Last week I joined Erfan Hesami on his Pipeline to Insights podcast to talk through something I think about constantly: how do you decide what deserves your attention when AI moves faster than anyone can keep up?
We got specific:
➡️ How to tell whether a model is genuinely better or whether someone just built a good setup around it.
➡️ Why the tools that dazzle you in a five-minute look are so often the ones you’d never depend on for daily work.
➡️ Why so many weekend projects feel productive while teaching you nothing. The one change that turns them into something you actually learn from.
➡️ And for the data engineers: why an agent’s reliability comes down to how clearly you describe the tools you give it.
This episode is for technical builders, operators, or leaders trying to give their team room to experiment without chasing every new model release.
