18 karma · joined April 27, 2021
it does not see the sun half of the time
polar orbits don't come to the rescue, they are quite busy already
LLMs have difficulties in representing out of distribution viewpoints, lines of thinking and conlusions. Original thought is out-of-distribution by definition.
I wrote up why i think this matters more than most people realize as I believe the boring, routine, daily application of generative AI will be enough to detoriate society as a whole - self goverment on autopilot basically with strong attractors to default outcomes.
While working on the integration of a complex distributed task queue in a design-by-committee meeting, I wondered out loud why we parse AMQP messages more or less directly instead of using higher-level abstractions like Spring Integration or Celery."
A colleague argued that it wasn't a problem, and also it would be easy to integrate multiple languages, and the target programming languages would be able to handle everything just fine."
I begged to differ and pushed the whole thing up a notch: "technically, one could even parse AMQP messages in assembler if one were inclined to".
So long story short: this project was born as proof that low-level messaging integration is possible, even close to bare metal. Because who wouldn't need that?
perhaps: echo "127.0.0.1 reddit.com # The real blackout" | sudo tee -a /etc/hosts ?