The trick to keeping the agents effective is complexity management.
You know you’ve hit a complexity barrier when your agent claims to have implemented something but it’s just.. missing. That’s your sign it’s time to clean things up - maybe simplify it to a pattern that funnels many things through one place. Then the models can work again.
This complexity reduction is a key step in engineering anything really. It makes for a healthier codebase.
And of course, incremental improvements in models will let them go further before hitting their own specific complexity barriers.
I truly believe there IS a durable role for “AI whisperers” of this kind. Folks who can tamp down the complexity as something is built, to keep it under the complexity ceiling of present-day models - and as the crazy pace of the industry lifts that ceiling again and again, folks who know where the limits are so they can work as fast as possible within them.
It’s a very different type of engineering driven by a very different set of constraints. It’s unfamiliar and it’s fascinating watching engineers grapple with it. But it’s here and we have no terminology for what it’s doing to our work, so some people are frightened while others get on with it.
Maybe my analogy is off, but I've been thinking of it like high-end restaurants not being threatened by fast food chains.
I'm trying to understand what I'm missing here. You describe these "AI whisperers" who can tamp down complexity, but doesn't that still require strong fundamental skills and deep knowledge of how to build software? Or am I thinking about this wrong?
If it eventually lifts this ceiling high enough that everyone can build with AI, how does this not simply become minimun wage work?
In which case, what's the unique value proposition?
The way I see it, either:
1. AI gets so good that everyone can build anything (in which case, why pay premium for whisperers?)
2. Or it stays limited enough that you need real engineering skills (in which case, we're back to valuing traditional expertise)
You know you’ve hit a complexity barrier when your agent claims to have implemented something but it’s just.. missing. That’s your sign it’s time to clean things up - maybe simplify it to a pattern that funnels many things through one place. Then the models can work again.
This complexity reduction is a key step in engineering anything really. It makes for a healthier codebase.
And of course, incremental improvements in models will let them go further before hitting their own specific complexity barriers.
I truly believe there IS a durable role for “AI whisperers” of this kind. Folks who can tamp down the complexity as something is built, to keep it under the complexity ceiling of present-day models - and as the crazy pace of the industry lifts that ceiling again and again, folks who know where the limits are so they can work as fast as possible within them.
It’s a very different type of engineering driven by a very different set of constraints. It’s unfamiliar and it’s fascinating watching engineers grapple with it. But it’s here and we have no terminology for what it’s doing to our work, so some people are frightened while others get on with it.
Best you be among the latter.