These systems are a function of their data more than anything else. They're a function of the training data. And basically, the big models are trained by scraping the internet and pulling in basically all human-generated training data, all human-generated text, and increasingly video and audio and everything else. And there's just literally only so much of that.
We're running out of human knowledge.
They're so gated on data that they're literally going out and hiring thousands of programmers and doctors and lawyers to actually hand write answers to questions for the purpose of being able to train their AIs.
This is not the future that was feared when OpenAI's ChatGPT burst onto the scene almost two years ago.
Collection
[
|
...
]