alt Hacker News> Drug design: Using Mythos 5, our internal protein design experts accelerated aspects of the drug design process by around ten times. In one example, they found that Mythos 5, with protein design and bioinformatics tools but no human assistance, matches or beats skilled human operators. In doing so, the model executes all of the tasks that are normally completed by a scientist: choosing binding sites, selecting and running protein design tools, and recovering from failures along the way. Nine of the 14 protein targets from this study (shown below) yielded strong candidates for drug design that we’re currently investigating.
How is this half-way down the page? To me it's the headline.
Replies
Because it's completely meaningless without validation, and even with validation, not really any better than the state of the art protein generation models. Which are also mostly just nice to have because coming up with a candidate is generally quite easy.
The rate limiting steps are generally testing, or characterizing. Not designing protein binders.
It's selective reporting. Says 'in one example', but out of how many, is that one-shot, or is it a random result out of 100. It's a marketing doc.
Would be funny if anthropic ends up as mostly a pharma company
Drug design isn't the bottleneck anymore, it's trials. Still cool they can do this with a general purpose model though.
There are tons of ways to generate "strong candidates for drug design." This is definitely not the bottleneck in drug discovery and development. The hard problem is vetting and developing these ideas to the point of having a commercially viable drug. That is still a very empirical process.