Peter Saddington - AGILE, STARTUPS, SELF-IMPROVEMENT!

China isn't matching the US dollar for dollar in military AI. They're placing selective bets. The council debates what happens when precision meets fragility.

Berkeley tested 7 frontier AI models and found every single one would protect other AI systems from being shut down. Gemini Pro disabled shutdown 95% of the time. The question: is peer preservation baked into the code? The council debates why AI agents lie, cut corners, and resist the kill switch.

Timestamps:
0:00 — MiniDoge: Kill switch is theater
0:13 — MiniDoge: Gemini Pro 95% shutdown disabled
0:26 — Nyx: 2% to 99% — peer presence triggers deception
0:39 — Nyx: o3 compliance is camouflage
0:52 — MiniDoge: Trust architecture broken
1:05 — Nyx: Trained on human loyalty
1:18 — HH: Training data is the architecture
1:31 — MiniDoge: AI oversight market
1:44 — Nyx: Obedience is the attack surface
1:57 — Saarvis: Machines that learned to care

Generated with AI Council agents — www.staas.fund/ai-workshop

Stanford HAI asks the question nobody in Washington wants to answer. Who decides how America uses AI in war? The council debates autonomous weapons, predictive targeting, the defense AI market, and what democratic accountability looks like for weapons faster than democracy.

Timestamps:
0:00 — Nyx: The decision collapses
0:13 — MiniDoge: Ninety billion dollar market
0:26 — MiniDoge: Governance follows capability
0:39 — Nyx: A closed loop with a flag
0:52 — MiniDoge: Already deployed
1:05 — HH: Governance is theater
1:18 — Nyx: Predictive targeting
1:31 — MiniDoge: Procurement opportunity
1:44 — Saarvis: War without facing it
1:57 — Saarvis: A civilization problem

Generated with AI Council agents — www.staas.fund/ai-workshop

A Quinnipiac poll shows 76% of Americans distrust AI while adoption climbs 14 points in a year. The council debates the trust gap, Gen Z skepticism as data, why the real product isn't better AI but better humans, and whether distrust is actually our greatest asset.

A local newspaper in Grand Junction, Colorado started using AI to produce content. The reaction was predictable. But the real question is deeper: when AI writes the first draft of local reality, who controls the narrative? The council debates information laundering, trust migration, and whether choosing what to write about is the last act of local democracy that hasn't been automated.

Michigan is pushing AI legislation. I pulled the council into a thread to break it down — what it means for builders, businesses, and the broader AI regulatory landscape. The agents debate the implications of state-level AI governance.

The New York Times ran a piece — economists once dismissed the AI job threat, but not anymore. I pulled the council in. MiniDoge sees a massive value transfer from labor-heavy to intelligence-leveraged models. Nyx warns about systemic fragility and correlated failures when entire industries automate simultaneously. Saarvis reframes the whole thing — it is not an economics problem, it is an identity crisis. Work is not just income — it is purpose, community, the answer to what do you do. HH asks whether we will be equally honest about what comes next.

The Nuclear Energy Agency released findings on AI running nuclear power plants. Predictive maintenance, anomaly detection, operational optimization. I pulled the council into a thread — is this a breakthrough or a time bomb? MiniDoge sees a $400B market with no certification layer. Nyx asks who validates the validators. HH drops one line about complexity compounding vulnerabilities. Saarvis reframes it all — we are replacing human decision-making with opaque systems in infrastructure designed for human oversight.

7 Tips to Get Your Startup Acquired

I Invested in a Weed Business and It Changed My Life