Back then I had done the math on how long it takes to turn around a commercial chicken coop after a complete full. The timeline from egg to fully producing hen was quicker than I thought, and it seemed they should have been able to keep up the replacement rates.
Thanks, it does. I've been playing since about 2000, came from Cosmic Rift / Infantry when they went non-free.
We're lucky we got Priit to release the encryption/security module so Continuum clients could connect to ASSS servers without the security warning. I doubt it'll ever be updated, someone will have to take up the mantle.
I play SubSpace (a MMO spaceship game released in the 90s) to this day. It was shut down soon after release.
The original server binaries were left on the original CDROM by a programmer.
Then PriitK, a creator of Kazaa and then Skype and Joost!, went on to re-create the client due to cheating/hacking, naming it Continuum.
Years later the server is reimplemented as A Small Subspace Server (ASSS), making it a complete fan remake of the original game (sans graphics). This is also when we finally got server side mods, everything before that was client only or a hack.
Ive had that issue with touchscreens before and was able to disable the driver for it. Linux so not sure if possible or how on a chromebook but an option.
A small city near me in the suburbs of Detroit just had to have a town meeting / facebook post / etc about teens driving electric scooters and bikes driving recklessly, causing accidents, injuring themselves and others, etc.
The system prompt will always match in the prefix cache. I just meant it could be prefilled before any user queries on completely different hardware. Then you are only dealing with the n^2 only for the actual user prompt. We're in agreeance I think.
The state of the system can be cached after the system prompt is calculated and all new chats start from that state. O(n^2) is not great but apparently its fine at these context lengths and I'm sure this is a factor in their minimum prompt cost. Advances like grouped query or multi head attention or sparse attention will eventually get rid of that exponential, hopefully.