What happened
Meta reportedly scrapped an ambitious plan to become an 'AI-native' company. The proposal, detailed in a report from Ars Technica AI, involved replacing significant human roles with AI agents. The most aggressive version of the plan included slashing headcount on some teams by as much as 60 percent. The project was ultimately abandoned due to its complexity and internal resistance.
How the room's reading it
The report is being read as a reality check for the 'AI-first' mantra. On social platforms, many engineers see it as proof that replacing complex human workflows with agents is far from a solved problem — especially at Meta's scale. The discussion highlights the immense organisational and technical hurdles. Some analysts frame it as a sign of maturity, suggesting the initial hype around autonomous agents is now meeting the practical friction of implementation. The consensus is that while AI can augment teams, the idea of wholesale replacement remains a distant, and perhaps undesirable, prospect.
Sailfish's take
We see this as a classic case of chasing a narrative instead of a product. The goal shouldn't be to replace people — it should be to ship better things, faster. We've found the most effective use of AI is to give small, talented teams superpowers, not to dismantle them. Trying to automate entire job functions with today's agentic tech is a recipe for organisational chaos and brittle systems. The real work is in the messy middle, integrating AI thoughtfully into existing workflows. If you're being asked to build a 'human-out' system, this is the week to push back.