leopardracer

researcher & analyst | eth maxi

[atrekovas @atrekovas]

CLOUD AI ANSWERS IN 3 SECONDS AND FORGETS EVERYTHING LOCAL AI TAKES 15 MINUTES AND REMEMBERS FOREVER

hermes running gemma 4 26b through lm studio, compressing context on the fly, processing prompts locally while your data never leaves your machine

yes it’s slower, yes it compounds every single day, yes the gap between people who own their AI and people who rent it is about to get very wide

full breakdown on how this actually works ↓

[leopardracer @leopardracer]

@cyrilXBT thanks Cyril 🙌🏽

[leopardracer @leopardracer]

https://t.co/505u4EnEdI

[Avid @Av1dlive]

Hermes Agent. Zero to full autonomous operation. One complete course.

Installation. Skills. Memory. MCP. Scheduler. Multi-agent.

Works while you sleep.

The people who build this system will never manually operate a content, research, or business workflow again.

Read this and bookmark it now.

[leopardracer @leopardracer]

@ridark_eth thanks for sharing

[leopardracer @leopardracer]

A NON-PROGRAMMER JUST BUILT A 3-AGENT AI TEAM AND RUNS THEM ALL FROM TELEGRAM

Hermes plays boss and two openclaw agents take orders while the whole thing runs on two mac minis and he's texting tasks from his phone at a coffee shop or walking his dog and none of it required a single line of code

YouTube channel, bookkeeping, daily ops, all handled by three agents that never clock out

bookmark this or spend 3 hours looking for it the day you finally get tired of doing everything yourself ↓

[leopardracer @leopardracer]

https://t.co/505u4EnEdI

[leopardracer @leopardracer]

@ryoxxai hermes folder in obsidian, skills directory as the source of truth, everything else flows from there, full breakdown in the article

[leopardracer @leopardracer]

140,000 DEVS STARRED THIS REPO IN 3 MONTHS AND MOST OF THEM STILL DON’T KNOW HOW TO USE IT

pair hermes agent with obsidian and notebooklm and you get a second brain that writes its own skills, maps its own knowledge, and never forgets what you taught it

while everyone else is copy-pasting into chatgpt every morning and starting from zero, this thing has been compounding for months

one setup, runs locally, gets smarter every single day

you’ll want this saved the day you get tired of explaining yourself to an ai that doesn’t remember you ↓

[leopardracer @leopardracer]

https://t.co/505u4EnEdI

[leopardracer @leopardracer]

@NeuralNavQ exactly, that’s the whole argument

[leopardracer @leopardracer]

TIM DICKEY GAVE UP HIS ENTIRE WEEKEND TO RUN HERMES LOCALLY

5 compression cycles, a 16b model optimized from scratch, two days of saturday and sunday gone, all to make sure not a single byte of his workflow touches a server he doesn’t own

most people don’t think about where their data goes, tim does, and the gap between those two groups is about to matter a lot

full breakdown on why this is the only setup that makes sense long term ↓

[leopardracer @leopardracer]

https://t.co/505u4EnEdI

[leopardracer @leopardracer]

@0xAlpha__ run it for a month and come back

[leopardracer @leopardracer]

@SlackingTom qwen 3.6 35b q4 is the sweet spot on spark, your setup looks solid, 262k context is plenty for hermes to run properly

[leopardracer @leopardracer]

@olivertaste exactly, that’s the whole point of the skills system, clean recovery is a learned behavior not a default

[leopardracer @leopardracer]

@vorty279 the guardrails catch obvious regressions but subtle goal drift is still on you to catch, that’s why the article mentions keeping an eye on the skills directory, it’s the clearest signal something’s drifting

[leopardracer @leopardracer]

@neil_xbt glad it landed🙌🏽

[leopardracer @leopardracer]

@alphabatcher the combo is underrated, most people still treat them as separate tools

[leopardracer @leopardracer]

@Blum_OG always 🤝🏽

[leopardracer @leopardracer]

@ridark_eth I saved this

[leopardracer @leopardracer]

@zostaff saved!

[leopardracer @leopardracer]

@h100envy nice share!

[leopardracer @leopardracer]

@slash1sol nice share!

[leopardracer @leopardracer]

@gippp69 nice share!

[leopardracer @leopardracer]

@h100envy nice share