A former Meta PM argues that Product Managers shouldn't waste time landing production diffs at the company's scale. Daniel D. McKinnon claims this practice amounts to "snacking" - mistaking motion for progress when PMs should instead fix the prioritization system. The post originated as internal Meta feedback amid "clout-chasing posts" celebrating PMs shipping minor features rather than addressing structural problems.

McKinnon observes PMs code like slow E3 engineers while costing IC7 salaries, accumulating tech debt through pet projects. Despite having connections with senior tech leads who review their code, this approach fails to deliver high leverage. The author argues PMs who land production features are often just "flexing on LinkedIn" rather than creating substantial impact at Meta's scale.

Instead, McKinnon suggests PMs should code to better communicate ideas, understand systems, run realistic experiments, and leverage unique expertise. The concrete recommendation: all PMs should prioritize building EVALS in AI. This provides immediate value while avoiding the inefficiency of PMs acting like junior engineers rather than focusing on product prioritization.