that (not even the "reasoning" models). Medical diagnosis relies heavily on knowledge, pattern recognition, a bunch of heuristics, educated guesses, luck, etc. These are all things LLMs do very well. They don't need a high degree of accuracy, because humans are already doing this work with a pretty low degree of accuracy. They just have to be a little more accurate.
- Davidzheng 1y ago
  I think at this point it's an absurd take that they aren't reasoning. I don't think without reasoning about code (& math) you can get to such high scores on competitive coding and IMO scores. Alphazero also doesn't need training data as input--it's generated by game-play. The information fed in is just game rules. Theoretically should also be possible in research math. Less so in programming b/c we care about less rigid things like style. But if you rigorously defined the objective, training data should also be not necessary.
  - Jensson 1y ago
    > Alphazero also doesn't need training data as input--it's generated by game-play. The information fed in is just game rules This is wrong, it wasn't just fed the rules, it was also fed a harness that did test viable moves and searched for optimal ones using a depth first search method. Without that harness it would not have gained superhuman performance, such a harness is easy to make for Go but not as easy to make for more complex things. You will find the harder it is to make an effective such harness for a topic the harder it is to solve for AI models, it is relatively easy to make a good such harness for very well defined programming problems like competitive programming but much much harder for general purpose programming.
    - Davidzheng 1y ago
      the harness is obtained from the game rules? the "harness" is part of the algorithm of alphzero
      - Jensson 1y ago
        > the "harness" is part of the algorithm of alphzero Then that is not a general algorithm and results from it doesn't apply to other problems.
    - GolDDranks 1y ago
      Are you talking about Monte Carlo tree search? I consider it part of the algorithm in AlphaZero's case. But agreed that RL is a lot harder in real-life setting than in a board game setting.
- jll29 1y ago
  Having worked briefly in the medical fields in the 1990s, there is some sort of "greedy matching" being pursued, so once 1-2 well-known symptoms are recognized that can be associated with diseases, the standard interventions to cure are initiated. A more "proper" approach would be to work with sets of hypotheses and to conduct tests to exclude alternative explanations gradually - which medics call "DD" (differential diagnosis). Sadly, this is often not systematically done, and instead people jump on the first diagnosis and try if the intervention "fixes" things. So I agree there are huge gains from "low hanging fruits" to be expected in the medical domain.
- iteria 1y ago
  Being a walking encyclopedia is not what we pay doctors for either. We pay them to account for the half truths and actual lies that people tell about their health. This is to say nothing about novel presentations that come about because of the genetic lottery. Same as an AI can assist but not replace a software engineer, an AI can assist but not replace a doctor.
jacquesm 1y ago
Humans are statistically speaking static. We just find out more about them but the humans themselves don't meaningfully change unless you start looking at much longer time scales. The state of the rest of the world is in constant flux and much harder to model.
- gizajob 1y ago
  I’m not sure I agree with this - it took humans about a month to go from “wow this AI generated art is amazing” to “zzzz it’s just AI art”.
  - jacquesm 1y ago
    That's culture, not genetics.
  - MSFT_Edging 1y ago
    To be fair, it was more a "wow look what the computer did". The AI "art" was always bad. At first it was just bad because it was visually incongruous. Then they improved the finger counting kernel, and now it's bad because it's a shallow cultural average. AI producing visual art has only flooded the internet with "slop", the commonly accepted term. It's something that meets the bare criteria, but falls short in producing anything actually enjoyable or worth anyone's time.
    - Earw0rm 1y ago
      It sucks for art almost by definition, because art exists for its own reason and is in some way novel. However, even artists need supporting materials and tooling that meet bare criteria. Some care what kind of wood their brush is made from, but I'd guess most do not. I suspect it'll prove useless at the heart of almost every art form, but powerful at the periphery.

ben_w 1y ago

Sure, that does make things easier: one of the reasons Go took so long to solve is that one cannot define an objective score for Go beyond the end result being a boolean win or loose. But IRL? Lots of measures exist, from money to votes to exam scores, and a big part of the problem is Goodhart's law — that the easy-to-define measures aren't sufficiently good at capturing what we care about, so we must not optimise too hard for those scores.

Jensson 1y ago
> Sure, that does make things easier: one of the reasons Go took so long to solve is that one cannot define an objective score for Go beyond the end result being a boolean win or loose. Winning or losing a Go game is a much shorter term objective than making or losing money at a job. > But IRL? Lots of measures exist No, not that are shorter term than winning or losing a Go game. A game of Go is very short, much much shorter than the time it takes for a human to get fired for incompetence.
- ben_w 1y ago
  Time horizon is a completely different question to what I'm responding to. I agree the time horizon of current SOTA models isn't particularly impressive. Doesn't matter in this point.

FrustratedMonky 1y ago

"artificial and constrained training materials" Are you simply referring to games having a defined win/loss reward function? Because pretty sure Alpha Go was ground breaking also because it was self taught, by playing itself, there were no training materials. Unless you say the rules of the game itself is the constraint. But even then, from move to move, there are huge decisions to be made that are NOT easily defined with a win/loss reward function. Especially early game, there are many moves to make that don't obviously have an objective score to optimize against. You could make the big leap and say that GO is so open ended, that it does model Life.

neltnerb 1y ago
That quote was intended to mean -- "artificial" maybe I should have said "synthetic"? I mean the computer can teach itself. "constrained" the game has rules that can be evaluated and as to the other -- I don't know what to tell you, I don't think anything I said is inconsistent with the below quotes. It's clearly not just a generic LLM, and it's only possible to generate a billion training examples for it to play against itself because synthetic data is valid. And synthetic data contains training examples no human has ever done, which is why it's not at all surprising it did stuff humans never would try. A LLM would just try patterns that, at best, are published in human-generated go game histories or synthesized from them. I think this inherently limits the amount of exploration it can do of the game space, and similarly would be much less likely to generate novel moves. https://en.wikipedia.org/wiki/AlphaGo https://en.wikipedia.org/wiki/AlphaGo > As of 2016, AlphaGo's algorithm uses a combination of machine learning and tree search techniques, combined with extensive training, both from human and computer play. It uses Monte Carlo tree search, guided by a "value network" and a "policy network", both implemented using deep neural network technology.[5][4] A limited amount of game-specific feature detection pre-processing (for example, to highlight whether a move matches a nakade pattern) is applied to the input before it is sent to the neural networks.[4] The networks are convolutional neural networks with 12 layers, trained by reinforcement learning.[4] > The system's neural networks were initially bootstrapped from human gameplay expertise. AlphaGo was initially trained to mimic human play by attempting to match the moves of expert players from recorded historical games, using a database of around 30 million moves.[21] Once it had reached a certain degree of proficiency, it was trained further by being set to play large numbers of games against other instances of itself, using reinforcement learning to improve its play.[5] To avoid "disrespectfully" wasting its opponent's time, the program is specifically programmed to resign if its assessment of win probability falls beneath a certain threshold; for the match against Lee, the resignation threshold was set to 20%.[64]