I’ve been experimenting with OpenClaw pretty early on and had a different reaction than most of the hype.
First, I think what Peter did was important. It showed you don’t need massive resources to build something that feels agentic and useful.
But after using it (over the past few months), I’m not convinced it represents a fundamentally new way of working.
What I see is:
- a UI layer
- orchestrating automations and scripts
- wrapping capabilities that already existed
It gets things done, but it feels more like different (e.g. not better) packaging than a new paradigm.
What’s interesting is the narrative around it. When people like Jensen Huang (who is a driven genius BTW) talk about these systems, it’s framed as a major shift in how work will happen. "Everyone needs an OpenClaw Strategy"
But in practice, it still feels like:
humans defining intent, stitching workflows, and dealing with edge cases.
It's not even a nicer interface.
So my current take is:
useful tool for spending lots of money, not a game-changing shift.
Curious how others see it.
Are tools like this actually changing how you work? Are they making existing workflows more convenient? Is it just fun technology?
I built Knit because I was tired of jumping between tools just to explain what I wanted to change in my code. All I wanted to do was point to the screen while my app was running and tell my coding agent to fix THAT.
The only way to do that was to take a screen shot or write down the UI element and then type it into my CLI or Coding Agent tool.
So I made Knit.
Knit allows you to do a screenshare with your coding agent — you describe what you want (even loosely), and it helps turn that into structured actions.
It’s early, but it already supports:
- Screenshare
- Audio (STT)
- Turning vague requests into actionable changes
- Working across different coding agents
- Helping reduce the time from idea to working code
I’m not trying to turn this into a company — I just want to make developers faster and remove friction from building.
BTW, I've testing the Codex CLI well. I'm sure you'll see bugs in Claude, Opencode, and the API integrations.
Also, I'm still waiting on Chrome to approve my Extension. That will make life (e.g. Knit) even better.
First, I think what Peter did was important. It showed you don’t need massive resources to build something that feels agentic and useful.
But after using it (over the past few months), I’m not convinced it represents a fundamentally new way of working.
What I see is: - a UI layer - orchestrating automations and scripts - wrapping capabilities that already existed
It gets things done, but it feels more like different (e.g. not better) packaging than a new paradigm.
What’s interesting is the narrative around it. When people like Jensen Huang (who is a driven genius BTW) talk about these systems, it’s framed as a major shift in how work will happen. "Everyone needs an OpenClaw Strategy"
But in practice, it still feels like: humans defining intent, stitching workflows, and dealing with edge cases.
It's not even a nicer interface.
So my current take is: useful tool for spending lots of money, not a game-changing shift.
Curious how others see it.
Are tools like this actually changing how you work? Are they making existing workflows more convenient? Is it just fun technology?