Working on DNS, trying to avoid adding a microsecond to processing a packet or enlarging a packet by a few bytes. AI: we need our own (nuclear) power station to be able to train and run this model. (I know a LLM can process many queries, still the difference in attitude is staggering)
Reminds me of the recent post by 
@dabeazDavid Beazley who has been using ed(1) for a rather beastly coding project and reports that, despite the quantity of time spent coding,
> ed…has logged a total of 3.25 seconds of CPU time in this endeavor.
https://mastodon.social/@dabeaz/115820592207784755
svelte coding is becoming a lost art. 😞
If you have a fediverse account, you can quote this note from your own instance. Search https://mastodon.bsd.cafe/users/gumnos/statuses/115826293479009752 on your instance and quote it. (Note that quoting is not supported in Mastodon.)