100%! I’ve been seeing zero sum tension across my timeline. Structural biologists vs computational scientists. Big company vs biotech. Research vs industry This is a collective win! While far from perfect it opens up a world of research & discovery heretofore inaccessible.https://twitter.com/sbarnettARK/status/1333916174401363968 …
-
Sure but given this came via CASP I don’t understand the default position taken by so many. I’m less so worried about Deepmind benchmarking this against well accepted benchmarks. More so worried about if this will be released and/or open sourced. And who gets access when & how.
This is also an odd case where another firm replicating (sidestepping the access problem) is gated by engineering effort not compute effort (which imo, is pretty modest here compared to ex. recent nlp) exact opposite of the gpt3 problem
