Beyond the Rating
Texas school ratings need context. Look up a campus or district, compare it with schools serving similar students, and see how the overall grade is built.
Look closer.
Make it useful.
FIELD LABQuestions → Evidence
→ Something you can use
→ Something you can use
Tools for making sense of education data,
built around real questions.
01 / SELECTED WORK
Context changes the picture.
A school rating. A funding figure. A writing score.
Each becomes more useful when you can see what’s behind it.
4 projects
Texas TIA Explorer
Find nearby approved Teacher Incentive Allotment systems. Search by ZIP code or district and compare public designation and allotment data.
ECR Score Explorer
Make the differences between writing scores visible. Compare anonymized student responses, examine writing moves, and connect what you notice to what to teach next.
ECR Outside In
Eight statewide essay-score distributions, read from the edge inward. Explore a grade, inspect the numbers behind the rings, and export a view to share.
02 / FIELD NOTES
The thinking
along the way.
Questions from the work, decisions worth explaining, and things still taking shape.
The Bottleneck Was Never the Prompt
Cassie Kozyrkov argues that judgment is the last skill standing. I think she's right, and I'd add one thing she leaves on the table.
· 5 min read
Read the note →03 / RECENT BOOK
Paperback · 2026 · ISBN 9798189673832
PUBLISHED WORK
Prompting is the floor. This book is about the ceiling.
Moving Beyond Prompting
How to Stop Asking AI Questions and Start Assigning It Work
In August 2023, I asked an AI for twenty dinner ideas my picky eaters might tolerate. Three years later, I was reviewing overnight reports from an AI agent running on a secondhand server in my closet.
So I exported every conversation I had ever had with an AI—16,366 prompts across ChatGPT, Claude, Gemini, and an agent of my own—and analyzed them to find out what had actually changed.
It wasn’t better prompting. What changed was everything around the prompt: what I handed the machine, what I stopped asking it, and what I started assigning it instead. The book follows that progression through four stages.
- 01Prompting
- 02Grounding
- 03Delegating
- 04Orchestrating
View the book on Amazon Author’s store link · Not a paid link
04 / WHAT I’M READINGREADING NOW · UPDATED 2026-09-08
Notes from the margins
The Thinking Machine
A running set of notes on Jensen Huang, Nvidia, and the operating ideas behind the company’s rise. I’m most interested in the moments where technology changes the shape of a market—and where strong judgment still matters.
“Parallel computing divides a big problem into smaller pieces, then solves them all at once.”p. 96
View the book on Amazon Paid link
As an Amazon Associate I earn from qualifying purchases.
THREE IDEAS I’M CARRYING FORWARD
Changing who the product is for.
CUDA begins as a reclassification move: gaming hardware becomes scientific infrastructure. The important shift is not only better chips; it is making the same capability useful to people with a different class of problem.
Finding the speed of light.
Start with the theoretical minimum time if everything went perfectly, then work backward toward reality. The idea separates physical limits from delay created by queues, priorities, and bureaucracy.
Capability without comprehension.
Neural networks can learn useful behavior and outperform people in narrow domains while remaining opaque—even to their creators. Performance is evidence of capability, but not necessarily understanding.
Updated periodically from my reading log. This entry draws from notes recorded August 30–September 8, 2026.
05 / BEHIND THE WORK
Better questions.
Useful tools.
Room for judgment.
I’m Micah Taylor, an educator in North Texas. My work is in assessment and accountability: asking whether something is working and bringing evidence to the answer.
These projects apply that habit to public data, interactive tools, and questions worth a closer look. The work comes first here. You can explore it, inspect its sources, and decide what is useful in your own context.
Show the evidence.Keep the context.Leave room for questions.