this post was submitted on 05 May 2024
27 points (100.0% liked)
TechTakes
1389 readers
138 users here now
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I haven't read it yet (and I might not even bother) but Google's probably lying about their ML models again
"I can predict the structure and interactions of all of life’s molecules"
it does seem mildly useful https://www.science.org/content/blog-post/screening-against-alphafold-models-maybe-better-it-looks
note that google's thing is pattern matching, and if there's few similar protein structures in literature, then alphafold is working on poor quality data. comments do notice a red flag or two
Can’t see the comments on there unfortunately, but interesting start to that piece. Will read it in parts today as I get time
if you're using firefox, disable enhanced tracking protection to see disqus comments
Ah my solution to this is different: if the things that I use to protect my shit are blocking some $x, ah well that can just get fucked then
Too exhausted by all the bullshit to try debug someone’s fuckery and specialcase them. Only exception is if it’s something I need (and then I’ll do so very grumpily, and quite possibly loudly)
got you
i mean specifically this one
The rest of criticisms is in article
hahaha that's an amazing comment, and yeah agreed/understood
also note it's from january so not about the new thing
also note that the big failures come from faulty biology, not from faulty chemistry. You can set innumerable piles of cash on fire if you choose to plug a wrong enzyme
elaborating on that. when you can hit a protein cleanly (without some side effects (called off-target, there are also on-target side effects)) but you hit wrong protein, this doesn't have to be a complete loss, but it's likely. maybe you learn something, either about biology of that all or about something else that it interacts with, or maybe you have found a cure to some other disease instead. but this is the actual hard problem of drug development, especially when the biology of it all is little known. there's a few of these drugs that started this way, fentanyl, viagra, thalidomide (as anticancer), minoxidil come to mind, and at worst you have a new tool for biologists. alphafold won't help with that