R&D

every public record for the twin cities, joined into one warehouse. pick a dataset, filter it to you, download the file. free, no account

what governments keep asking for that nobody makes anymore: 6,904 asks from 208 public agencies, each linked to the agency's own filing, with where to buy it, the service manual and who can still service it

every building minneapolis has okayed for demolition, on one map: the address, who is tearing it down, the year it was built and what is likely inside. for scrappers and junk haulers

say what you sell and get the minneapolis businesses that just hit a change: a build-out, a new liquor licence, a first health inspection, a teardown, a layoff notice. each with why now and who to ask. free, csv download

a minnesota address in, a printable draft sba records search with risk assessment out: the government lists the federal rule calls for, at the right distances, history back past 1940, and a suggested risk call. for an environmental professional to review and sign

a nursing aide talks through the shift instead of charting with both hands full

writes family caregivers the words to say instead of correcting a loved one’s reality

the same day planned twice, for seven kinds of small fleet. watch the usual plan draw itself, then par draw over it, priced with published wages, diesel and rates

an mcp server that knows where the public records are, and how each one lies

a lightweight model running on a raspberry pi

voice assistant for a raspberry pi: wake word, whisper, claude, piper

a local agent that answers questions about your screen activity

counting packages of varying types from one camera

gpt-4v joined to instagram

all of the literature research in one dashboard: how a novel's mood moves, how genres mutate, who influenced whom, whose prose sounds alike. filter it, follow a story, or ask it a question

one small gpt each for shakespeare, melville and mccarthy, trained from scratch on that author alone, then opened and measured

the rate at which fiction’s genres form and mutate, modeled as a growing network

who read whom, drawn as a graph

literature-mutations, phase one

subtract the words everybody uses; what’s left is a graph of literary likeness

swap only the adjectives in a famous line and hear how much the writer’s words do

sentiment paragraph by paragraph across a novel, and what the null result said