Any subject matter experts care to chime in with your gut feeling about where this sits in the "promising potential treatment" to "AI psychosis" spectrum?
>Maybe he publishes and it goes somewhere, I have a feeling getting approval for trials will be challenging with his setup as it is now.
Yeah, drug development is one of those things where practically the only way to make it to production is to let a big incumbent with a big enough war chest to drag it across the finish line buy it.
If it isn't a potentially big money-maker said big incumbent won't touch it though. Will more likely try to shut it down especially if it overlaps with any of their more ludicrous offerings. And given the properties that this has (easy to manufacture, stable, soluble, low/no toxicity, etc) it's ripe for a shutdown because it'd make AD treatments far too cheap.
"Potentially promising but likely to fail to replicate in human/non-mouse trials" is the easy default for new drug announcements, it's the modal outcome for drugs in this level of development, we've cured mouse cancers dozens of times over the years after all. The guy seems to have relevant expertise in the field so less likely to be purely AI driven nonsense.
What does the guy's basement have to do with it? We live in the internet age, we can do anything anywhere. Was there not any other more relevant information to fluff up the headline? By the way, why are we still using twitter?
"Meanwhile, LLMs (mainly ChatGPT Pro) were deeply integrated into virtually every step of the discovery process - I wouldn’t have been able to complete this project without them"
> All of the in vitro screening was performed by an OpenTrons OT-2 liquid-handling robot programmed by Claude Code. Meanwhile, LLMs (mainly ChatGPT Pro) were deeply integrated into virtually every step of the discovery process - I wouldn’t have been able to complete this project without them
That does not seem to be an accurate description of what has happened here.
What it looks like is someone with significant biochemical experience and a Harvard PhD has created some kind of drug or chemical that he thinks will be effective for treating Alzheimer's, and that he mentions using Claude Code to help him program some of the complex chemical engineering machines that he used along the way.
Maybe he publishes and it goes somewhere, I have a feeling getting approval for trials will be challenging with his setup as it is now.
I do like the "do chemistry the old fashion way" ethos though. I wish he added a blurb about why he's doing it this way
Yeah, drug development is one of those things where practically the only way to make it to production is to let a big incumbent with a big enough war chest to drag it across the finish line buy it.
https://xcancel.com/DouglasYaoDY/status/2070904914050797582
He used AI to program robot arms.
What it looks like is someone with significant biochemical experience and a Harvard PhD has created some kind of drug or chemical that he thinks will be effective for treating Alzheimer's, and that he mentions using Claude Code to help him program some of the complex chemical engineering machines that he used along the way.
However, it does seem to have a few undesirable side effects.