The company closes; working life goes on sale
Employees leave more than code when a startup closes. Messages exchanged while debugging, drafts revised to meet a deadline and meeting calendars remain in the company’s possession. An investigation by MIT Technology Review describes how these traces of working life become assets sold to AI laboratories. Keelan Miskell sold the code from his closing cryptocurrency startup for about $25,000; he expects another transaction covering Google, Notion, Figma and Slack content to bring roughly $15,000. Code and the everyday collaboration that produced it are priced separately. Even after the business stops operating, the way its employees worked together can become material for another product.[1]
For buyers, the attraction lies in how work happens rather than only in a polished final answer. Surge AI builds reinforcement learning environments for tasks including debugging and legal work; its Nick Heiner describes looking for difficult decisions and situations where people disagree. Sunset’s pricing considers employee numbers, time spent operating, content volume and tools used. The sale therefore turns more than the delivered result into commercial value: it also monetizes the time and relationships involved in reaching it. That explains the interest in employee experience. It does not, by itself, establish that real workplace data produces better training outcomes than synthetic data.[1]
It would be incomplete to read this market solely as employee labor changing hands. Paying a closing company’s debts and recovering value from its remaining assets are real needs too. The more than 150 deals reported for SimpleClosure’s Asset Hub show that a group of buyers has emerged within that liquidation process. My inference is that this recovery function simultaneously creates a distribution problem: material controlled by the company gains value again without requiring a new negotiation with the people who produced it. The alternative explanation is ordinary liquidation of an asset the company already owns. Yet an ownership calculation does not settle who receives the proceeds and who bears the risk.[1]
Proceeds and risk have different destinations
The investigation says sale proceeds generally go to senior creditors and investors rather than employees. Privacy risk, meanwhile, does not disappear because the company has closed. Both sides of a sale say they remove personal information; lawyer Greg Ewing points out that calendars can still contain details about medical appointments and childcare. This is the other side of distribution: commercial benefits follow the company’s payment hierarchy, while the cost of intimate information circulating can remain with the individual concerned. Removing identifying information and preserving a worker’s control over their own life are different operations.[1]
Ifeoma Ajunwa of Emory University therefore argues for obtaining employees’ explicit consent. Her recommendation should not be treated as an already established legal requirement for every such sale. It nevertheless reveals a concrete decision point: data cleaning can be purchased as a technical service, while involving the person who produced the content is a separate governance choice. A company’s control over its archive does not mean the employee has a say in each new use. In my assessment, it is more coherent to count workers as producers of the collaboration being sold, alongside treating them as holders of personal information that needs protection.[1]
Completing the distribution account means considering three parties together: the selling company and its creditors, the laboratory buying material for training, and the employee whose labor and private life appear in it. Liquidation revenue is tangible for the first party; the size of any training benefit is an outcome the buyer still needs to demonstrate. The worker’s share and voice do not arise automatically from a price negotiation between those two sides. A high sale price is therefore insufficient as a measure of a good agreement. A more revealing measure asks which material is transferred with whose consent, and who shares in the benefit of that transfer. Recognizing collaboration’s value also requires keeping its producers visible in the distribution account.[1]