It’s probably no secret to those who do a lot of focused work in space, but when it comes to generative AI, it quickly becomes apparent that how a user interacts with generative models and systems is at least as important as the underlying training and inference . technology. The latest, and I think the best example comes through OpenAIs ChatGPT, which launched this week as a free research preview for anyone to try.
In case you haven’t seen the buzz around ChatGPT, it’s basically an implementation of their new GPT-3.5 natural language generation technology, but implemented in such a way that you just chat with it in a web browser as if you were at the slacking off with a co-worker or interacting with a customer service representative on a website.
OpenAI has already made waves with its DALL-E image generation technology, and its GPT series has garnered attention (and occasional existential dread on the part of writers) with each successive release. But the latest chat-style iteration has seemingly increased its appeal and audience, in some ways moving the conversation from “wow, students are going to use this to submit bad but workable term papers” to “wow, this could really help me with debugging code I want to put into production.”
Examples so far in the wild seem to show that it really is get much better at the word paper, but that there is still work to be done when it comes to avoiding some typical pitfalls for AI chatbots, including presenting misinformation as fact. But it’s clear that engagement is huge, and people seem to be much more impressed with ChatGPT than at least with GPT-3 (although that’s partly due to the fact that it’s a new release with improved core inference technology as and a new interaction paradigm).
My own example of why I think this is so powerful is topical, albeit mundane: I asked ChatGPT to give me all the different Pokémon Type strengths and weaknesses, and it delivered exactly what I always hope Google will do every time I take part in a Tera Raid in the new Pokémon Scarlet game and have to try and remember what’s going against what.
To know:Note that I wasn’t fancy at all with my question; it’s about as easy as I can get while still being clear on my request. And the result is exactly what I’m looking for – not a list of things that are likely to help me find what I’m looking for if I’m willing to put in the time, which is what Google returns:
The potential for something like OpenAI’s ChatGPT to eventually replace a search engine like Google isn’t a new idea, but this delivery of OpenAI’s underlying technology is the closest yet to how that would work in a fully fleshed-out system, and it should be late Scare Google.
ChatGPT is also the best expression yet of something that startups and entrepreneurs eyeing the space should already know: the gold rush in generative AI will be driven by developing new, defensible businesses built around how it appears, less than what’s under the hood.