Note: An LLM was not used in writing this microblog entry.
This entry is part of a microblog series called LLMs
I experimented a little more with the /infill endpoint on llama.cpp for simple code filling with a local LLM. It works as expected. I’m not a prepper (about the bubble bursting or so), but I believe it’s worth having first-hand manual knowledge in the stack, to be informed on the technical side. I am regularly using Gemini to lookup things that I used to use google and Wikipedia for. I am debating setting up a closed, local instance of this.
A topic on my mind a lot recently is the Spectacle, in particular, the mediation aspect of mediums (previously raised on this diary under McLuhan topic), which is summarised well here:
experiences, or the first-hand fulfillment of authentic desires, to individual expression by proxy through the exchange or consumption of commodities, or passive second-hand alienation, inflicted significant and far-reaching damage to the quality of human life for both individuals and society.
“AI” (generative LLM as of this year), is often replacing authentic interactions with people (writing reports, sending automated messages, meeting notes, “pick a gift for my nephew”, generating code, opening legal cases, etc.), work colleagues ask their AI of choice rather than their colleagues, for help or brainstorming. I read somewhere, though I can’t recall at the moment, the joke that some are using LLMs to “eat ice cream for them.” Funny summation.
Other reading: On not becoming a cyborg