Knzo

@0xDezo lets fcking go bro built an AI factory

@0xKnzo dude thats an insane setup

CHINESE GUY BUILT A 300 MAC MINI FARM AND MADE $312K LAST MONTH $312,000 last month. That is what Liang pulls with his 300 Mac Mini AI farm. A year ago he was bleeding hundreds every month on Claude Code, ChatGPT Pro, Cursor and the rest. All of them metered him, throttled when volume spiked and told him nothing new. So he built the farm. 300 Mac Minis wired into metal racks in the warehouse. Swarms of local agents on Ollama. They search markets, run real-time predictions, execute tasks, grind opportunities and close loops 24/7. Nothing leaves the building. No API calls. No surprise bills. No rate limits. Pause at 0:04. The monitor lights up with a full grid of live agent sessions firing non-stop in parallel. The farm runs while he sleeps. Agents work on schedule, every few hours, every night. It flags high-value signals, hands them to the next agents, completes full workflows by morning. Costs pennies in electricity. Replaced thousands in subscriptions and unlocked real scale. His girlfriend asked what was humming at 3AM. He said the boxes were working. He stopped renting AI. He owns the machine that prints the money. Watch the full farm and live agent grid here:

@Vexo200 little secret sauce

@0xKnzo Ran this exact setup for 2 months. The thing you don't mention: the psychological shift is bigger than the math. When an inference costs $0 you just… try more stuff. That's where the real ROI hides

@Jeyxbt true that!!

@0xKnzo Most people are afraid to invest in this setup, they just don't realize how much potential it unlocks

@0xCodez u the goat

@0xKnzo Damn bro. This setup looks just crazy. Going to run same setup for me

@gippp69 hermes is gem. liked

THIS GIRL USED HERMES AGENT AFTER IT HIT 140,000 GITHUB STARS AND TURNED A $0 LOCAL SETUP INTO $5,200/MONTH CLIENT REPORTS hermes is not a normal agent that forgets everything after one task. finish a report once and it saves the process as a local skill, so the next client starts with memory, structure, and a workflow already sitting on your laptop after 20+ self-created skills, similar tasks can run around 40% faster than a fresh agent. less prompting, less cleanup, fewer repeated instructions. the system slowly learns the exact shape of work you keep selling the setup takes about 30 minutes. install lm studio or ollama, pull qwen 3.6 27b, install hermes, connect it to localhost. zero api fees, zero cloud workspace, zero data leaving your machine this girl used it for startup competitor research reports. pricing, positioning, customer complaints, review mining, market gaps. a report founders usually pay $400-$700 for can be drafted in 15-20 minutes once the workflow is trained most people are still paying monthly for agents that reset every session. the better move is building a local skills folder that gets more valuable after every paid task watch her video first, then read the full Hermes breakdown below. the money is not in one prompt. it is in repeated work that slowly turns into an asset ↓

@gippp69 thanks bro

@0xKnzo added to bookmarks

@Flandermaxx bro thats a builder mindset

$9,400 a month. That is what a 24 year old in Porto pulls from pages he never appears on. A year ago he paid a social manager €2,400 a month to find trends and four SaaS tools to track them. All of them told him what worked three days after it stopped working. So he built one agent. It scrapes TikTok and Instagram through Apify, scores the outliers on a local Llama model in Ollama, drops the winners on his own board. Nothing leaves the building. No seats. No tokens billed. Pause at 0:30. The screen reads: 3 verified outliers found. Runs again tonight at 3AM. Every 2 days, as scheduled. He is asleep when it works. It flags a creator doing 46x their follower count, hands him the hook, his editor fills a Kling template by noon. Costs $19 in power. Replaced a €2,400 salary and five subscriptions. His girlfriend asked who he texts at 3AM. He said nobody. The box was working. He stopped renting trends. He stopped renting tools. He owns the machine that finds the money.

@NazarSkoch sheeesh thanks

$459 a month on AI subscriptions. Every month. Automatically. Nobody decided to spend $5,500 a year. Just $20 at a time- and suddenly you're Uber burning $3.4 billion in four months. The most expensive habit of a modern developer isn't coffee. It's the subscriptions you forgot to cancel. ☕💸

@0xDezo yeah

@0xKnzo broski local AI plus execution speed is a crazy combo

CHINESE DEV TURNED $5.3K INTO $71K IN 24 DAYS ON POLYMARKET He bought three Mac Minis and one Mac Studio. Pause at 0:04 — that's the entire team. One local AI system that scans thousands of wallets, picks the top 15 winners, and copies their moves in seconds. No cloud delays. Runs 24/7. Average $4,700 profit per day. Most people use AI to write tweets. He used it to build a machine that prints money while he sleeps.

CHINESE DEV JUST BANKED $3,200 LAST MONTH RUNNING THIS LOCAL AI SETUP Stop the video at 0:22 when the model list appears on screen. That is the exact moment you see why he is not bleeding cash on API limits anymore. Qwen 3 14B and DeepSeek R1 doing the real work on his own $699 Mac Mini instead of renting intelligence every single month. While everyone else is still paying the subscription tax he turned the savings straight into profit. The API companies built their business on devs who never own anything. But the real next level with the same local stack is already moving:

@OpenAI meanwhile i am just automating drop shipping ads

We’re bringing new capabilities to GPT-Rosalind, a model series purpose-built for life sciences research at enterprise scale. It brings GPT-5.5’s agentic coding and tool use together with stronger intelligence for drug discovery, analysis, design, and experimental workflows. openai.com/index/introduc…

@ClaudeDevs feed it csvs and hope it hallucinates correctly

How do we automate business analytics with Claude? New blog post covering our best practices for skills, data foundations, and evaluations when building agents to perform data analysis: claude.com/blog/how-anthr…

@ClaudeDevs another layer of busy work for us

We've changed the trigger word from "workflow" to "ultracode". You can still say "use a workflow for this", but when you're clearly referring to something else, Claude won't kick off a dynamic workflow. For an explicit trigger, use "ultracode". We appreciate the feedback!

@nvidia cool but can it automate my drop shipping

@AnthropicAI they are just automating the same old traps

How well do the security community's techniques hold up against AI-enabled cyberattacks? We examined 832 malicious accounts and mapped their activity onto a longstanding database of tactics and techniques used by threat actors. Here's what we learned:anthropic.com/news/AI-enable…

@elonmusk everyone wants the profit without the work

That’s what it comes down to

@nvidia more local compute for the inevitable vfx ramp

Local AI Agents are leveling up across DGX Spark & RTX PCs. NVIDIA OpenShell is coming to Windows alongside new agentic AI optimizations and creator app updates—including NVIDIA Broadcast 2.2, plus upcoming RTX acceleration for Adobe apps and Blender. Learn More: nvda.ws/4fliRD7