Meta's AI-assisted coding targets
- Meta's creation organization set a target for 65% of engineers to write 75% of code with AI assistance in H1 2026. - The internal goal frames AI as integral to who writes code and how teams collaborate on engineering tasks. - The target highlights Meta's push to scale AI-driven developer workflows and change internal engineering processes. (x.com)
Meta has set a mid-2026 target for one of its biggest product groups: 65% of engineers should produce more than 75% of their committed code with artificial intelligence help. (msn.com) The goal applies to Meta’s “creation org,” the division tied to core consumer products including Facebook, Instagram, Messenger, and WhatsApp, according to reports citing an internal document reviewed by Business Insider. The same reporting said Meta’s scalable machine learning group had a February 2026 target of 50% to 80% AI-assisted code. (livemint.com) The internal targets went beyond one team. Companywide goals cited from the document included 55% of software engineering code changes being “agent-assisted” and 80% adoption of general AI tools among mid- to senior-level engineers in central product teams by the fourth quarter of 2025. (letsdatascience.com) In plain terms, AI-assisted coding means engineers prompt software to draft, edit, or refactor code instead of typing every line themselves. Meta’s target measures the share of code that gets checked into projects, not just whether a chatbot was open in a browser tab. (livemint.com) Mark Zuckerberg had already signaled this direction in April 2025, when he said he expected artificial intelligence to write most of Meta’s code for some efforts within 12 to 18 months. The new internal benchmarks put numbers and deadlines on that prediction. (engadget.com) Meta has been spending at a scale that matches that push. The company reported $72.22 billion in capital expenditures for full-year 2025, a figure that included heavy investment in the computing infrastructure needed to train and run artificial intelligence systems. (atmeta.com) The coding targets also arrived during a broader reset of Meta’s workforce. In January 2025, the company said it would cut about 5% of staff, or roughly 3,600 employees, in performance-based layoffs as it pushed managers to move faster on underperformers. (cnbc.com) That combination suggests Meta is treating AI fluency less like an optional productivity tool and more like a job requirement for large parts of engineering. By the first half of 2026, the question inside Meta is not whether engineers use AI to write code, but how much of their code they can ship with it. (letsdatascience.com)
