While I was on YouTube today, I came across a video by Alex Finn talking about how amazing the new Hermes agent is.
Hermes agent is something that came out a while ago, and is marketed as an agent that gets better the more it learns about you.
(Just like…any other agents?)
What surprised me is the fact that Alex was talking about Hermes at all. Because from the moment OpenClaw came out, Alex’ channel was already all about OpenClaw.
Every single day, it was OpenClaw this, OpenClaw that.
I get it. There’s strong interests there, so YouTubers SHOULD be talking about that.
Just like how many YouTubers are talking about Hermes right now, even though it was already here so long ago.
What’s most interesting about the video to me however was what Alex said about OpenClaw.
Apparently, whenever the OpenClaw team pushes an update, things about the app break.
Which is very interesting because it mirrored exactly the symptom I saw when I pushed an update to my NLC customers last week, which broke some stuff (which I then had to spend a lot of time doing recovery work).
The difference is that I was scrambling day and night to fix those broken stuff, and strategized to never again have an update break my customer’s NLC.
This includes:
Creating a dedicated workspace for all customer’s projects and works
Preparing within said dedicated workspace a folder architecture based on best practices
Putting guardrails in place so that customers’ personalizations do not conflict with NLC’s main update areas (this is what broke OpenClaw)
Among many other things I worked on all last week, these 3 were the core work that I’ve done.
But it dawned on me today listening to Alex Finn that basically, while I was burning the midnight oil doing plumbing work for my customers, the OpenClaw community is pretty chill with all these issues.
Guys like Alex don’t have any issues with updates breaking OpenClaw, because they know how to easily recover their work.
People like myself, I absolutely hate it when my flow gets disrupted and I have to figure out how to regain stabilization.
Which is why I work so hard to fix issues ASAP for my customers. Because I imagine myself on the receiving end of it, and I would be nothing but furious that my flow was disrupted.
Today, someone asked me “What’s the difference between OpenClaw and NLC?”
The man asked a good question - and I can totally see why from his perspective it should be an easy thing to answer.
But it is not.
You have to have experienced how its like to be running OpenClaw, Hermes, Claude Code CLI, and so on, and be the kind of person who HATES these experiences, in order to truly understand.
And a huge part of me just want to reply “NLC is more reliable.” But what the hell does that even mean, “more reliable?”
Yet this is the fullest truth.
This is why I am STILL working on the boring plumbing of NLC.
For the past 2 weeks, I have been focused mostly on stabilizing the app, stamping out edge cases that breaks the experience, and so on.
I just really can’t be arsed to focus on building the sexy stuff when I know that the basic stuff isn’t smooth.
This attitude I have may cost me effectiveness in marketing.
But I would rather do this, than market well and take on business that I can’t handle because my core deliverable is wonky.
So right now, as we speak, my agents are doing absolutely nothing other than the boring work.
The 15 machines I already installed are experiencing minimal friction right now.
My mission is to make minimal friction the default experience for the next 150.
If you are someone interested to have the power of agents working for your business, minus the hassle of having to maintain the setup yourself…
If you want a powerful AND reliable agentic setup, maintained by someone obsessed with the minute boring work that makes your system predictable..
…reply here.
I will put you on a wait list the moment the next 10 slots open up for my NLC setup.
FELIX
P.S. Feel free to ask for details too.