>> seismic interpretation here
Strong disagree here. Lets put aside the math and focus on money.
I dont know much about seismic interpretation, but I know a lot about Radiology+CV/ML. I was CTO+CoFounder for three years full time of a venture-backed Radiology+CV/ML startup.
From what I can see, there is a huge conflict of interest w/r/t Radiology (and presumably any medical field) in the US. Radiologists make a lot of money -- and given their jobs are not tied to high CoL regions (as coders jobs are), they make even more on a CoL-adjust basis. Automating these jobs is the equivalent of killing the golden goose.
Further, Radiologists standards of practice are driven partly by their board (The American Board of Radiology) and the supply of labor is also controlled by them (The American Board of Radiology) by way of limited residency spots to train new radiologists.
So Radiologists (or any medical specialist) can essentially control the supply of labor, and control the standards of best practice, essentially allowing continued high salaries by way of artificial scarcity. WHY ON EARTH WOULD THEY WANT THEIR WORK AUTOMATED AWAY?
My experience during my startup was lots of radiologists mildly interested in CV/ML/AI, interested in lots of discussions, interested in paid advisory roles, interested in paid CMO figurehead-positions, but mostly dragging their feet and hindering real progress, presumably because of the threat it posed. Every action item was hindered by a variety of players in the ecosystem.
In fact, most of our R&D and testing was done overseas in a more friendly single payer system. I dont see how the US's fee-for-service model for Radiology is ever compatible with real progress to drive down costs or drive up volume/value.
Not surprisingly, we made a decision to mostly move on. You can see Enlitic (a competitor) didnt do well either despite the star-studded executive team. Another competitor (to be unnamed) appears to have shifted from models to just licensing data. Same for IBM/Merge.
Going back to seismic interpretation -- this cant be compared to Radiology from a follow-the-money perspective because seismic interpretation isnt effectively a cartel.
Happy to speak offline if anyone is curious about specific experiences. DM me.
Strong disagree here. Lets put aside the math and focus on money.
I dont know much about seismic interpretation, but I know a lot about Radiology+CV/ML. I was CTO+CoFounder for three years full time of a venture-backed Radiology+CV/ML startup.
From what I can see, there is a huge conflict of interest w/r/t Radiology (and presumably any medical field) in the US. Radiologists make a lot of money -- and given their jobs are not tied to high CoL regions (as coders jobs are), they make even more on a CoL-adjust basis. Automating these jobs is the equivalent of killing the golden goose.
Further, Radiologists standards of practice are driven partly by their board (The American Board of Radiology) and the supply of labor is also controlled by them (The American Board of Radiology) by way of limited residency spots to train new radiologists.
So Radiologists (or any medical specialist) can essentially control the supply of labor, and control the standards of best practice, essentially allowing continued high salaries by way of artificial scarcity. WHY ON EARTH WOULD THEY WANT THEIR WORK AUTOMATED AWAY?
My experience during my startup was lots of radiologists mildly interested in CV/ML/AI, interested in lots of discussions, interested in paid advisory roles, interested in paid CMO figurehead-positions, but mostly dragging their feet and hindering real progress, presumably because of the threat it posed. Every action item was hindered by a variety of players in the ecosystem.
In fact, most of our R&D and testing was done overseas in a more friendly single payer system. I dont see how the US's fee-for-service model for Radiology is ever compatible with real progress to drive down costs or drive up volume/value.
Not surprisingly, we made a decision to mostly move on. You can see Enlitic (a competitor) didnt do well either despite the star-studded executive team. Another competitor (to be unnamed) appears to have shifted from models to just licensing data. Same for IBM/Merge.
Going back to seismic interpretation -- this cant be compared to Radiology from a follow-the-money perspective because seismic interpretation isnt effectively a cartel.
Happy to speak offline if anyone is curious about specific experiences. DM me.