Will Generative AI save us from bad heuristics? Will it create new ones?
James Gimbi
Rough notes from an email I sent myself; revisited in 2025
We value impressive sounding numbers because they are a proxy for quality. Page views, word counts, units “sold”. Maybe AI eliminates the need to rely on those poor heuristics?
I’m one of those who thinks that chat GPT is going to be like the advent of the calculator. It’s going to allow people to express their intentions very clearly and in ways that are very easily understood very very quickly.
A lot has been said about what this might made in terms of flooding content with cheap unnecessary verbiage. Displacing writers. So forth.
I am sure that there will be negative consequences. But one thing that stands out to me is that it may represent a real alignment of what good speech is and what we look for to recognize it.
Right now we put a great deal of emphasis on the quality of the output. We put a great deal of emphasis on the quality of the speech and the words and how well it strikes the cultural moment.
But generative AI is a grand equalizer. Quality of speech, word choice, cultural relevance, voice, clever presentation and so forth. They are all going to become cheap. They’re all going to become readily accessible. Any fool is going to be able to put together a hell of a good press release, statement, or article about whatever garbage idea they may have.
I think that this will eventually lead to a refocus on intent. I think there is potential that this resolution of talent for writers may turn attention back to intent and inputs and prompts. People will care what you prompt ChatGPT with more than they will care what ChatGPT gave back. We may see a renewed period of recognizing quality over sophistry.
I’m optimistic about what generative AI could mean for how we relate to one another. For how we understand guilt and intention and intellect. If nothing else I’m just plain excited.