Went on a troubleshooting (dokuwimi xmlrpc issues) binge with Gemini Flash 2.0 Thinking Experimental. Absolutely amazing and fun. Most fun I've had with a. 'Problem' in computing for ages!

Holy Hell langflow Q&A is insanely easy! JustWorks tm

anybody played with n8n, I'm hoping vectorize some mardown and us AI to chat with my docs

15 days has passed, where is the JFK files?

🍓🍓🍓🍓🍓🍓🍓🍓🍓🍓🍓🍓🍓🍓🍓🍓🍓

Wow I've been sleeping on notebooklm

I feel like this Ai shit is gonna go crazy, and I've been ignoring a huge tidal wave that is coming crashing in. I'm struggling to make it do really useful things for me. At this point it doesn't seem like I will likely find a use for it in ways that have momentary value vs cost analysis. I think I should just dive in, butt I think that wave is gonna hit before I can get ahead of it.

I'm watching deepseek-r1 8b <think> and I want to yell at it because it is insanely stupid. Not even paying attention to the rules of my prompt. It's pitiful. LMstudio et all needs a way to 👎 and scold these models, stopping it is not harsh enough. Feck

I feel like these models are hyped to be PHD level, but in reality they perform like elementary school students. Obviously I'm not using the large parameter versions, but damn those versions are not in anyway usable for anything real. Thus the closed cloud models are still king unless you can afford to run the full 100b+ parameter models at home. I need more powa!

In my experiments, I am painfully frustrated with the hardware on hand for AI. Ollama, LMstudio, Jan, gpt4all. Let's say 30 tokens max with 8b. Tried out some basic local docs stuff with a single PDF/CSV, most 8b models can even read a basic spreadsheet without failing miserably

Its like you buttcoin maxtards dont even know fundementals of technology. BITCOIN is not magic, there really isnt much to it that changes fuck all. It will be used against you before it saves the world. FUCKING get it through you head, BITCOIN is a CULT

Yeah I wanna to get into AI, but god damn these GPU prices are insane

Lets talk about Suchir Balaji

Bwhahaha chatGPT Gov

Why my damn internet so bloody slow right now

Is pip always this damn slow when downloading packages

AI and quantum computing is far more important than crypto, fight me

Enough said

I love the idea of running AI models locally, but it seems like ultimately, scale is a huge issue, the larger the parameters/larger the model the more hardware you have to throw at it. At home, obviously you probably aren't gonna have more than one GPU. Certainly not 100s of thousands of gpus like the big boys have in their datacenters