Ask HN: Should engineers who use less inference be paid more?

2 pointskeithwhor2mo ago2 comments

Employee A: $2,000 / mo inference spend (eg Claude code). Ships 5 features a month. High impact. No bugs.

Employee B: $100 / mo inference. Otherwise identical.

Does Employee B deserve a higher salary than Employee A? How should employers think about these variable costs? In some orgs, engineers are being aggressively encouraged to lean on AI as much as possible. But for highly competent contributors shouldn’t that $2,000 / mo in surplus go to eg their kids’ daycare instead of an inference provider?

At what point does inference spend become a red flag? Are there organizations that are punishing people for not using enough?

I’m sure there’s an argument a manager might make that Employee B should be more productive. But at some point you’re just incentivizing waste, right?

2 comments

tjr2mo ago

I think that limitless programmer productivity is a questionable goal.

Extrapolating these numbers, perhaps Employee B is so skillful that, if B did use $2000/month, B could get not 5 features done but 100 features.

That sounds great as an isolated metric, but does the user want 100 new features each month? Does marketing want to advertise 100 new features each month? That sounds like quite a bit.

Or what if B spent $4000/month and completed 200 features? Or spent $10,000/month and completed 500 features?

The software product pipeline has never really had to consider this question too much, because it has usually been the opposite problem: planning to ship 10 features in a month, and only completing 8. Well, if we can now complete 100 features in a month... is that good? Do we actually want to ship that?

Perhaps some discussions need to be had, but there probably should still be some reasonable limit on how many new features to put into a software product in a given timeframe. And then, sort out expected engineering productivity to match.

But to the initial core question: I would think a company / department / team might have some overall monthly budget for AI tool usage. If one person uses more than another, I don't know it really makes sense to give extra cash to those who used it less. Like, if some engineer needed a $2000 compiler tool for their project, and another used GCC for free for their project, does the GCC user get $2000? Not sure things traditionally worked that way.

olivierroy2mo ago

With monthly subscriptions like Claude Code Max, inference spend shouldn't be high enough for this to matter at the moment.

But if an employee is really spending that much on inference, the results they achieve with AI is part of their performance and their inference spend should be considered part of their total compensation.

j / k navigate · click thread line to collapse