Jarvis has always been my fantasy. (Still is) Drive connector went haywire for me. initially I should have payed attention. Nope 3 weeks kicked my ass.
The 'mistakes first' framing is underrated — most tutorials skip straight to the happy path and leave beginners confused when things break. Curious what the biggest time-sink was for you: the prompting, the tool integration, or getting the loops right?
I think the biggest one was things that I wanted to do fast and had very little idea about them. The first time I introduced night shifts to my AI agent, it actually cost me a day or so to fix things, because I was just not too optimistic about the route. The route at night went too wide. I would say it was terrible, and I had to fix it afterwards.
ha, night shifts for AI agents is a dangerous game. did the same with an auto-merge bot once — woke up to 14 PRs merged and 3 broken builds. the fix always takes longer than the thing you were trying to speed up
Good question. Opposite actually. Claude Code, Cortex, OpenClaw are agents, so the basics (the loop, tool use, context, memory) become the mental model you need to use them well. Without it you hit a wall the first time one misbehaves or you want it to do something it doesn't do out of the box. I run most of my day on Claude Code and knowing what's happening under the hood is the difference between a $20 toy and a system that runs my nightshift unattended.
Wow this guy writes awesome
Bueno
AI agents are going to become as common as websites. The real advantage won't be building one it'll be knowing where they create the most leverage.
I heart you, man!
Jarvis has always been my fantasy. (Still is) Drive connector went haywire for me. initially I should have payed attention. Nope 3 weeks kicked my ass.
Popculture had broken us all xD
Just don’t feed the BOWKYs personal info.
Just found this today. New follower 😊 Provided a ton of clarity!
Thank you! Sorry for such late reply, but my humble substack growth is crazy past 2 months :D
Thanks for your guide
I'm a beginner and this helps me to understand what to do and don't
i wish i get to know about this earlier 😭
Thank you for putting in the effort to make this actually helpful.
Thanks for comment - I was trying my best!
Thanks a lot!
cute
That's cool! We need more simple how-tos like this. Tried to come up with my own piece of work:
https://substack.com/@larsb3/note/c-258573963?r=6zu368
The 'mistakes first' framing is underrated — most tutorials skip straight to the happy path and leave beginners confused when things break. Curious what the biggest time-sink was for you: the prompting, the tool integration, or getting the loops right?
I think the biggest one was things that I wanted to do fast and had very little idea about them. The first time I introduced night shifts to my AI agent, it actually cost me a day or so to fix things, because I was just not too optimistic about the route. The route at night went too wide. I would say it was terrible, and I had to fix it afterwards.
ha, night shifts for AI agents is a dangerous game. did the same with an auto-merge bot once — woke up to 14 PRs merged and 3 broken builds. the fix always takes longer than the thing you were trying to speed up
Nighshifts are very bad, if you don’t know what you want :D
Been there xD
lol yeah the worst is when it does exactly what you asked but not what you *meant*. that's the real danger of unsupervised agents
The biggest beginner mistake is thinking an agent is “an LLM plus tools.”
That is only the visible layer.
A useful agent usually needs:
- clear task boundaries
- memory policy
- tool permissions
- failure handling
- verification
- logging
- human override
- retry logic
- cost control
Without those, the agent is not autonomous.
It is just a loop with API access.
The hard part is not making it act.
The hard part is making it act safely, cheaply, and correctly when the world gets messy.
Building your first AI agent is one of the best ways to understand what these models are actually capable of.
100%
Cool? Does this become less important with tools like Claude Code, Cortex and OpenClaw?
Good question. Opposite actually. Claude Code, Cortex, OpenClaw are agents, so the basics (the loop, tool use, context, memory) become the mental model you need to use them well. Without it you hit a wall the first time one misbehaves or you want it to do something it doesn't do out of the box. I run most of my day on Claude Code and knowing what's happening under the hood is the difference between a $20 toy and a system that runs my nightshift unattended.
Love this!
Thank you :)