AI ROI Scrutiny Is Arriving at the Enterprise Level
Uber's AI president publicly said AI spending is getting 'harder to justify,' and a separate HN thread confirmed Uber blew through its entire AI budget in a single quarter. The mechanism is instructive: Uber created an internal leaderboard ranking teams by total AI tool usage, which predictably caused engineers to burn tokens to improve their rank rather than to solve real problems. Goodhart's Law, applied to AI procurement.
Sam Altman also walked back his earlier predictions about AI causing mass unemployment, saying he was wrong. HN was skeptical, with commenters pointing to the gap between what Altman says and what OpenAI is actually doing. The broader debate on HN about AI jobs hysteria surfaced a real data point: unemployment among recent college graduates is sitting at 5.6%, above the general rate, which some read as early evidence of AI's effect on entry-level knowledge work.
The pattern across all three threads: the AI spending euphoria of 2024 is giving way to actual accountability conversations at large organizations. The question is no longer 'are we using AI' but 'what did we get for it.'
So what?
If you are selling AI tooling to enterprises, the procurement conversation has changed. You need a credible ROI story, not just a demo. Vague claims about productivity improvements will not survive a Q3 budget review. Come in with specific time savings, error reduction rates, or revenue impact tied to actual usage data.
Read these
Uber president says AI spending is getting 'harder to justify'
Uber blows through its AI budget in 1 quarter
Sam Altman: I was wrong, AI unlikely to lead to jobs apocalypse
Don't Subscribe So Casually