Others wrap it in euphemism; these said it plainly
In this column I usually dig into corporate language, because layoffs always get wrapped as 'restructuring,' 'focus,' 'agility.' This week two companies tore the wrapping off. Amsterdam-based hotel software firm Mews cut 15% of its 1,350-person staff — about 170 people — and founder Richard Valtr named the reason directly: AI is 'fundamentally changing the economics of hospitality... capable of handling much of the execution layer of that work intelligently and at scale, and more profitably.' Underline those two words, 'more profitably.' The cut came six months after a $300 million funding round and a $2.5 billion valuation. The company isn't sinking; on the contrary, it has cash and cuts anyway.[1]
The second case is barer still. Coinbase head of platform Rob Witoff says 95-100% of code is now AI-written — it was 40% in February. That disclosure comes right after the company cut about 14% of staff, roughly 700 people, in May. CEO Brian Armstrong had said AI 'dramatically' changed the work, with engineers finishing in days what took weeks. Add the numbers: on one side record valuations and 77% profit jumps, on the other the sentence 'with fewer people now.'[1], [2], [3]
Whose pocket does the efficiency go into?
Let me be fair: I'd take Mews's plain speaking over the dozens of firms that wrap it up. Valtr calls the people he laid off 'exceptional stewards of growth,' notes customer-facing roles were largely spared, and that 36 positions are still posted. I grant those nuances. What I don't grant is the hidden subject the word 'efficiency' carries. Efficiency is never neutral; it favors someone and costs someone. Witoff says humans stay central in cryptography and security — good; but where did those 700 people doing the routine work go?[1], [2]
The same week, Nadella warned companies they 'pay twice' for proprietary models, reminding them the exhaust of their data flows into someone else's model. I take that warning one floor down: the exhaust of labor flows through the same pipe. What cuts an engineer's weeks-long task to days is a model trained on that engineer's past labor. The productivity gain is real — the numbers are there. Let my question be constructive: how much of the value created by that gain flows back to the people who made it possible? This week the industry wrote the layoffs too plainly to hide behind 'restructuring.' It might as well write the answer plainly too.[2], [4]