I don't know if it's so much that talented people are being locked out, as much as it is that communities everywhere, not just industry, are requiring a level of people skills that academic people lack but nonetheless thrive without.
There's quite a lot of people skills involved in surviving as an academic in today's environment. Imagine if you had to teach calculus to 150 random, uninterested teenagers (barely adults) every 12 weeks. There's some serious people skills involved in doing a good job at that (most people do actually try to teach well, I've known multiple people this year refused tenure based on rate-my-teacher ratings).
It's a different set of skills for sure, but being an academic isn't as socially challenged as the zeitgeist appears to believe.
i like academics, don't misunderstand me.
yes, or at least largely overlapping circles
what i am saying is, having people skills are the answers "yes" to all those questions. you can cynically call getting a job nepotism, or you can call it, well people like to work with their friends at the cost of measures of competency. and maybe, the core competency is being pleasant to work with or work for.
another place people struggle with this is executive compensation. if i told every DoD employee they could get a 10x better boss for only $20/y, every single one would, which is $58m in executive compensation. but the DoD CAN'T do that, and its leadership is TERRIBLE, so... do you see?
There is no communication there. No concepts for them to communicate. It is just math.
I'm not sure that models are complex enough to have a consistent internal representation of a concept the same way that organic brains can to communicate. I'm not sure of any quantitative science backing this up though. Models don't know anything across iterations yet.
Can you expand. They have both context and memory?
I'm sure there are research prototypes that work differently from this but I haven't seen any enter the mainstream yet.
Also, diffusion language models have a different evaluation order but I think they also do not really have internal thoughts or feelings because they also do not seem to have any sort of hidden state that encodes anything like that.
I've had worse. Mostly much better, but I've had worse.
You cannot share (effectively) if you cannot communicate in a way that others can understand.
Further the entire ecosystem that academics rely on to get what they need to do for their research (grants, and other funding, resources, and so on) necessitate them to convince people who control those, who do not necessarily understand the purpose of the work