March 31, 2026 – A former senior technical documentation engineer at Snowflake, Randy Urbano, took to LinkedIn to seek new job opportunities after being laid off. This came as the U.S. cloud – data platform provider announced in March that it was eliminating its entire technical writing and documentation team, affecting around 70 employees.
Before the layoffs, Urbano and his colleagues were mainly responsible for restructuring product documentation into a format that was more suitable for AI large models and program agents to access. Their work enabled machines to better understand the products, but ironically, it also seemed to pave the way for their own displacement. Snowflake simply stated that the move was to align the team with the company’s long – term strategy.
This case is a vivid example of the wave of AI – related layoffs sweeping across the global tech industry. These technical writers transformed fragmented product knowledge into machine – friendly content, creating a core capability for enterprises to leverage AI. However, they ended up being the ones to be laid off.
Similar layoff events are becoming increasingly common in Silicon Valley. Around the same time, Atlassian, the software giant known for its project management tool Jira, announced that it would lay off 1,600 employees, accounting for about 10% of its global workforce.
The layoffs were not limited to support functions but also extended to core departments such as engineering, product, and design. Even front – line employees deeply involved in the implementation of AI products were not spared. Mike Cannon – Brookes, the company’s co – founder, admitted that although AI was not directly replacing people, it had significantly changed the skill sets required by the company and the number of positions in certain areas.
E – commerce giant Amazon has also been making large – scale staff adjustments during its AI transformation. Since 2025, it has laid off approximately 30,000 corporate employees. At the same time, it is strongly promoting the use of its in – house AI programming tool, Kiro.
Some employees revealed that Kiro often generates code with defects, but they are forced to use it, getting stuck in a vicious cycle of “using AI to fix problems created by AI”. Moreover, the code and work content written by employees are ultimately used as training data for AI.
In addition, companies like Crypto.com and Pinterest have also initiated layoffs under the guise of AI transformation and resource reallocation, presenting organizational downsizing as a necessary step for technological transformation.
The core logic behind these layoffs is not that AI can completely replace human work. The criterion for enterprises is quite simple: as long as AI can break down complex work processes and enable a small number of people to complete more tasks, it is sufficient to justify reducing the workforce.
AI doesn’t need to be perfect. As long as it is useful enough, the “cold winter” of layoffs in white – collar knowledge work, which was once just an industry prediction, has now become a reality.
