Live data from Hacker News

Generative A.I. arrives in the gene editing world of CRISPR

nytimes.com

61–69 of 69 posts

Re: Generative A.I. arrives in the gene editing world of CRISPR

#63
post #17

Reading their blog post I wonder if an LLMs is really the best way to do this. If I got it right, they used the LLM to enumerate potential protein DNA sequences. Does that really need an LLM? Enumeration is not novel, nor are LLMs particularily good at it. If you want to computationally parallelize the search in a large enumeration space it would be much easier to simply, well, do that instead of taking a detour via…

It's not an English LLM, but a "protein" language model, where tokens represent amino acids or nucleotides. Learning a transformer language model on such data simply learns a distribution over sequences of tokens. It's a fine approach conceptually that in many ways is the "right" way or most elegant method, and not a stretch at all.

yeah but your training is bottlenecked by the lack of ground truth. Some things were(I presume)/will be easy to do with LLMs, like protein structure, because every part of every protein is source data (and there's millions of known structure). But suppose you want to estimate clearance, or ld50. How many proteins do we know their serum clearance? 1000? 10000 maybe?

Re: Generative A.I. arrives in the gene editing world of CRISPR

#64

What can possibly go wrong if we let ChatGPT edit our DNA?

This model has nothing to do with ChatGPT other than transformers. And as someone that could desperately use some advances in gene editing, this lowbrow dismissal is frustrating.

More and more I find the flippant dismissal of the amazing breakthroughs that are constantly streaming in kind of disheartening.

Nobody is impressed by anything anymore, nobody is excited about the future - everything’s just “terrible” and everything looking forward is bad. I hate that mindset and I’m sick and tired of cynicism poisoning practically every online space. Thank you for saying it’s “lowbrow” and giving OP a little bit of a hard time.

Re: Generative A.I. arrives in the gene editing world of CRISPR

#66
post #58

Earlier quoted context omitted.

I enjoyed the feeling when I made this connection talking with a startup doing this a while back. It's just a different "language" and although it's not a given that LLMs can operate in it, it's a reasonable thing to try, and it turns out they can.

Personally I think it was obvious that LLMs were going to be useful for protein modelling since the previous generation used HMMs very successfully. Pfam (a library of HMMs for classifying proteins into preexisting known families) is one of the most important resources we have because of the power of HMMs to model sequential language. I suspect we will need to move from sequential modelling to graphical modelling to…

> I suspect we will need to move from sequential modelling to graphical modelling to level-up again, though.

Out of curiosity, would you mind elaborating on this?

Re: Generative A.I. arrives in the gene editing world of CRISPR

#67

Earlier quoted context omitted.

This model has nothing to do with ChatGPT other than transformers. And as someone that could desperately use some advances in gene editing, this lowbrow dismissal is frustrating.

More and more I find the flippant dismissal of the amazing breakthroughs that are constantly streaming in kind of disheartening. Nobody is impressed by anything anymore, nobody is excited about the future - everything’s just “terrible” and everything looking forward is bad. I hate that mindset and I’m sick and tired of cynicism poisoning practically every online space. Thank you for saying it’s “lowbrow” and giving O…

the breakthroughs in tech are great, the way they are marketed, sold, and pushed in our lives by means of adtech everywhere... not so much.

Re: Generative A.I. arrives in the gene editing world of CRISPR

#68
post #58

Earlier quoted context omitted.

Personally I think it was obvious that LLMs were going to be useful for protein modelling since the previous generation used HMMs very successfully. Pfam (a library of HMMs for classifying proteins into preexisting known families) is one of the most important resources we have because of the power of HMMs to model sequential language. I suspect we will need to move from sequential modelling to graphical modelling to…

> I suspect we will need to move from sequential modelling to graphical modelling to level-up again, though. Out of curiosity, would you mind elaborating on this?

I don't work in the field so I'm probably just repeating something Hinton already said, but it seems to me like attempting to model things in reality that have graph-like structures (like interacting pairs of residues in a 3d protein structure) using sequences with finite context lengths is ultimately going to be less efficient than modelling graphs. My guess is this work is roughly describing that I think of: https://www.cis.upenn.edu/~mkearns/papers/barbados/jordan-tu...

it could also be I completely misunderstand context in sequential models and what I'm describing is already being used, or has been evaluated and has been unsuccessful.

Re: Generative A.I. arrives in the gene editing world of CRISPR

#69
post #67

Earlier quoted context omitted.

More and more I find the flippant dismissal of the amazing breakthroughs that are constantly streaming in kind of disheartening. Nobody is impressed by anything anymore, nobody is excited about the future - everything’s just “terrible” and everything looking forward is bad. I hate that mindset and I’m sick and tired of cynicism poisoning practically every online space. Thank you for saying it’s “lowbrow” and giving O…

the breakthroughs in tech are great, the way they are marketed, sold, and pushed in our lives by means of adtech everywhere... not so much.

Part of me says - and this is somewhat of a hot take - "then stop using platforms that mash ads into everything?" I don't know, like, I'm here for the rebel yell against global industrial capitalism - let's replace it with the solar punk world of our dreams, but... like, we're not going to do that any time soon, so instead of complaining about "how terrible everything is" let's try to do better. Let's be excited about this stuff. Let's be positive.
Post reply on HN