Speaking of negative signals the startup website linked in the substack comes across as one.
A lot of medical buzz words in what is becoming a very crowded area for AI (this is the third opthalmology related AI startup I’ve heard about this month) without any real substance to their claims, details or publications. Opthalmology is the farthest thing from my specialty (radiology) so I can’t comment whether the problem being described is worthwhile but there is also little to no detail provided.
The only substance is the EMNLP paper which is vaguely described as award winning although unspecified and was an interesting read but in my opinion objectively isn’t sufficiently technically impressive to be the whole pitch and hyped on the startup homepage and doesn’t really lend much direct credibility for the problem tackled.
Some of the links are also broken on mobile Safari.
Finally, the author of this article and co-founder is described as “MD Candidate” on the startup page yet his CV states he finished his masters in 2023 and was at Lyft. If this means “admitted to medical school but hasn’t started” suggesting that status imparts any qualification or medical expertise is a huge red flag.
I fully appreciate that this may be a stealth startup but that website is being linked in the post so I interpret that as an advertisement and intended to be read.
Hope to be proven wrong when more comes out but I’d suggest refining that website a bit more.