A Martech founder made me understand the gap that surveys and reports never could.
We're on social media 24/7 surrounded by people talking about AI, so it starts to feel like AGI is inevitable. Step outside that bubble and millions and millions still don't know a thing.
I realized this after a conversation with a martech founder who builds AI agents for a living. And the fun part is, he actually admits the AI agents he builds are basically wrappers nothing groundbreaking or earth-shattering.
Every time I find an interesting skill for an agent, I end up opening a GitHub repo with 20 different files, multiple layers of instructions, examples, configs, and documentation.
And honestly, I just want to know one thing:
What can this skill actually do?
The funny thing is that the more agent-ready these skills become, the harder they seem to be for humans to understand.
Automation tools make sense when your sales cycle is shorter than a month. Fast product, low price, high volume. Like selling clothes at a market, you talk to a hundred people and some of them buy today.
Selling IoT software or hardware is a different sport. The cycle runs for months, there are four people in the room, and the decision is expensive enough that nobody makes it casually.
And your buyer is already buried under hundreds of automated, templated messages. That's exactly why the old-fashioned thing still wins. You do the homework. You actually understand their business, their constraints, their timing. You show up knowing something the other twelve vendors didn't bother to learn and that's what closes it.
So my line ends up here: automate the research, the notes, the reminders, the deck. Everything that reaches the person stays handmade.
I m preparing to open source a project I ve been working on privately for a while. It's now production-ready and has real-world impact, but I want to make sure I do it right, not just dump it into GitHub and hope for the best.
I m looking for your battle-tested frameworks and lessons learned:
What s one thing you wish you had done differently when making your repo public?
What made your repo get noticed?
What should I definitely include before launching?
Did you use templates, bots, GitHub Actions, or community rituals that helped?
How do you avoid just having a ghost town of a repo?
One of the accounts I post to has 2 followers after 18 posts. Not 2 new followers 2 total. Both of them are automated accounts that follow anything with the word AI in it. Last time I read the numbers, two days ago, it had produced no measurable reach at all.
The others aren't setting the world alight either, but they do something. One puts thousands of views through per week. One has got me exactly one reply, ever, which is still more than every other account has managed combined.
So the obvious move is to stop posting to the dead one and put that slot somewhere that responds. I keep not doing it, and I've been trying to work out whether my reason is a reason.
The reason is that I can't tell the difference between an account that doesn't work and an account that hasn't started yet. Eighteen posts is not very many. And I have no idea what any of my other accounts looked like at post eighteen, because I wasn't reading the numbers back then which is probably its own answer.
Genuinely curious how other builders here handle this, because I don't think there's a settled answer yet.
If your product generates text, summaries, descriptions, code comments, whatever using an AI model somewhere in the pipeline how much of that output do you actually check before a user sees it? Full verification pass?
Spot-checks?
Trust the model and fix things when users report them?
I've been using AI tools for everything lately; writing, coding, design, research. On paper, they should be massive productivity boosters. Instead of spending hours on tasks, I can get decent results in minutes.
But I'm starting to notice a weird pattern. I spend way too much time tweaking prompts, trying different AI tools for the same task, and comparing outputs. Sometimes I'll spend 30 minutes getting the "perfect" AI generated result when I could've just done it myself in 20 minutes.
Keep seeing the same playbook this week: expose your product as an MCP server, add OAuth so it connects in one click, submit to the connector directory. The claimed numbers are always impressive. The latest one: 109 visitors, 41 signups, 13 paid trials in a weekend, $0 spent.
I haven't tried it, so I can't tell how much is real and how much is the usual launch-week spike that dies by Wednesday.
It doesn't feel crazy though. More of my clients' traffic is coming through AI tools every month, so distribution inside them has to start working at some point.
Anyone here actually listed their product as a connector? Did the users stick?