[ryohkyo]

Perhaps he meant overhyped; I am just guessing. The creator of GPT-3 also said that the model is more hyped than it deserves because it's simply looking at the existing data to forecast the next words.

Hype or not, there are some questions that deserve some serious consideration. It is hard to quantify how "personalized" the content GPT-3 spits out actually is. It may output the same content on every N-th try on average. If so, we have copyright issues at hand.

Two people using GPT-3 who happened to feed it similar parameters and got similar results can sue each other for copywrite infringement. On the other hand, an author who genuinely created an article can sue someone using GPT-3 created content without that person knowing he has infringed on the author's copyright, and vice versa.

It's all wonderful on paper. In practice, a whole host of issues can pop up when using this indiscriminately.

[robbedpeter]

It's looking at the existing data in the context of everything it's been trained on. It's forecasting in a very sophisticated way. We're a couple years away from seeing fully realized applications with these models that exploit the most clever and nuanced uses. Moore's law is trundling along in the meantime, and we're 6-8 years from seeing these massive transformer models show up in a flagship phone.

It might be overhyped, but it's also a massively broad-use tool that is going to provide value for many years to come. We've hardly scratched the surface of its utility.