CEO Dara Khosrowshahi slowed hiring to offset the expenditures.
If other companies report the same, the bubble pops. 🫧
Uber COO Andrew Macdonald said he's not seeing proportional productivity gains from increasing AI costs. https://bit.ly/4e3w4PC
Many users condemned Uber's AI token spending as unreliable and wasteful without clear ROI or consumer features, while some noted real productivity gains from adoption.
No Digg Deeper questions have been answered for this story yet.
Uber’s COO has said that it’s getting “harder to justify” its AI costs because there was no way to show a link between AI spend and any meaningful increase in useful features. This is the first time I’ve seen a company say this directly.
https://www.businessinsider.com/uber-coo-andrew-macdonald-ai-token-spending-harder-justify-2026-5
If enough other companies report the same, the bubble pops. 🫧
Uber COO Andrew Macdonald said he's not seeing proportional productivity gains from increasing AI costs. https://bit.ly/4e3w4PC
My interview with @praveenTweets for @theinformation sparked internal discussions at Uber about the return on AI costs, per @BusinessInsider
Developer productivity is hard to manage.
Uber’s COO has said that it’s getting “harder to justify” its AI costs because there was no way to show a link between AI spend and any meaningful increase in useful features. This is the first time I’ve seen a company say this directly.
https://www.businessinsider.com/uber-coo-andrew-macdonald-ai-token-spending-harder-justify-2026-5
Pretty sure 50% of internal token spend is completely useless, but right now it's hard to know which 50%.
As an admin I'd love a dashboard that breaks down each person's spend into summarized clusters. Much easier to spend more when you can draw a clear line to value.
Uber’s COO has said that it’s getting “harder to justify” its AI costs because there was no way to show a link between AI spend and any meaningful increase in useful features. This is the first time I’ve seen a company say this directly.
https://www.businessinsider.com/uber-coo-andrew-macdonald-ai-token-spending-harder-justify-2026-5
@edzitron Interesting
@edzitron An image I use in my software engineering class:
I think that speaks less to token costs and more to the fact that there’s never been a good singular sellable metric to show developer productivity. That is the original sin.
Many leaders have cited and still cite “number of diffs” as the metric which has been known to be a bad metric for 30+ years.
So when that metric fails and costs go up, I’m not surprised you get these headlines.
This is an organizational issue imo. Like of course if you give unlimited tokens to SWEs they are going to get off on spinning up fucking pet projects, its like crack
You gotta distill down your processes and pluck them one by one with AI
Uber’s COO has said that it’s getting “harder to justify” its AI costs because there was no way to show a link between AI spend and any meaningful increase in useful features. This is the first time I’ve seen a company say this directly.
https://www.businessinsider.com/uber-coo-andrew-macdonald-ai-token-spending-harder-justify-2026-5
BI report here: https://www.businessinsider.com/uber-coo-andrew-macdonald-ai-token-spending-harder-justify-2026-5
@echantech1 Yeah it’s not the costs it’s just the amount they have to spend on the service and its relation to any measurement of any kind
@andyinsdca right but the question is whether it’s increasing productivity in proportion to costs and COO is saying no.
My original interview here: https://www.theinformation.com/newsletters/applied-ai/uber-cto-shows-claude-code-can-blow-ai-budgets
MSFT cancelling Claude Code sub internally is most likely fake news, but companies having a hard time justifying AI spending is real. However, it's mostly a skill issue. AI turns a 10x engineer into a 100x one, but also turns an incompetent one into a coding slopping machine.
🦔Uber's COO Andrew Macdonald said on Saturday that the company is having a harder time justifying its AI spend. After CTO Praveen Neppalli Naga went viral in April for admitting Uber burned through its 2026 Claude Code budget in four months, senior engineering leaders concluded higher token usage was not translating into proportionally more useful product. Macdonald said the link between AI consumption and shipped features is "not there yet." CEO Dara Khosrowshahi confirmed on the earnings call that Uber is slowing hiring to fund its AI spend. Duolingo also walked back its decision to include AI usage in performance reviews last month.
My Take Uber is the first major enterprise where the C-suite has publicly admitted, on the record, that the AI productivity story is not closing for them. That matters because Uber is not a skeptic. The company went all-in on AI tooling, set internal targets, and burned through its annual research and development budget in four months trying to make it work. The conclusion from the people running the experiment is that tokens consumed and value shipped are not the same number, and management is finally noticing.
Duolingo's reversal lands in the same week for a reason. CEO Luis von Ahn said employees were asking whether they needed to use AI just to use AI, which is Goodhart's Law showing up in a performance review system. When usage becomes the metric, employees optimize for usage, not output. Microsoft canceled internal Claude Code licenses, Google AI Pro stripped credits from paid subscribers, and now Uber is admitting the ROI does not close at scale. The narrative has shifted in the last 30 days from "AI productivity is here" to "AI productivity is harder to measure than we thought." The companies pushing tokenmaxxing internally are now the same companies signaling cost pressure externally. The IPO calendar for OpenAI and Anthropic is going to get a lot more complicated if the largest enterprise customers keep saying this out loud.
Hedgie🤗
@edzitron To be fair to Uber, at least they're asking the hard questions internally. Most companies are just burning cash and hoping the board doesn't notice.
@bloomai_com FOMO. but
For whatever reason, many engineering leaders have sort of failed at the basic engineering practices they enforce at their own companies.
If a L4 engineer came out and said "Let's replace the whole codebase with Rust and we have no way of measuring progress or success/failure" , they'd be laughed out of the room.
But when a leader does it with AI, its okay?
I really don't know why people are surprised that crap like this is happening.
@edzitron Skill issue
Anthropic itself roles out insane new features weekly
Or product maturity issue — maybe there’s just not that much more to invent with rideshare
@GaryMarcus I feel like I saw a post on X earlier today where Uber had 5000 engineers using AI...ah, here it is!
@GaryMarcus Then why are there record compute contracts and tokens produced every day?
