Humans + AI… or Just AI + AI?
- egonzalez267
- 6 days ago
- 2 min read
Disclaimer: This post blew up on LinkedIn and clearly hit a nerve. It racked up over 25,000 impressions and sparked a lot of conversation in the insights and AI community.
Written by Kevin Karty, Ph.D. and Founder at Intuify
I want to believe the future is humans + AI. I really do. But I can't shake this unsettling suspicion that Humans + AI is just a short stepping stone to AI + AI. And I think I'm not alone.
Last year I posted on a presentation by Clorox's Oksana Sobol that was the best evidence I've seen that Human + AI is better than either alone. I STILL think that's true: https://lnkd.in/e_6TKnSC
The worry I have is that I'm not sure it will be true in 3 years. Here are a few things that have changed in just the last year:
1) The conventional wisdom was that "prompt engineering" was THE critical skill for the next 10 years. That's soo 2024... These days, if I want an optimized prompt I can just go to ChatGPT (or any other) and ask it to write my prompt for me.
2) At IIEX, the buzz was all about "agentic AI". This was really nothing new - it was a pretty simple job to just stack LLMs on top of each other so that AIs with different functions can be governed by another AI to help avoid common issues and do multi-step tasks.
3) A few days ago, LinkedIn exec Aneesh Raman wrote a public letter warning about something that I and others in market research have been worrying about for a while - the loss of entry-level jobs in MR that can build skills. So it's not just the Cassandras anymore. :)
Let me be clear, we're still pretty far from replacing humans - topline reports written by AI still kind of suck. They're bland and generic. BUT... will they be in 3 years?
At Intuify, our tools are optimized to support the "human in the middle". But a lot of Venture Capital money is betting against the human in the middle, placing wagers on end-to-end AI tools. It will be interesting to see how this shakes out.