Honestly part of the challenge here is that they’re not really doing anything truly groundbreaking that everyone else doesn’t figure out on their own weeks later. That’s not an environment that fits the traditional “publish” mindset and is a big problem for these big labs as everything slides towards being an undifferentiated commodity.
We’ve really not had massive advancements in the last few years that weren’t quickly discovered/copied by everyone at the same time, just bigger models. Everyone has been just playing around with the same mathematical parlor tricks from the original breakthrough.
Until that changes I expect we’ll continue to see this arms-race-to-the-bottom as everyone is just battling to make a commodity vs true innovation that gives one startup a legit IP advantage over others.
We’ve really not had massive advancements in the last few years that weren’t quickly discovered/copied by everyone at the same time, just bigger models. Everyone has been just playing around with the same mathematical parlor tricks from the original breakthrough.
Until that changes I expect we’ll continue to see this arms-race-to-the-bottom as everyone is just battling to make a commodity vs true innovation that gives one startup a legit IP advantage over others.