Almost 4 years into the generative AI era, and we’re still treating productivity as the main metric. Specifically, time saved. And you know what? Time saved isn’t the right metric for judging AI because that time is quickly filled with rework, verification, and fixing bad output.
Tag Archives: Artificial intelligence
Why I’m Bullish on Humans
There’s been a lot of conversation about the existential risks posed by AI. Some of it is warranted. We should think about worst-case scenarios. We should pressure-test what could go wrong. We should think seriously about alignment, control, incentives, concentration of power, and what happens when systems become more capable than the people using them.
Which AI Model Is Best Doesn’t Matter. Here’s What Does.
A few days ago, Theo shared his AI model tier list. He placed Fable 5 in its own category. I agree; Fable 5 is a beast. But, you know what? Who cares!
The Biggest Constraint Is the One You Never Questioned
Are you really questioning the status quo? A 10x improvement isn’t a 1% improvement repeated ten times. It’s a step change. The biggest innovations in history didn’t come from doing more of what already existed, just a little better. They came from questioning assumptions that everyone else accepted as true.
Stop Measuring Outputs. Start Managing Inputs.
Yesterday, I participated in a series of round tables on topics related to, surprise, AI. One of those discussions was about Executive Alignment on AI Priorities. One comment stuck with me.
If You’re Just Applying AI To An Inefficient Process, You’re Automating Inefficiency
If you’re just applying AI to an inefficient process, you’re automating inefficiency. A while back, I shared a process mapping framework to help identify where AI can create the most value.
Are Expensive Frontier Models Necessary for Everything?
Don’t use frontier models for non-frontier problems. That’s it.