Personal index / ongoing
Adam Hartley
I study medicine and build tools for difficult decisions.
Medical student at Trinity College Dublin, founder and hands-on builder. I am most interested in tools that help people make difficult decisions with better evidence. I also row with DUBC, play piano and sing.
At a glance
Right now
Studying medicine at Trinity College Dublin.
Developing rowing analysis tools through OARA.
Finding out whether TOP can forecast a real AI workflow before it runs.
01 / Current work
What I am working on now
Three entries, at three different levels of certainty.
Log 001
TOP
TOP is an experiment in forecasting the low, likely and high cost of a repeated AI workflow before it runs. A public tool already helps people understand selected AI usage exports locally in their browser. The pre-run forecaster exists internally, but its accuracy and usefulness are still being tested.[1]
- Working question
- Can one real workflow be forecast well enough to change a decision?
What is proven so far?
Historical usage analysis works locally in the browser. Reliable pre-run forecasting has not yet been proven.
Log 002
OARA
OARA is a set of evidence-grounded rowing analysis tools. OmniRace helps examine races and crews, while OmniErg makes erg data easier to read and act on. The goal is useful evidence for athletes and coaches, without asking the numbers to say more than they can.
- Practical test
- Does the analysis improve the next conversation, session or selection?
Log 003
New Age Solutions
New Age Solutions is the home I am building toward for work that earns its place through real use. It is an early direction, not a finished company story.
- Operating rule
- Let evidence make the structure necessary, not the other way around.
02 / How I work
A few operating rules
Click one. The short version is the title.
01Evidence before the claim
A working prototype and a reliable result are not the same thing. The sentence should change when the evidence does.
02Privacy by default
Sensitive data should stay local or tightly controlled unless there is a clear reason and permission to move it.
03Build, test, revise
I prefer a real trial and an honest result to a polished story with nothing underneath.