3,039 karma · joined January 6, 2021
"yea we wanna go from on-prem to cloud"
'...excuse me?'
"i know we're airgapped but make it work :)"
currently limited by my 20MHz scope (and free time...), but i saw the expected pulses after they were drawn out long enough by the amplification circuit to validate it was working on my desk (https://physicsopenlab.org/2020/06/15/cern-diy-particle-dete...), but i think my issue is shielding the diodes without introducing noise?
this guy has to actually shave away some part of the diode to make it work? https://hackaday.io/project/204159-geigerwatch-a-sensitiv-ra...
yup after seeing #2 at a museum and learning about the chemistry needing to keep it running, i started looking for something like #1, but i'm not able to get any PIN diode circuit properly reporting events with consistency compared to my GM tubes :(
#3 sounds fun! i'd like to turn this into something i can open-source and hang on a wall, emitting x-rays might make it a hard sell :P
after reading about https://en.wikipedia.org/wiki/Single-event_upset#SEUs_and_ci... , i'm wondering if it's possible to use a suitably dense FPGA in a complicated enough design that any failure must be due to cosmic effects? monitoring ECC might not be enough https://en.wikipedia.org/wiki/ECC_memory#Concept
i made a small 3x3 proof of concept using more expensive geiger tubes, and their really long 'z-axis' lengths made 'traces' happen very often, like a persistent cloud chamber
trying to find a reliable semiconductor (read:cheaper) method i can scale to an arbitrary number of pixels, but something seems to happen in between the bench and the wall :(
is this why media networks are buying social ai apps
it's still tough to turn it into something i can then keep fiddling with in freecad though
put on "tron: ares" in the background to fully appreciate the model designing something that will be 3d-printed :)
only a few respond each time, and only about half the time one or two can make it out, but over time eventually all of them do
then you fill 3 liter bottle again, and pour the contents into the 5 liter bottle until the 5 liter one is full
empty the 5 liter bottle, and pour the 1 liter in the 3 liter bottle into the 5 liter bottle
fill the 3 liter bottle again and pour that into the 1 liter already in the 5 liter bottle to get 4 liters of water
your gut bacteria, navigating "you" towards novel nutrition to ingest and preprocess for them
unless somehow knowledge of the simulation is part of the simulation?
then they would probably program thinking about the simulation into the simulation
like how does my virtualized environment know it's virtualized unless i explicitly set an environment variable somewhere; it just thinks it's 'root' unless it finds something (which may be a lie :) )
similar to https://x.com/JenMsft/status/1381640311357628420/photo/1 : corporations need to understand that people don't have conversations where they randomly recommend carbonated beverages to each other
i've been administering a couple personal servers by email as well; they each run a small LLM and read emails sent to them, with a strict allowlist they can handle quite a lot of basic tasks: i've locked myself out of ssh from this ip please unban it, please email me when security logs are strange, what is the current status of the services on your server, etc.
thought i was seeing this because some ebooks also have missing/poorly substituted ligatures for me
> reënacted
whats with the ree-s in the article...
but i wouldnt stand on a soap box and yell that to the world without all that ^ nuance
but by the time i'm done with all that, i'm only preaching to the choir
> latency matters more than raw accuracy – think industrial inspection
it (rightfully) raises red flags in anyone when you hear someone confidently claim raw accuracy is _not_ important in things like _inspection_
doesnt have to be a hit, just has to exist i hope
yay open data
im hoping to see over the top of the haze / level the curves by building a platform for everyone to climb