Working with AI
Capitalism changes what AI tooling means for engineers. Here is what I see from inside an AI-forward team.
Engineers have always built things to make our own lives easier. There is a saying that what you want on a team is a lazy, annoyed engineer. They find a repetitive process, fix it, demand no kudos, and quietly make everyone else more efficient. Sometimes that fix is glue work. Sometimes it is a new tool. Sometimes it is a full service that automates something nobody else wanted to think about. Most of that work goes unnoticed. It compounds anyway. Teams get faster, companies get cheaper to run, and the people who built the thing move on to the next one.
LLMs are the same pattern at a much broader scale. That is what makes this cycle different. Capitalism has strong opinions about efficiency gains. Every company in the world is right to try to make their teams faster right now. Leaving that value on the table is how you get out-shipped.
The problem is the math does not add up yet.
Take Jensen Huang at his word. He says a 500k engineer needs 300k in tokens. That is 800k per engineer per company. That number is more than many senior engineers make. It is in the range of an SDE2 or a strong SDE1. The industry is not paying that today, and I am not sure it ever will at that rate. We are in the middle of a bet no one has priced correctly.
In the meantime, companies are getting handed an easy justification for downsizing. "AI will fill the gap." It will not. AI can write code. It can make things show up. It cannot connect the dots, come up with ideas, or be left to its own devices. Not at this stage, and probably not for many years. The ceiling on what AI does well is still a human who knows what needs to be built and why.
I work at an AI-forward company, and it is genuinely refreshing. We use it as a velocity booster on the stupid parts of the job. Tasks that used to take two hours ship in ten minutes. Redundant work that wore people down is gone. That is a real win, and it should have been automated a decade ago.
What we do not know yet is the honest economics. Nobody has solved per-token output against engineer output. Companies will claim cost savings because AI increased PR count or lines of code. Anybody who has worked on a team knows what happens when you put a metric in front of people. We hit the metric. We game it. That is not suspicious. That is how incentives work, and no one should be surprised.
The thing I am actually worried about is entry-level jobs. The easy tickets were the ramp. That is where a new engineer learned how the codebase fits together, how the team actually makes decisions, and how a production system behaves at 3am when the dashboards look weird. AI can close those tickets now. It will not close them with the context or the learning attached. The bug gets fixed. The entry point for the next engineer does not exist.
Five years from now, the engineers who understand how things really work are going to be worth more, not less. Knowing how systems connect, how a team ships, how to tell a real problem from a gamed metric, that is the job. AI raises the floor on output. It also raises the price on judgment. That is a trade I will take.