Linux org - target roughly 20 percent of Linux engineering spend on experiments; bring proposals Peter can say yes to; risk gets discussed, not ruled out

Situation

Opening the Reno LinuxEng/RAT onsite to the full org, Peter named the missing piece as a culture that experiments and is willing to fail. Answering the objection that a public mistake in enterprise Linux is fatal, he argued the current way of working will be disrupted eventually, and set a target: about 80 percent of time on what is winning, about 20 percent of Linux engineering dollars on experiments - this team or a dedicated hired team. He asked for concrete proposals (for example three engineers on a parallel approach) and set the risk norm: write down if we do X, Y happens, then discuss mitigating Y rather than refusing the conversation.

Reasoning

Confirmed by Peter. Earlier attitude-level pushes (fail visibly, prototyping culture) did not change behaviour because the CVE load consumed every hour, so this time he named a budget share instead of an attitude. The burden of design moves to the team: Peter funds, they propose. Public pristineness is less of an asset than it was - customers like Google now see enterprise distros as piles of vulnerabilities - so being outrun is the larger risk. The onsite was the one moment the whole org was in a room as CVE volume, agentic engineering and very large deals arrived together. Stated as a target he wants, not a locked budget line.

Additional Context

Greg followed with the Ginux/HDAE experiment as a worked example of disposable code and cheap failure. Someone noted QA is done by authors; Nathan framed the gating thought-experiment (what must be true to remove humans from review).

Observed Evidence

80/20 target stated to the whole org; whether you do it or we hire a team and lock them in a room, I do not care, but we are not doing it; nobody has come to me asking for three engineers - paint me those pictures; I do not want we cannot take any risk to win the day.

Confidence Breakdown

Reasoning Depth Analysis