You Can’t Unscramble an Egg

How technical choices can impact you way down the line in AI.
Here: LiDAR vs. camera.

Technically, I won’t even try to do a better job than Marijn Markus at explaining the technicalities. I’ll leave the link in the comments.

BUT: Sometimes, we still have Finance or IT teams, looking at us weird (or even judjing) when we tell them we start an AI Agent by building... automation.

What we build are the skills that at a later stage are leveraged continuously and with reliability by an AI Agent...
Not a random and generic "tool", an impactful member of the team who knows what it is doing!

The boring stuffs is massively overlooked in the current buzz, yet, we start to see even at our clients how we are preferred to other partners on the long term because it is exponentially beneficial...

We build the fundations, the connectors to all the software you have internally, the skills your team needs from their colleagues, the memory that needs to be leveraged... And at last, when useful, add the most tailored model for the usage...

An easy way to understand it?
You can’t unscramble an egg. 🥚

Once everything is mixed together, it’s much harder to separate what went wrong, fix the underlying ingredients, or rebuild the process properly.

If you put a powerful AI on top of messy data, disconnected systems and unreliable processes, you haven’t magically fixed the foundation.
You’ve just built an intelligent messy colleague.

So yes, sometimes the best way to build AI is to start with something that doesn’t look like AI at all.

Have a great day,
Alexis

—

Originally posted on LinkedIn:

https://lnkd.in/p/ds2WmAeq