t0yv0
t0yv0@vmst.io
<p>Gardener that programs. Dabbler in computing esoterica. FP in the past (F#, Haskell, Scala), Go/Python more recently, these days diving into ASIC design software in every language in that world including TCL.</p>
Posts
-
Post #1828302
https://lnkd.in/eir8_tMc I didn't know light can do this. Here we come, microring modulators.
-
Post #1592621
RE: https://infosec.exchange/@ubernostrum/116374532775532247 This was a great read. Thanks for writing this down. Actually thoughtful discourse on LLM use in programming.
-
Post #1082428
More fun-facts from the silicon world (TIL): PFETs and NFETs are in theory mirror images of each other because p-holes can be seen as carriers of current, instead of electrons themselves; however! in practice NFETs have higher transconductance because electrons have greater mobility than p-holes in silicon so NFETs are preferred for most designs.
-
Post #903609
Work today. Referencing Lethal Trifecta helped design a better clean-room environment. An honest checklist of defenses. Thank you https://simonwillison.net for providing a vocabulary and mental reference point for agentic security
-
Post #873941
til mmdr is a drop-in replacement for mmdc mermaid-cli compiler that does not require chromium and therefore removes a huge headache in my nix-darwin build tree
-
Post #829589
A bit of a drive from Virginia but now here we are in Saint Louis.
-
Post #741704
Another day trying to synthesize working legacy software integrations with agents. Good to have access to domain experts that can call the agent's bluff. I was not necessarily able to call it. Still not quite sure I got the intuition right on what's a circuit schematic vs what is a netlist and what is the crucial difference. But it will click soon I hope.
-
Post #689465
Two tools at work are scriptable in TCL. This is going to be a fun computing history trip.
-
Post #607048
Design Of Analog Cmos Integrated Circuit , 2Nd Edition has arrived in the mail! Here we go.
-
Post #581550
Jupyterling - a little jupyter - born out of frustrations explaining the kernel state when teaching programming to my daughter - you can try it live https://t0yv0.github.io/jupyterling/ -- longer note in https://t0yv0.github.io/2026/03/20260307-jupyterling.html
-
Post #530690
https://leodemoura.github.io/blog/2026/02/28/when-ai-writes-the-worlds-software.html what I keep thinking about Claude translating zlib to Lean and proving it correct is that the Lean translation was proved correct but the C library is in the one deployed. Subtle but important. System software sometimes has perf edges that are not captured by the proof model, or scientific software has convergence edges, so a rewrite is risky to deploy as it will break something even if it is mathematically more...
-
Post #527978
@thumphries@hachyderm.io we are entering strange times for work ethics when routine intellectual work is largely automated... Novel results won't be but that was always hard and too unpredictable to sustain 9-6 commercial jobs. So as regards these jobs for AI operators, is it work or grift? Is human judgement really a nice and valuable touch or that's going to be automated also for 99%? What do we do next and what to do with our identity? Fun times. Good to be kind while we sort out new...