My obsession with the anticipation of the future comes from the moment I became a manager, a few years ago.

After seeing the impact of ChatGPT in Microsoft between 2022 and 2023, I understood that AI would be far more challenging for everyone than people were preparing for.

I thought that if I could see it coming, I could coach others to adapt earlier.

It took a toll on me, but I believe I succeeded.

My team became the best in Anaconda at the time at building AI products. Everyone who I had the privilege of managing, leading or mentoring has been ahead of the curve ever since.


My biggest surprise came by the end of last year: how so many engineers pivoted to vibecoding, and with them, so many organizations. The market went from resistance and disbelief to rushing into it with desperation.

It is, I believe, an over-correction, but I digress.

Let’s go back to two years ago.

My methods were therapeutic. I would have individual meetings with people, sometimes for hours, asking them “where do you see yourself in the future,” “is there any vision of the future you perhaps dream of.” That way I would learn what motivated them.

I still don’t believe people learn anything new because somebody tells them about it. Instead I have found more effective to assume they have to be already convinced that the new thing can be leveraged to advance their prior interests.

So, for example, if a data scientist, behind digging and digging through datasets, was actually motivated from usage metrics, then I could lead them to evaluations: well, now see whether those metrics mean anything, what’s the distribution of outputs that users are seeing?

A core challenge that I had was limited understanding of how to communicate appropriately. The skills I used before were raw performance as a programmer. Now I needed storytelling.

Before I learned how to tell stories, I would for example adorn our meetings and documents with art that perhaps could elucidate visions of the future.

To the degree that I can tell stories today, I owe thanks to large language models for helping me understand how English is structured in different contexts — not to mimic them, but to develop my own voice, voice which I am only arriving into very recently.