AI means engineering can fix almost anything now.
That does not make the job easier. It makes prioritization the job.
Value starts with no
If you say yes to everything, you devalue everything.
The moment you say no, the thing you chose becomes gold. The rest becomes dust.
Until tomorrow morning, when you look at the evidence and prioritize again.
That is the work now. Not proving that something can be built. Deciding whether it deserves to exist.
I am more ruthless with my team
If you can fix everything, why are you fixing useless stuff?
Why are you polishing a page nobody visits? Why are you tuning a query that is already fast enough? Why are you building a feature before proving that anyone needs it?
“We can do it” is no longer an argument.
Tell me what changes if we do it. Tell me what we are refusing to do instead.
Stop flying blind
Every change needs a number before it needs code.
Where is the dashboard? Where are the logs, traces, errors, conversion rates, response times, and costs?
If the dashboard does not exist, why are you still touching the system?
You should be able to measure the problem before the change, watch the change land, and prove what moved after it shipped.
Otherwise you are not prioritizing. You are guessing faster.
The new engineering bar
AI made execution cheap. Judgment became expensive.
Your value is no longer how much you can fix.
It is what you refuse to fix and whether the thing you chose was worth it.
