4 ms·

I don't agree, even a human chef with decades of experience in all the tools and ingredients they're using is not going to get every recipe right on the first t

I don't agree, even a human chef with decades of experience in all the tools and ingredients they're using is not going to get every recipe right on the first try. I'm not saying an AI recipe will never be good, but it is going to make mistakes sometimes. If you're wondering whether a task is theoretically possible for an AI, I think a good rule of thumb is to ask "could a human domain expert do this given a few days and as much research material as they need". If the task here is "develop a novel recipe without trying it", the answer is no. You can't predict the interactions of every possible combination of ingredients and processes no matter how much time and experience you have.

petra 3y ago
>> You can't predict the interactions of every possible combination of ingredients and processes no matter how much time and experience you have. That's just the thing machines do better than humans. Not sure about LLM's though.
famouswaffles 3y ago
You misunderstand what a language model is. What a human can or cannot do is frankly irrelevant. A GPT (Generative pretrained Transformer) is not a simulator nor an imitator. It is a predictor. Predictors try to reverse engineer the computation that could have led to a certain output so they don't just grok what is explicitly stated in a dataset but also what is implied by its structure. As an example, In a language model trained on protein sequences alone and nothing else, you will find biological structure and function emerge in the inner layers. https://www.pnas.org/doi/full/10.1073/pnas.2016239118 https://www.pnas.org/doi/full/10.1073/pnas.2016239118 In fact, with a capable enough architecture and sufficient data (quantity, qualit, variance), there is nothing a predictor couldn't divine with enough training. You assert >Actually generating a recipe is obviously much harder, and should be effectively impossible unless you include a humanoid robot that can cook and taste. In the pursuit of recipe prediction, a predictor will learn to internally model taste because taste is implicit in the data. You don't need the hypothetical perfect predictor to demonstrate your assertion as false because a capable predictor already does. >If you're wondering whether a task is theoretically possible for an AI, I think a good rule of thumb is to ask "could a human domain expert do this given a few days and as much research material as they need". You are wrong. Train a language model on descriptions of the functions of proteins and an equivalent protein sequence and you get a language model that can predict novel functioning sequences from function descriptions alone. https://www.nature.com/articles/s41587-022-01618-2 https://www.nature.com/articles/s41587-022-01618-2 Not only is this not something a human expert can achieve in a couple days, It's not something a human expert can achieve at all. It is a Super-Human ability.
- uoaei 3y ago
  I don't think the commenter's understanding of LLMs is relevant here. It's what the authors of TFA think LLMs are for and, arguably more importantly, what kinds of problems (nails) they think about when presented with the idea of applying LLMs (hammer). It's a bit ridiculous that this is the best they can come up with, and I don't really think the idea is worth defending, especially by telling other people that they just don't get it.
  - famouswaffles 3y ago
    His/Her primary assertion is that it is impossible for a LLM to generate novel recipes without physically tasting it first. His/her assertion is wrong and one easily corrected by understanding a LLM is and does. It's not a perfect predictor so it's not impeccable for all conceivable situations but GPT-4 can already generate novel recipes that taste nice. This is not some far flung science fiction ability.
    - ska 3y ago
      This is true. On the other hand, the recipes aren't typically good, and generating novel ones isn't really what people mostly want to do... so the value seems limited.
      - famouswaffles 3y ago
        That's part of it yes but people tend to forget that the road to capable prediction is not instantaneous. It's possible the data is enough to get a good predictor but the one you tested simply wasn't good enough yet.
        ska 3y ago
        Possibly, but the easily available data itself seems is pretty poor for systemic reasons.
    - ForestCritter 3y ago

petra 3y ago

>> You can't predict the interactions of every possible combination of ingredients and processes no matter how much time and experience you have. That's just the thing machines do better than humans. Not sure about LLM's though.

famouswaffles 3y ago

You misunderstand what a language model is. What a human can or cannot do is frankly irrelevant. A GPT (Generative pretrained Transformer) is not a simulator nor an imitator. It is a predictor. Predictors try to reverse engineer the computation that could have led to a certain output so they don't just grok what is explicitly stated in a dataset but also what is implied by its structure. As an example, In a language model trained on protein sequences alone and nothing else, you will find biological structure and function emerge in the inner layers. https://www.pnas.org/doi/full/10.1073/pnas.2016239118 https://www.pnas.org/doi/full/10.1073/pnas.2016239118 In fact, with a capable enough architecture and sufficient data (quantity, qualit, variance), there is nothing a predictor couldn't divine with enough training. You assert >Actually generating a recipe is obviously much harder, and should be effectively impossible unless you include a humanoid robot that can cook and taste. In the pursuit of recipe prediction, a predictor will learn to internally model taste because taste is implicit in the data. You don't need the hypothetical perfect predictor to demonstrate your assertion as false because a capable predictor already does. >If you're wondering whether a task is theoretically possible for an AI, I think a good rule of thumb is to ask "could a human domain expert do this given a few days and as much research material as they need". You are wrong. Train a language model on descriptions of the functions of proteins and an equivalent protein sequence and you get a language model that can predict novel functioning sequences from function descriptions alone. https://www.nature.com/articles/s41587-022-01618-2 https://www.nature.com/articles/s41587-022-01618-2 Not only is this not something a human expert can achieve in a couple days, It's not something a human expert can achieve at all. It is a Super-Human ability.

uoaei 3y ago
I don't think the commenter's understanding of LLMs is relevant here. It's what the authors of TFA think LLMs are for and, arguably more importantly, what kinds of problems (nails) they think about when presented with the idea of applying LLMs (hammer). It's a bit ridiculous that this is the best they can come up with, and I don't really think the idea is worth defending, especially by telling other people that they just don't get it.
- famouswaffles 3y ago
  His/Her primary assertion is that it is impossible for a LLM to generate novel recipes without physically tasting it first. His/her assertion is wrong and one easily corrected by understanding a LLM is and does. It's not a perfect predictor so it's not impeccable for all conceivable situations but GPT-4 can already generate novel recipes that taste nice. This is not some far flung science fiction ability.
  - ska 3y ago
    This is true. On the other hand, the recipes aren't typically good, and generating novel ones isn't really what people mostly want to do... so the value seems limited.
    - famouswaffles 3y ago
      That's part of it yes but people tend to forget that the road to capable prediction is not instantaneous. It's possible the data is enough to get a good predictor but the one you tested simply wasn't good enough yet.
      - ska 3y ago
        Possibly, but the easily available data itself seems is pretty poor for systemic reasons.
  - ForestCritter 3y ago