very interesting! the only point we would bring up for discussion is the idea of immovable. Meta is dominant for hospitals running their own infra. Epic + Mayo Clinic just integrated GPT-4 deeply, so the real idea is bifurcation not Meta dominance.
Thought about this and there is a more nuanced take. My two portcos with Mayo on the slide deck are both asterisks. One with Mayo as an investor is a SAFE note in exchange for data+patient access. The other as a customer they got strongly preferential pricing. This is a known thing that they do consistently and have found multiple times in data rooms during diligence.
I wonder what sort of preferential terms Mayo extracted for the signalling lift of being a customer ahead of the IPO.
Agree. It's the smart move, but Mayo (and anyone else) will jump ship to another provider once its convenient.
The interesting thing about Meta "dominance" is that they derive no revenue from LLama variants or distillations. Not even in a RedHat consulting / implementation way. When Abridge deploys their internal models, that's margin that they get to keep.
TBH, I'm glad they're doing it - and it aligns with the original OpenAI thesis - but I don't know what the endgame is.
very interesting! the only point we would bring up for discussion is the idea of immovable. Meta is dominant for hospitals running their own infra. Epic + Mayo Clinic just integrated GPT-4 deeply, so the real idea is bifurcation not Meta dominance.
Thought about this and there is a more nuanced take. My two portcos with Mayo on the slide deck are both asterisks. One with Mayo as an investor is a SAFE note in exchange for data+patient access. The other as a customer they got strongly preferential pricing. This is a known thing that they do consistently and have found multiple times in data rooms during diligence.
I wonder what sort of preferential terms Mayo extracted for the signalling lift of being a customer ahead of the IPO.
Agree. It's the smart move, but Mayo (and anyone else) will jump ship to another provider once its convenient.
The interesting thing about Meta "dominance" is that they derive no revenue from LLama variants or distillations. Not even in a RedHat consulting / implementation way. When Abridge deploys their internal models, that's margin that they get to keep.
TBH, I'm glad they're doing it - and it aligns with the original OpenAI thesis - but I don't know what the endgame is.