[ AI & ENGINEERING ]
What AI is actually good at in a business — and what it does to engineers
AI is not a replacement for people. It is a very fast colleague who is confidently wrong often enough to matter.
Two conversations get mixed together whenever this comes up: what AI does for a business, and what it does to an engineer. They have different answers, and the second one is the one worth being careful about.
What it does for a business
Decisions made from data rather than instinct, because a model will read a dataset no analyst has time for.
Operations that cost less, because the repetitive middle of most workflows is exactly what automation is for.
Personalisation, because a system that knows what a customer did last can offer them something better than the average.
Prediction, because history usually rhymes and the pattern is in there.
And risk caught earlier — fraud in particular, where the signal is a shape in the data rather than anything a rule was ever written for.
What it does to an engineer
Four habits, in order of how much they matter.
Understand before you delegate. Accepting a solution you could not have written is how a codebase fills with things nobody can debug at three in the morning.
Use it honestly. Attribution and responsible use are not paperwork; they are what keeps a team's output trustworthy.
Keep learning. Skills atrophy when a tool does the reps for you, and the loss is invisible until the day the tool is wrong.
Treat it as a tool. A very good one — but a hammer that suggests where to hit is still a hammer.
Originally published
This first appeared on LinkedIn in February 2024: linkedin.com/in/amar-beka-2771ab1b3