I feel like this is similar to how Stalin and Hitler really liked to envision massive fighting machines the size of buildings. “If APC < tank < bigger tank, then I just need to build THE BIGGEST TANKS.” It’s like some kind of disease that infects people who aren’t actually doing any of the engineering, just overseeing it, and then taking the most simple minded shortcut they can think to how to make everything more “winning.”
It’s already happening. Llama 4’s release wasn’t just a mess, but a dishonest one they cheated at/lied about. They fired a bunch of researchers, are shuffling teams around to focus more on ‘products,’ and seem to be hiring on Tech Bros instead.
In other words, the dysfunction is well underway. Which is kinda sad, actually, as the Llama team basically pioneered locally runnable LLMs (as opposed to AI Bro API models) with pretty lean resource access.
I feel like this is similar to how Stalin and Hitler really liked to envision massive fighting machines the size of buildings. “If APC < tank < bigger tank, then I just need to build THE BIGGEST TANKS.” It’s like some kind of disease that infects people who aren’t actually doing any of the engineering, just overseeing it, and then taking the most simple minded shortcut they can think to how to make everything more “winning.”
It’s already happening. Llama 4’s release wasn’t just a mess, but a dishonest one they cheated at/lied about. They fired a bunch of researchers, are shuffling teams around to focus more on ‘products,’ and seem to be hiring on Tech Bros instead.
In other words, the dysfunction is well underway. Which is kinda sad, actually, as the Llama team basically pioneered locally runnable LLMs (as opposed to AI Bro API models) with pretty lean resource access.