Activity
Mon
Wed
Fri
Sun
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More

Owned by Mike

TopAgent Experts and TopAgentU is an Exp Realty network of some of the TOP real estate Agents, Teams, Coaches and Marketers in the world. Partner UP!

6 contributions to TopOfMind AI Builders
New Classroom course is up: Hermes Skills That Ship Into a Business.
https://www.skool.com/topofmindai/classroom/58be1266 If you already have Hermes or OpenClaw running and you keep installing skills that never touch a real workflow, this is for you. We cover the security gate, clarify-before-act, handover docs, skill creator without bloat, teach/learn for business memory, a morning brief that serves the company, and multi-model use that does not burn tokens for sport. Every module has a short TLDR, a real SMB scenario, numbered steps, homework, plus a faceless explainer video and a diagram so you are not stuck in walls of text. We refuse the art-director / thumbnail / design-slop path. This is SMB ops: follow-up, missed calls, intake, briefs, light CRM hooks. Security is not optional. Start at Module 0. Watch the clip. Do the homework. Post what you refuse to install next. - Keith
0 likes • 23d
@Keith Motte question on the morning brief module: how do you keep it short enough that the owner actually reads it every day? It seems like the kind of thing that slowly grows into a wall of updates nobody opens.
$100K to build a coding agent that runs on your own machine
Every coding agent most of us use today depends on a big model running in someone else's cloud. This challenge asks what it would take to get that same capability offline, on consumer hardware. The task is to post-train Gemma 4, Google's open model, into a coding agent that can find its way around a real repository and draft fixes to real software engineering tasks on its own. Humans still handle architecture, logic, and review. There are two tracks: 1. Developer Agent Competition ($65K). You post-train Gemma 4 into an agent that resolves real engineering tasks. 2. Paper Track ($35K). You submit new, unpublished research, or you write up how you approached the competition. Total prizes are $100,000, and the entry deadline is November 25, 2026. This lines up with what we keep saying in here. An agent is only as useful as the loop around it: a plan before it touches files, tests it has to pass, and a human at the gate. A local agent makes that loop cheaper to run, and you don't have to send your code to a third party. A few questions for the group: - Is anyone entering? I'd like to put a small team together. - Who has done fine-tuning or reinforcement learning on an open model before? - What hardware are you running locally? Not a vendor brochure. Just a good problem. — Keith
0 likes • 24d
@Keith Motte not entering, but your hardware question hit home. I have two older Intel i7 iMacs with 32GB each, which is only enough for tiny local models. So my plan is to run the agents on them and let cheap cloud models do the thinking, like Kimi, MiniMax, DeepSeek or GLM through their subscriptions or Ollama Cloud. Is that a sensible split, or is there a catch I'm not seeing?
What's the blast radius if this goes wrong?
Two AI agents launched this summer that want full access to your email, calendar, and credit card. I'd tell most business owners to wait. Instinct is the one people are raving about. You text it like a person, and it books the reservation, negotiates the bill, and follows up without being asked. It's genuinely impressive. It also had a rough first week in public: → A user disconnected Gmail and still received an inbox summary hours later → A founder emailed himself hidden instructions as a test, and the agent followed them → Another founder reported it sent an email on her behalf without checking first Meta's Muse launched last week with a more careful design. It works on its own isolated computer, can't see your passwords or payment details, and lets you require approval before it acts. It's free to start, with paid tiers at $20 and $100 a month. But it's Meta, and trust is earned over years, not in a press release. I spent years leading Enterprise Support at AWS. The question we asked before anything touched a customer's production environment was simple: What's the blast radius if this goes wrong? An agent connected to your inbox, your clients, and your card has a very large blast radius. My advice for small business owners right now: 1. Try these on a throwaway account, not your business email 2. Set approvals to "always ask" 3. Keep them away from payroll, client data, and payments for now These tools will get there. The capability is real. But capability and trustworthiness aren't the same thing, and your customers are trusting you, not your AI. Would you give an AI agent access to your business inbox today? Why or why not?
0 likes • 28d
Reading it, yes. Sending from it, no, not without me seeing the message first. The founder who emailed himself instructions is the part that worries me most. Anyone can email you, so anyone can talk to your agent. Mine is set up so anything it reads in an email is information, never an order. If an email asks it to do something, it has to come back and ask me. Your third point is the right line. Payments stay a human job.
Agent Constraints and Security & Preventing Deletion/Destruction
Looking for Skills or Instructions for MD files for Hermes or Openclaw to prevent Deletion/Destruction but the agent Anyone have anything that they have used that works well?
0 likes • 29d
Writing the rules in a file didn't work for me. The agent would read them and still do what it wanted, because nothing actually stopped it. What worked was putting the stop in the code. Everything goes through one script, and by default it only tells me what it would do. It won't do it unless I add a flag that says go ahead. Two other things. If anything is missing or broken, have it refuse everything rather than allow everything. And check your limits still hold when something goes wrong, not just when it goes right. Mine claimed it was pausing between actions and it wasn't, and I'd never have caught that on my own. What's yours able to get at?
Coding Agent Platforms for Local Business Work (Oct 2026)
Quick practical comparison for anyone building or selling AI agent systems (especially with Hermes, OpenClaw, or coding agents) for local businesses. Cloudroom Multi-harness (Claude Code, Codex, Pi, Hermes, etc.), open-source Rust core, use your own model subscriptions (no markup). Free limited cloud agents + self-host option. Early-stage but strong on flexibility. Orgo Full persistent cloud computer per agent. Works with almost any agent. Good isolation model when you’re running agents for multiple clients. OpenHands Fully open-source, strong self-hosting, model-agnostic. Highest control, more ops work if you self-host. Official platforms (Cursor Cloud, Claude Code Cloud, Codex Cloud, Devin) More polished and lower friction, but higher lock-in. Better if you stay inside one ecosystem. Recommendation for this community: Orgo for client isolation + Cloudroom or OpenHands for the agent work. Self-host the open-source options when you want control over data and cost. What are you currently using to run agents for clients or internal systems? Drop your stack below. Image
Coding Agent Platforms for Local Business Work (Oct 2026)
0 likes • 30d
One thing worth adding from running these unattended: the machine matters more than the harness. Mine sits on a dedicated box that comes back by itself after a reboot. The two with disk encryption turned on just wait at the unlock screen until someone physically types a password, which defeats the point of an always-on agent. Worth settling that before picking a platform, because none of them solve it for you.
1-6 of 6
Mike Grbic
1
4 points to level up
@mike-grbic-5820
Billion Dollar Agent, 7,500 real estate transactions, Coach, Investor, Biohacker, Ironman Triathlete

Online now
Joined Mar 22, 2026
Powered by