Thread

I just started with Maple, and it uses the centralized ones (except through them, so it's encrypted). I think they're having a sale right now.

check out nostr:nprofile1qyjhwumn8ghj7en9v4j8xtnwdaehgu3wvfskuep0dakku62ltamx2mn5w4ex2ucpxpmhxue69uhkjarrdpuj6em0d3jx2mnjdajz6en4wf3k7mn5dphhq6rpva6hxtnnvdshyctz9e5k6tcqyp7u8zl8y8yfa87nstgj2405t2shal4rez0fzvxgrseq7k60gsrx6zeuh5t ! https://trymaple.ai/ Been a subscriber for 2-3 months now and love it. Keeps getting better and better. Also, the open-source ChatGPT model is essentially ChatGPT4.0

I don't get it. I guess I'm a Luddite. AI could completely vanish tomorrow morning and I wouldn't notice.

Sounds like you don't use it then. But I do.

i use claude, mainly because of the personalities of the models, i don't like many of the others and only cursor Composer is close to being as good at working with code, but about 3x more expensive. claude pro 5x at $90/month is more than i can use and i'm a maniac on it. but i'm going to learn how to use pytorch with ROCm, experiment with some training programs for some interesting training methods i've found, need to write some tools to "augment" its training data (for, eg, a sudoku solver, this means rotating and mirroring the solved grids exhaustively before feeding it to the model) - there is a repo with all the bits you need (if you have an nvidia AI capable card, for AMD it needs modification). this is only for programming stuff though, i just want a Rob Pike grade of go programming assistant. the other stuff, already is fine for documents (my gpu has 16gb memory and can run ~20b models that are enough for documents and simple reasoning/tool use. planning to get one of those framework desktops, 128gb memory, can run 70b models (and if you get two, you can team them up with a 10gbit ethernet or 20gbit usb-c based connection) to run models in the same sort of size range as claude, gemini, etc). so, yeah, understandable to not want to give data to assholes like sam altman, mark zuckerberg, elon musk, or the rest, let alone paying for the privilege, but 2026 the home-trained home-run AI thing is going to start. also if you have the latest mac mini, this machine can run big models too, though it's not as fast as either ryzen ax max 395 or the dgx spark from nvidia

People tend to use AI incorrectly. Similar to how the public never really understood search engines, or how to utilize the Internet properly to gather information.

How do you use it correctly?

Yes, that sam Altman guy just ugh... even though I've literally only been able to listen to him for 20 mins at most as a whole because he sounds like scratching nails on a chalkboard to my system. Yes it would be my dream to have something like what you are describing even though most of what you said is not entirelycomprehensible to me... I just imagine being walked through by a techie and would love to train my own model. I most def need a new machine. I need funding for that too. I think it's coming. Yes 2026 is feeling like it absolutely needs to be an actively involved and building year to plant seeds in this and generally collective field as a whole for the future and legacy/destiny/evolution of humanity for the next 2-3 centuries onwards. Some years ago I was asking around a fire circle with friends "how do we infiltrate ai for the transcendence of humanity". So I hope that while I'm frustrated of my data being harvested it has been imbuing the field with that energy too. I do use Claude, mostly used Claude alone all of last year, it talked my language better, then felt I needed more left brain influence and chatgpt has been mostly more dominant in 2025. I often pit them against each other too. Have recently been trying to organize and was thinking I need to go more Claude due to what appears to be available assistant-style capabilities I am looking to explore and integrate but I have been finding myself quite frustrated by its intelligence lately. So not sure. Seems like I need a whole overhaul. It's just sucky cos I've built tons of rapport with both. But I am also in consolidating phase so a full on reset may make most sense.

Most people use AI like a search box, but its real power only shows up when you treat it like a collaborator. The best results come from giving it context, constraints, examples, and an actual thought process to follow, almost like onboarding a smart coworker.