language model server. just language model. On Fri, Dec 8, 2023 at 08:28 Undescribed Horrific Abuse, One Victim & Survivor of Many <[email protected]> wrote:
> problem: metamodels can make worse dissociation. because we are new enough > to see them as more powerful maybe could be wrong > > On Fri, Dec 8, 2023 at 08:26 Undescribed Horrific Abuse, One Victim & > Survivor of Many <[email protected]> wrote: > >> apologies for confusing us >> >> 0819 >> >> there was some question around coding more effectively than >> hunt-and-peck-on-mobile-by-hand >> >> one runs into two issues, inhibitions regarding different kinds of >> metatooling, and general life difficulties (breaking devices, poverty) >> >> noting we have another computer. putting code on this computer would be >> helpful. we might also make plans for ai assistants or helpful ai work (not >> certain here), and split time between tasks and these. >> >> the other computer may not run either, unsure. >> >> this is also just a side task. >> >> 0821 >> >> frarg more inhibitions. we’re taking on the goal holding now of more >> effective behavior. >> >> one idea was using openai like business people do. it doesn’t seem that >> useful. what about paying for a machine learning server, like on akash >> (warning, triggery property) or vastai >> >> with vastai it is possible (and a little difficult) (and also this can be >> done with akash) to spin up a machine learning server and run something for >> just a [minutes] >> >> 0823 >> >> we have idea: we could work on making an inference backend (a simple one) >> that runs on akash or vastai without regard to which one (missing >> vocabulary: wondering whether there is a god, shape near guessing >> influencers are engaging in religious influence but that eventually >> contract expires from amount of funds paid) >> >> anyway maybe work on server script that shuts down server if no api >> requests made by launcher after delay, (maybe twice launch time). means >> interface with backend implementation >> >> 0826 >> >
