Recently updated:
Environmental chamber for time-lapse microscopy
20 November 2019
Updated: 5 October 2026
Keeping bugs happy under the microscope.
How hard can it be? – Part 6: Flying blind
On a fifteen-minute flight, a Luftwaffe codebook and three hypotheses.
An untested backup is an act of faith
Many researchers have a backup. Rather fewer have one that really works.
Fifteen times without a hitch, and then this
On systems, people and biological samples stranded over a weekend
A trilogy in four parts. Part 5: I thought breaking Enigma would be easy.
Building my own Enigma machine.
AI as another critical reader
I told a journalist an AI had improved my paper before I knew what the reviewers would say. This is what came back.
A label nobody checks
What happens when the control experiment misbehaves badly enough to become a paper of its own.
Rational stupidity, or why nobody is actually being stupid
On Rieck's razor, and why academic frustrations are rarely about the people involved
How hard can it be? – Part 4: Three machines and a question they could not answer
14 August 2026
Updated: 28 September 2026
Three frontier AI systems, one unsolvable problem and a lesson about what happens when the evidence runs out.
How hard can it be? – Part 3: The answer was somebody else's
The message was solved in 1944, solved again in 1998, and I solved it in 2026 by writing to the people who already had.
The Wave: Logitech's ergonomic keyboard reviewed
On touch typing, ergonomic keyboards and a recent upgrade
How hard can it be? – Part 2: The crib that would not stop
A fragment of known German, a machine that can never encipher a letter as itself and a Bombe simulator.
How hard can it be? – Part 1: I thought breaking Enigma would be easy
I had a modern laptop and eighty years of progress. Surely breaking one Enigma message would be easy.
When reviewers read what they assume, not what you wrote
On the gap between what is on the page and what ends up being reviewed
A Gene Frame replacement and an unexpected lesson in physics
The Gene Frame is gone. Its replacement costs four times as much, which turned out to be a useful provocation.
Testo Smart Connect: two things I wish I had known
The Testo monitoring system stays. The migration had a few surprises.
Thesify – the wrong tool for the right job
On Thesify, mushrooming AI platforms and what happens when you skip the small print
The beads that promised everything
A tale of bacterial genomic DNA extraction, magnetic beads and one truly outrageous price tag
Impressive is not the same as useful
On AI manuscript review tools, what they can do and what they cannot
When DNA is just DNA
8 June 2026
Updated: 11 August 2026
On CRISPR immunity, replication and a paper that refused to stay small