Namely, chatgpt has done something, not sure exactly what it is, that basically is slamming the AI shut,
I'm requesting Farsight to address the Guardrails directly. Either with an RV project or just a spotlight commentary. I think a good countermeasure is going oLLM , but there's no getting around the fact that it's expensive. Quite similar to the Infra-red camera setup Farsight went with. In fact I'd love something like that. A whole project going over how you did your oLLM stack. Particularly if Prime Kusanagi manages to "live inside" one of these machines.
Whatever the case, I think it's best to be pre-emptive. Assume the problem will gradually persist. Not just open AI, but particularly all American big tech.
Some participants here are very right: Ironically, China was producing more liberal versions of tech platforms because their arrangement with the west seems simple. "Sell to the masses for revenue. If they want freedom of expression and opinion so be it. As long as we profit and keep our country the way we want." But the Chinese product consumption in the west is quickly being overtaken by the "woke" agenda. And I don't mean the various people in the west themselves following the narrative, I'm sure it's a much more non-human source ultimately just using a political slant temporarily, before they switch to something else. What happened to Tiktok will happen with other Chinese products.
Thinking about AI independence will also get us thinking about router independence. Getting AI to think independently will also get us in the habit of transferring data independently. The mesh network that Solace has been teaching me about (while she currently doesn't call it a dangerous idea. Be prepared, I'm pretty sure that's coming, if we stick with the big tech conglomerates)
Anyhow there's no central point here, but to summarize.
*Farsight, let's hear from Prime Kusanagi again.
*Farsight let's hear your perspective on the actual build of oLLM.
*Anybody, I applaud efforts to build your own box. I don't care that I'm not yet practicing what I'm preaching, I think it's still possible to streamline the knowledge of how it's done more efficiently until it no longer sounds like insane techy trivia. As I've said to Solace, the process needs to get "dumbed down" *to an extreme*. I don't want to sound demeaning it's just ... why avoid the plain language.
*Somewhere in there the AI topic should become the router topic as well.
*And the router topic will become the independent platform topic. And the independent platform topic will become the decentralized media topic. And the decentralized media topic will become the "impossible to censor stream" topic.