OpenClaw is sorta dead now, and how to avoid the next AI flavor-of-the-day
I'm not a sore winner, but I'm going to enjoy my victory lap
We’re all old enough to remember February, 2026 (3 months ago) when OpenClaw was literally all the rage, right?
Well, it’s de-facto dead. Not literally. You can still run it, although I’m not sure what you’d do with it, but nobody is searching for it.
I’m not a sore winner, but when I proudly posted in February that OpenClaw and AI Browsers were Try, Oh My, Goodbye, I got some flack from people that called me “cynical” and “talking my book” by saying that AI and automation is really still a job for developers (which I have on my team).
Now, I’m going to give you the formula as to why I knew, without ever downloading OpenClaw, that this was going to be a hype cycle that died as soon as it lived.
My goal for you is that you can avoid getting wrapped up in the next hype cycle, whatever it may be. I hope you can focus your energy on stuff that matters.
First, and most importantly, everyone was talking about it but nobody was actually using it for anything meaningful.
I asked everyone who told me about it “What are you using it for?” And I sorta didn’t get answers? Anytime the answer is hard to decipher, it doesn’t mean there’s something so complex that I couldn’t possibly comprehend it. It means that the solution doesn’t make sense.
The number 1 thing that people said was “I text it to do stuff while I’m away,” to which my response was “What?” And they didn’t say anything that you or I couldn’t do from our phones already.
Notice how I don’t list any use cases? It’s because there weren’t any that were specific to OpenClaw that drove meaningful value.
Contrast that to simple ChatGPT Deep Research. You ask it a question and trigger Deep Research mode, and then 20 minutes later it delivers you a 5-10 page report about that topic. It’s easy to say and easy to envision exactly how you’d use it. If someone asked me “What’s the last thing you used ChatGPT Deep Research for?” I’d say “I was doing research on Palantir’s business model and contrasting it to Oracle’s.” Easy and done.
In another, similar, example, before ChatGPT was just GPT. Companies using GPT to write marketing copy, such as Copy.ai, had real revenues and raised money at 8-figure valuations. While that’s not to say “it was clear ChatGPT was going to be a hit,” it was clear that the underlying technology was already making a difference. The same was not true with OpenClaw.
Second, the technology is wayyyyyyyy to slow to be anything you use regularly
This was true with AI browsers, which I had tried. Finding flights on Google flights takes approximately 15 seconds. Finding them using ChatGPT Atlas, last I tried, took 7 minutes. Perplexity Comet was no different. OpenClaw, from what I had heard, was the same.
People would say “Well, it takes longer, but you’re not doing the work. You can do other stuff.”
While true, I still need to cross the task off my list, and in doing a task like booking flights, I, ultimately, need to press “book.” It’s not a task that’s getting done overnight and then being reported to me. Using AI should make my life faster and more efficient, not simply move work around.
Will the technology get better and faster? Probably. Does that matter in the long run? Probably not, because, as I stated in the aforementioned essay, AI, really, should be invisible. Needing to manage its slowness and mistakes is not a task I have bandwidth for.
So how can you avoid the next hype cycle, and what will it be?
“What do you use it for?” Is always the question I ask.
The legitimacy of a hype cycle comes down to one thing: does the thing being hyped create meaningful-enough value that you will change your habits to use it?
Usually, meaningful-enough value consists of time savings or efficiency savings.
If you have an assistant spending 1 hour per day creating your morning report (and thereby working on Sunday evenings), an AI automating that is a huge benefit, but if you already have it mostly automated, using OpenClaw to retrieve it for you makes no sense.
If every day you need to check a particular database or website or Excel file and then transfer the data into another system (for example, data from form responses being transferred into your CRM), then you probably already used Zapier or Claude Skills to do that, and OpenClaw doesn’t make sense.
I spend most of my time telling people “relax, just focus on understanding what you team does and automating the tasks that suck.”
The AI-flavor-of-the-day is mostly hype.
Admittedly, one person’s hype is another person’s time savings (for example, I know of someone saving several meaningful hours per week using Claude Cowork), but, for most people, hype is just wasted energy.
I hope this is my last time, for a while, writing about hype cycles and saying “just focus on one or two things that matter.” I speak up so much because wasted energy for you can easily become wasted energy for your teams. By all means pay attention to the flavor of the day, but there’s near-zero harm in waiting 2-3 months before adopting the technology. I’ve learned enough lessons the hard way; I’m okay waiting to see what people do first.