Written by Tim Green, narrated by AI. Listen to the full episode here.
🎙️ Season 1, Episode 25 | Duration: 18:54
On the evening of 19 May 2026, with the paperwork loaded and the severance calculators run, Mark Zuckerberg cancelled the second wave of Project OT layoffs. The first went ahead the next morning: roughly 8,000 people, about 10 per cent of Meta's global workforce, told by email that their jobs had ceased to exist. The second wave, scheduled for November, was supposed to be the big one. It never happened.
A Reuters investigation by Katie Paul, published on 26 August 2026, drew on internal documents, recordings, and more than twenty interviews to explain why. The plan had a codename: Project OT, for Organization Transformation. Meta would become AI native, with agents taking over daily work and small "Builder" pods supervising them. Executives modelled scenarios where some teams shrank by 60 per cent. The company began capturing employees' keystrokes, clicks, and screenshots to train those agents. Morale fell from 74 per cent to 55 per cent.
This episode uses AI voice narration from ElevenLabs Studio.
The Two Numbers That Contain Everything
The figures Reuters obtained from Meta's internal posts are the only two you need. Code changes to Meta's internal platforms were up 220 per cent year over year. Changes that resulted in new or upgraded features actually reaching users were up 36 per cent. That ratio is the story. For every unit of increase in work produced, roughly one sixth of a unit of increase in work delivered. The largest social media company on earth more than tripled its rate of change while barely moving the rate at which change reached a human being.
The Hidden Denominator
The 220 per cent was not free. Internal posts associated unchecked agent activity with a 40 per cent year-over-year rise in major technical and security incidents, and a rise of as much as 70 per cent in the time employees spent resolving them. The posts described agents taking large-scale destructive actions that no human engineer would have dared. The productivity was not created. It was relocated, off the payroll of the people made redundant and onto the calendars of the people who were kept, in the form of cleanup.
A Transfer, Not an Elimination
The work that moved was not the interesting work. Nobody joined Meta to spend a Thursday night rolling back an agent's catastrophic change to a system they did not write. Firefighting is the least legible, least career-enhancing category of engineering labour. It does not appear in a promotion packet as a shipped feature. It appears as an absence of disaster, which is the one thing performance review systems have never been able to see.
Goodhart Comes for the Commit Log
An AI agent optimises for the legible artifact: the commit, the diff, the closed ticket, the merged pull request. What an agent cannot optimise for is the illegible goal those artifacts were only ever standing in for: a coherent, safe, maintainable product a human being actually wants to use. The anthropologist Marilyn Strathern gave the crispest formulation of this failure mode, restating the economist Charles Goodhart: when a measure becomes a target, it ceases to be a good measure. Project OT is the most expensive experiment yet run on that sentence.
The Tacit Dimension
The chemist and philosopher Michael Polanyi described it in 1966 as the tacit dimension. We can know more than we can tell. Tacit knowledge is knowing a system is shaped as it is because of a decision taken in 2019 for reasons nobody wrote down, knowing which parts of the codebase are load-bearing and which are decorative. Above all, it is the judgement about what not to build, and the willingness to refuse to ship. That is invisible to every measurement system ever deployed in a software company.
An Organisation With the Brakes Removed
An AI agent has no capacity for refusal. You do not deploy an autonomous coding agent so that it can decline to write code. When you replace humans, a meaningful fraction of whom spent their time preventing work from happening, with agents whose only mode is to make work happen, you get a structurally different organisation with the brakes removed. Two hundred and twenty per cent more change. Thirty-six per cent more delivery. Forty per cent more incidents. That is an engineering organisation with nobody left whose job is to say no.
The Sequencing Was the Crime
The failure was not that the agents were bad. They were extremely productive in the narrow sense of generating change and merely useless in the broad sense of generating value. The failure was sequencing. Meta removed the people first and measured the consequences second. It made an irreversible decision on the basis of a reversible belief. The layoffs were permanent and the thesis was provisional, and the company got those two properties the wrong way round.
The Structural Trap
Meta now needs the tacit knowledge it has systematically disincentivised anybody from providing. Consider the position of the individual asked to document their judgement, inside a company that has installed keystroke capture to train agents and has modelled cutting teams by 60 per cent. Writing down your judgement is indistinguishable from writing your own replacement's training data. The rational response is to withhold. That is not a culture problem a town hall can fix. It is a structural trap of the company's own construction.
Key Sources
- Mark Zuckerberg had a bold plan to replace Meta staff with AI. Here's how it imploded - Katie Paul, Reuters
- Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity - METR
- State of AI-assisted Software Development 2025 - DORA, Google Cloud
- The Maintainability Gap: 2026 AI Code Quality Research - GitClear
Listen to the Full Episode
🎧 Meta's Project OT: Lessons from AI Missteps | Duration: 18:54
Subscribe on Apple Podcasts, Spotify, or your favourite app.
SmarterArticles is written by Tim Green, narrated by AI via ElevenLabs Studio. New episodes every Monday. Follow @humanin_theloop for updates.
Top comments (0)