Yeah - I think this is the schism. Sam is clearly a product person, these are AI people. Dev day didn’t meaningfully move the needle on AI, but for people building products it sure did.
The fact that this is a schism is already weird. Why do they care how the company transforms the technology coming from the lab into products? It's what pay their salaries in the end of the day and, as long as they can keep doing their research work, it doesn't affect them. Being resented about a thing like this to the point of calling it a "absolute embarrassment" when it clearly wasn't is childish to say the least.
this is sort of why henry ford left the company he founded before the ford we know, i think around 01902. his investors saw that they had a highly profitable luxury product on their hands and wanted to milk it for all it was worth, much like haynes, perhaps scaling up to build dozens of custom cars per year, like the pullman company but without needing railroads, and eventually moving downmarket from selling to racecar drivers and owners of large companies, to also selling to senior executives and rich car hobbyists, while everyday people continued to use horse-driven buggies. ford, by contrast, had in mind a radically egalitarian future that would reshape the entire industrial system, labor-capital relations, and ultimately every moment of day-to-day life
for better or worse, ford got his wish, and drove haynes out of the automobile business about 20 years later. if he'd agreed to spend day and night agonizing over how to get the custom paint job perfect on the car they were delivering to mr. rockefeller next month, that wouldn't have happened, and if fordism had happened at all, he wouldn't have been part of it. maybe france or japan would be the sole superpower today
> as long as they can keep doing their research work, it doesn't affect them
That’s a big question. Once stuff starts going “commercial” incentives can change fairly quickly.
If you want to do interesting research, but the money wants you to figure out how AI can help sell shoes, well guess which is going to win in the end - the one signing your paycheck.
> Once stuff starts going “commercial” incentives can change fairly quickly.
Not in this field. In AI, whoever has the most intelligent model is the one that is going to dominate the market. No company can afford not investing heavily in research.
Thinking you can develop AGI - if such a thing actually can exist - in an academic vacuum, and not by having your AI rubber meet the road through a plethora of real world
business use cases strikes me as extreme hubris.
Or the obvious point that if you're not interested in business use cases then where are you going to get the money for the increasingly exorbitant training costs.