Impact Lab · Track 2
The decisions above your job
Whether people can afford to work here is decided in rooms most of them will never see — and by the time a decision is visible, it is already a bill.
Someone wrote
Same room, same cards. These are the ones that reached past the job and asked who decides.
“Who makes the calls?”
“Who owns the data?”
“The information + access gap will widen. People will lose work before getting enough access to change or save their careers.”
Written on cards at a Claude Conversation. Unsigned, and quoted exactly.
What people said
The July room was asked what AI means for their job. It answered about governance, ownership, privacy and access — who decides, who owns the output, and who gets in early enough for it to matter.
Underneath that is a harder question this track exists to ask: whether it is still possible to build a life in this city on this kind of work. Costs outrunning wages. Public money spent without a say. Retraining aimed at jobs that no longer exist. Rules written for a previous decade.
The hope was small and concrete: to see it coming, and to know where to push.
Second paragraph is Travis's framing, not the room's. None of the 28 July cards mention rent, wages or the cost of living. Kept because it is the track's premise; marked because the cards above are verbatim and this is not.
The opening
This track is the same question about work, asked of the institutions that shape it. Directions worth pulling on.
Early warning, not audit. Plenty of people examine spending afterward. Almost nobody makes the decision legible while it is still a decision.
Public money against public outcomes. Workforce training, industry incentives, development subsidies — a lot is spent on employment, with little visible about whether it worked.
The cost of staying. Housing, transit, childcare, permitting. What it takes to keep working here, and what is moving.
Rules written for a previous economy. Licensing, classification, benefits eligibility — much of it predates how people actually work now.
One question, many jurisdictions. Dozens of cities in this county, no common format. Nobody can currently ask them all the same thing.
Data to consider
City and county open data portals, which have real APIs. Budget and audit publications, often better than people assume. Legislative and council records, which is where decisions live before they are decisions. Lobbying and contract disclosures. Regional cost-of-living and housing series. Workforce program reporting. Federal spending flows and nonprofit filings.
Most of this is technically public and practically unreadable. That gap is the opportunity — you are not unlocking data, you are making it legible.
Don’t
Don’t build a corruption detector. Unusual spending patterns have boring explanations, and a tool that implies wrongdoing will eventually be wrong about a real named person.
Don’t do sentiment analysis on meeting transcripts. It is the reflex and it produces nothing anyone can act on.
Don’t drift into general government accountability. The test: does this connect back to somebody’s ability to work and live here? If not, it belongs in a different room.
Know what you are taking on
This track has less on-ramp than the others. Scope hard: one decision, one program, one question.
Pick one decision that affects whether people can afford to work here.
A budget line, a program, a rule, a fee. One.
Answer four things before you build: who decided it? When? In what venue? What was the alternative?
If you cannot answer all four, you have picked the wrong thing — or you have found the actual problem, which is that nobody can answer them.
Then: what would have told an affected person this was coming, early enough to do something about it?
Bring this to a lab
Impact Labs are one day, in a room, on a real problem. The next one is on the calendar.