We Have Reached Peak Chatbot
Making bigger, and more hobbled, LLMs just isn’t the way forward.
The internet is full of opinions, some good, some bad, some crazy, but mostly uninformed and absolutely saturated with rage — I’ll let you classify this particular article into any one of those bins at your leisure.
Today I’m going to outline my own thoughts on the way notion “AI” has progressed recently through my own experience of subscribing to OpenAI, being exposed to Bing, and trying to avoid Gemini.
Frankly, I’m so sick of having to disable the damn things in almost every application I start up that I have “AI fatigue” and can’t wait for things to change.
LLMs, or what some people fancifully call “AI,” remain the Big Thing right now (as they have been for around a year or so in their current form), but I genuinely believe that we’ve reached peak velocity of the hype train with what can be done with large scale, eye-wateringly expensive to train, LLMs.
It’s interesting to have a Socratic Dialogue with one, something that I believe is their best use case¹, especially if you’re looking for inspiration, innovative ideas, or are just trying to flesh out an existing nebulous thought or two.
But, they’re an absolutely terrible idea for things like computer language code, actual facts, or any kind of…
