Spent the last stretch wiring MCP servers into an agent product instead of hand-building every integration. The pitch is great in theory: paste a URL, click connect, tools show up as a checklist.

In practice it's been a mixed bag. Some servers are exactly as plug-and-play as advertised. Others expose tools that don't match how the underlying API actually behaves, so the agent confidently calls something that returns garbage, and you don't find out until you're debugging why a real user got a wrong answer.

Ended up hand-building the connectors that really matter (the ones where getting the semantics wrong costs someone money) and using MCP purely for the long tail where "mostly works" is good enough.

Anyone else gone deep on MCP servers? Curious if others found the same gap between the demo and production, or if I just hit unlucky servers.

Every landing page teardown I've read treats social proof as a solved component.

Logo bar, three testimonials, a usage number. Slot it in, move on.

But it's the one element on the page you can't produce yourself. Copy, design,

Been down a bit of a rabbit hole comparing robot vacuums over the past few months (two dogs at home, so cleaning is a daily problem, not a weekly one). And the thing that surprised me is how much the marketing focuses on one spec suction power while the actual dealbreaker for most people turns out to be something else entirely: the brush roll.

We were getting slammed with support tickets from users stuck in password reset loops. When we checked the logs, we realized the issue wasn't broken code it was a 40 second email delay.

Users would leave the tab or click resend before the first link even hit their inbox.

About eighteen months into our product, we hit a wall that wasn't a feature problem. Every new thing we shipped took roughly three times longer than the last comparable one. Bug fixes kept reopening. Two engineers had built quiet, undocumented workarounds around the same broken assumption in our data model, and neither knew the other had done it.

The obvious answer was a rewrite. We didn't do it not immediately. We spent a month arguing about it instead, which in hindsight was the most useful month we spent that year, because it forced us to name the actual problem: it wasn't the codebase. It was one early decision about how we modeled user accounts that everything else had been built to work around.

We ended up doing a partial rebuild. Roughly 40% of the code, the part that touched that assumption. Took four months instead of the fourteen a full rewrite would have taken. Shipping slowed but never stopped.

I still don't know if that was the right call. A full rewrite might have cleared out debt we're still paying for. Or it might have killed us the graveyard of startups that spent a year rebuilding while a competitor shipped is not small.

We measured something today that I had never thought to look for, and I suspect most teams are not looking for it either.

Two of our customer sites, one a mechanical keyboard store and one a perfume store, came back after their containers were recreated. For roughly 30 to 60 seconds, both of them served exactly the same page. Not a similar page. Byte for byte identical, 281475 bytes on each. The title tag carried the name of a third brand entirely, and the canonical tag pointed at a localhost address on port 3000.

Whenever I search for a skill.MD file for a relevant skill there are tons of Github repositories that I find and none of them actually become useful to me because most of them have AI slop and actual hardcoded instruction which is necessary is somewhere buried in line 12 or 13 and the plug-in comes with a bunch of more files which I can swear I have never read.
What is the best way to share a skill that can be plugged to agentic workflows? And is anybody monetising this already?
With most of the skills that I find online, there is no true sense of ownership and most of them lack authenticity!!

Before vibe coding I would usually think twice before starting a new project

The setup alone could take a lot of time

Now I can test a small idea much faster and see if it is actually worth continuing

Some ideas still fail

I've noticed something lately: the more productive tools I add to my workflow, the more time I seem to spend maintaining the workflow itself.

One app for tasks.
One for notes.
One for bookmarks.
One for team communication.
One for documents.

Then I need another tool to connect all of those tools.

At what point does productivity software become another thing we have to manage?