By Erica C. Barnett
This week on Seattle Nice, we discussed tensions between Mayor Katie Wilson’s office and members of the City Council, whose frustration with a lack of collaboration between the second and seventh floors of city hall erupted last week when a Wilson staffer asked the council to hold off on passing a bill to implement the final part of Wilson’s shelter surge plan.
As I reported, Wilson was apparently unhappy with some of the amendments councilmembers proposed and wanted the council to change them. The council—already irritated that Wilson sent them the shelter bills without first securing a council sponsor and trying to elicit support—was not pleased that the mayor seemed to be ordering them around, and after a reportedly heated meeting between countil members and three Wilson staffers, the council passed the legislation, which Wilson had asked to be expedited as an “emergency” bill, with the (relatively minor) amendments intact.
The tension, Sandeep pointed out, has been brewing since well before the latest conflict; when Wilson fired former City Light director Dawn Lindell, some councilmembers were sensitive to union complaints and excoriated the mayor for what they called a rash decision. Just yesterday, Councilmember Bob Kettle took up that torch again during a discussion about a proposed one-year moratorium on data centers, saying, “We had top notch leadership with Seattle City Light, and this is a failure of our city right now.”
And speaking of data centers, our second segment is all about whether saying no to companies that want to build massive data centers to power AI is a good idea.
Sandeep argued that if Seattle doesn’t embrace the AI future, we may fall behind economically and turn into a hollowed-out shell of a city, like Rust Belt cities did in the 1980s. David some economists claim AI could help solve the affordable housing crisis and doesn’t want to dismiss possibilities like that out of hand. And I, as the resident Luddite, argued that we shouldn’t hitch our entire economy (and the future of our climate) to technology most people don’t like or want.
FYI: Seattle Nice Patreon donors got an early preview of our show this week. Supporting Seattle Nice gets you access to some of our episodes a day before they go out on the regular feed, along with occasional Patreon-only exclusives and the knowledge that your contributions go directly toward making Seattle Nice for you every week, including paying our editor, Quinn Waller.


I’m glad Erica is on the Ed Zitron train. Funny to hear the number-go-up argument from the boys.
The problem is not AI per se – AI could be incredibly useful and beneficial if trained for the proper purpose. The problem is how these super-data centers are conceived and designed. They probably should use advanced AI to design these data centers so that they do NOT increase the utility rates of working-class people. It requires some creativity, but I think all of the elements are there. Need to use solar-wind or even underwater cooling instead of the grid as a primary source of energy. Installation of small (nuclear) modular reactors (SMDs) to power these centers should also be explored if they are deemed safe.
And if the grid is used, need to have the data centers pay their fair share through a Minimum Billing Demand (MBD) tariff to force these facilities to pay for 100% of its contracted capacity every month which prevents utilities from recovering costs from residential users if the data center under-delivers. Instead of a massive new connection to the grid, these facilities could use the grid’s “last mile” of unused capacity that is already built and paid for, which would turn the grid into a cost-saving advantage. Active and dynamic load management of the grid is a must, including the use of battery energy storage systems (BESS) to do “peak shaving” that draws energy from the the batteries instead of the grid during expensive peak hours (similar to what is used in solar powered grids).
But just hooking up these large data centers to the grid and hoping for the best is a NON-STARTER.