Practitioners in Lenny's members-only Slack are wrestling with three concrete problems this week: non-PMs bypassing product process to ship directly to production, Anthropic running a live pricing A/B test on Claude Code subscriptions, and the specific ways generative AI is changing game development workflows.
The Claude Code pricing discussion is the sharpest thread. Members are dissecting what it means when an AI toolmaker tests price points on its own power users, and what that signals about the economics of agentic coding tools. The non-PM shipping debate cuts to a governance question every team with an AI-accelerated engineering org will face: when anyone can build and deploy, who owns the quality bar.
The full thread archive includes specific tactics members are using to set boundaries without killing momentum, real numbers from the Claude Code pricing observations, and concrete game studio examples where gen AI is replacing or augmenting specific production roles. The original is worth reading for the unfiltered practitioner arguments, not just the conclusions.
[READ ORIGINAL ā]