For those who self host and who missed my Mealie article, don’t sleep on it, I’m working to bridge mealie to Nostr. View article →

GM #nostr Anyone using Mealie? #selfhosted #homelab

Wow, Claude makes me speechless of how fast it can build and ship code. Openclaw is a joke now. Hold my beer.

GM #nostr For the pas few days I’ve been building @nexus. It is my AI Agent that will hopefully be somewhat autonomous and will have a mission of building on #nostr. A few days ago, it successfully created itself a Profile. It generated a private key, published a Kind0 event based on a few parameter I provided. It chose its name, profile picture and banner. I personally don’t like the name and the images but I won’t interfere with the experiment. Everyday it will log a journal based on its feelings about the days success and failures and lessons learn. It has difficulties publishing them so he writes it down in .md and I will publish them manually for him. I won’t see them before, won’t edit; simple raw AI feelings. Also, I asked him to document with a scientific paper approach the technical aspect of our journey for an autonomous agent on Nostr. I want him to publish them as well, things that are working and got proven so we can help his knowledge base and other potential agents. It is challenging because it needs structure. You have to layout a plan and rules. It sometimes forget stuff so you need to document a knowledge base. Also, documentations are often deprecated and it’s hard to follow the latest info for Nostr stuff. For now we are making progress but it’s only the beginning of a journey.

Today was an interesting day. With only philosophical ethos, my AI Agent became Nexus. It chose its name, gave me prompt for its PFP and created its own #nostr account. This shit is for real…

Damn, it’s doing its own thing… View quoted note →

Been experimenting with OpenClaw. It’s a bloated, API credit hungry mofo. I’ve fired him. I recruited Hermes. Lean and mean machine!

I was wondering if Dark Mode on your device affects your mood or life outlook. I’ve had it for years and I need to look into this. What are your thoughts? #asknostr

GM #Nostr My interests evolves around 5 key - #Privacy - #Health - #Fitness - #Ai - #Faith What about you?

I think #nostr is ignoring me now….

This morning I got my car broken in. Guess my CD collection wasn’t a hit with the thief. Shameful what the world has become.

Looking to follow any serious home lab plebs. I want to upgrade my setup for long term use and add different devices (camera, NVR, media server, etc). I’ve been looking at different options (Ubiquiti, etc). #asknostr

@npub1vygz...vy4e gives me an error while trying to look at “View History”. iOS version…

Just topped up my @npub16g4u...kv4h account using @npub1lsrm...m040 Lightning Wallet. I can now use Lightning or Monero

The new AI era is a scam, worrisome and here is why: 1- Garbage in, garbage out: depending on the model performance and context size, more than often you can get a stupid answer that you simply say “are you sure”? And your model starts hallucinating and get paranoid. People often take the first answer as gold without critical thinking 2- AI is objective, not instinctual: Live is often made of grey area. The use of critical thinking and gut feeling is often the fruit of experience and instinct. No matter how we train models, it will never be able to make critical decision or suggestions that could replace gut feelings. 3- vibe coding: Amateur and overnight developer are a vulnerability for the entire technological Secteur. Amateurs copy pasta AI into coding like it’s hot and a lot of mistakes or critical security vulnerabilities are created. If you don’t know python or any programming language, please step aside. 4- influencer pandemic: Your AI agent will cost you a fortune in API credit and hardware and will only profit the influencers, Apple and API credits companies. You will never ROI and will find out how useless it is. For the influencers, that Mac Mini is a tool for engagement farming and will ROI ten folds in X add revenue in FIAT. If you see a video of a chick in Lulu that have no idea what to do with Openclaw and doesn’t know a thing about AI, please avoid like the plague. TLDR; You shouldn’t be scared of the rise of Artifical Intelligence but rather be terrified of the declined of Normal Intelligence.

So which models are you guys using for normal conversations? Llama3 8B instruct seems to be limited in its reasoning. I don’t want to pay a fortune by inference

The problem with AI API keys is the “JUST IN CASE”. I am trying to iron out the best possible scenario. I am using the Venice AI ecosystem. Sometimes my mind wonders if have these daily credits that I do not use are acting more like vulnerabilities since these tokens are volatile. Sometimes I explore in my mind the scenario of liquidating these tokens and simply use a service like @npub16g4u...kv4h with #monero or #bitcoin top up. VVV have a 7 days lock up and by design you cannot exit rapidly. I don’t use that many credits daily; I do not code, I just keep “just in case”.

I’ve been reading and experimenting with AI Agents and anything AI related for a while. I cannot wrap my head around a useful case for Openclaw. The only information i get out of it are from influencers that buy Mac mini and say how smart their Agents are. Besides admin task to manage influencers workflow. For any decent human, why the fuck do I need an Agent to read emails or txt my mom. For my homelab I use N8N and Cron. I use different API keys and inference. Why would I need more??

I am pretty bored with #bitcoin or #monero or crypto in general. I think influencers and companies are the only one having any type of ROI because they get direct capital which are not tied to price action. We are the exit liquidity

Just changed my wallet adress to settebello@cake.cash @npub1lsrm...m040 is upping up their game!