An index of open questions
status
Is the Socratic method the best method with LLMs for getting what you want?
I guess so
Whenever you have some issues with incorrect answers from your LLMs, try the Socratic method! Try asking it questions that direct it to the right path, without actually hinting at the answer (which you yourself might also not know). As an example, if you are trying to debug a nasty bug in production, try asking it questions that it can investigate and/or test the answer for. It will almost always result in better outcomes, just like how Socrates did! I guess the biggest invention of humanity is still a footnote to Socrates :P
Is Claudese intentional?
yes
If you are able to somehow see the thinking traces (PS: omp), they are much more readable!
Can AI play NetHack?
surely
Once you set up the correct harness, it can. The model was never the missing piece — the scaffolding around it was.
Is Perl still relevant?
apparently
My AI agent just reached for Perl to do a lookup — unprompted, out of a whole shell’s worth of alternatives. So yes, apparently it is still relevant.
When will AI take over?
pretty soon
What will we do once it happens?
total chaos
Is there a techapocalypse waiting for us in the near future?
absolutely
How much environment can I fit into a couple of nodes?
as much as you need
How many minutes of songs can I fit into a single year?
more than I admit
Is everyone connected, no matter where you go?
YES
What is everyone's career plan?
no one knows