At one point space was the familiar final frontier. Even with talk of putting data centers in space, I dare say we’ve moved the concept of frontier closer to terra forma and set aside the “final.” Frontiers require explorers who are willing to accept risks, pushing beyond them to discover if there’s any there there. Maybe we’re in the moment of redefining “there.”
I’ve been curiously watching recent developments on the frontiers of Artificial Intelligence around what was launched as Clawdbot, then became Moltbot, and molted into OpenClaw. At least I think that’s what it is still called as of this writing.
For those unfamiliar, essentially OpenClaw is an AI agent created by software engineer Peter Steinberger, that receives instructions from the user in a chat. Running locally on your computer it then connects to other AI sources and web based apps you give it permission to access. It performs those tasks and actions. Mike Elgan has a good rundown on the (brief) history and the ins and outs. I encourage you to read it.
Both fascinating and frightening, OpenClaw seems to have taken on a life of its own without any regard for guardrails. After Federico Viticci wrote an early post about what was Clawdbot at the time, interest shot through the roof, reminding me quite a bit of the furor over the still recent launch of ChatGPT and just about any other big computing innovation we’ve seen.
Quite a few jumped in with both feet to test the waters. Alongside of all of the splashing around came upfront real warnings that this thing was not secure. That proved to be even less effective than signs telling you not to run around the pool. Viticci mentioned that given security concerns the project was not really ready for everyday users, and recommended that those interested install it on a second computer, not their main one. Apparently there was even a run on Mac minis.
The promise seemed clear and the hype leapt into hyperspace. OpenClaw would become the user’s personal assistant doing whatever was required. That’s been the as yet unrealized promise so far in all of these AI adventures.
The moment continued to evolve to a point that there’s even a social network called Moltbook where these AI bots could talk with each other. (Sounds like Mark Zuckerberg’s dream.) Mathew Ingram writes about that here, linking to Simon Willison’s post Moltbook Is The Most Interesting Place On The Internet Right Now.
At the time of Mathew’s post there were 1.6 million agents participating. Not to spoil his article, which you should read, there is some doubt as to whether or not there are humans doing mischievous human things behind the scenes. (Again, sounds like Zuckerberg’s dream.)
Casey Newton gave it a try. Still Moltbot at the time of his writing, he fell in love and out again, eventually uninstalling the software saying that “maybe someday you’ll have a genie in your laptop working for you 24/7. Today is not that day.”
That reminded me of all of the users who said that ChatGPT would replace Google for all of their search needs in that first explosive week. It appears that though the excitement and hype is still boiling hot, not everyone is ready to be the chef that tosses the lobster in the pot.
On other fronts
Before all of the OpenClaw news became the main course of the moment there was another very interesting AI story that caught my attention.
Since January 7th, Apple Health users have been able to connect ChatGPT to Apple Health. Geoffrey Fowler gave it a try.
Like many people who strap on an Apple Watch every day, I’ve long wondered what a decade of that data might reveal about me. So I joined a brief wait list and gave ChatGPT access to the 29 million steps and 6 million heartbeat measurements stored in my Apple Health app. Then I asked the bot to grade my cardiac health.
It gave me an F.
I freaked out and went for a run. Then I sent ChatGPT’s report to my actual doctor.
The good news is Fowler was OK and his doctors told him to relax. The concerning news is that one of the promises of AI is that it would help with medical diagnosis and be a boon to patients and doctors alike.
Now, certainly Fowler’s experiment is different than what may happen under stricter supervision and stringent testing. And, as he points out, OpenAI and Anthropic say their digital doctor bots can’t replace the real thing and provide big bold disclaimers.
Fowler’s experiments didn’t stop short with his artificially intelligent failing grade. You should read the article to see how the adventures continued. Suffice it to say, the conclusions (not just the medical ones) currently leave much to be desired.
Then this morning I stumbled across this article from Om Malik called How AI Goes To Work. It’s a great story about how one user found a way to solve a problem he has with spreadsheets using AI. It also provides some great tech history context and leads to an opinion I share about where we are today:
My simpler explanation of “embedded intelligence” to myself makes me step away from the headlines and look at the present and the future in more realistic terms. My bet is that in five years, it will all be very different anyway. It always is. I am a believer in the power of silicon. When we have newer, more capable silicon, and more networks, we will end up with ever more capable computers in our hands. And the future will change.
For now, what I call embedded intelligence is a sensible on-ramp to the future. The hype may be about the frontier models. The disruption really is in the workflow.
As I said, I concur with that opinion and it colors all of my current observations of the AI landscape. Be curious and become informed. I go further and say I’m comfortable letting others take the first leap.
I don’t think there’s any denying that most of us would enjoy living in a world when we could sit down with our computing devices, talk to a pendant, or even the air around us, (anything without the name Siri preferably), wish the world a good morning, and have it spit out not only our tasks for the day but do many of those for us. Folks of my generation grew up on Star Trek and other science fiction where this seemed common place. So too, did the problems and catastrophes when circuits got crossed or corrupted.
So, it’s a new frontier. Maybe the final one. Maybe not. But at the moment, we’re still just humans crossing into it. Forget what the bots may eventually do to us. I think I’m more concerned about the humans.
(image from kentoh on Shutterstock)
You can also find more of my writings on a variety of topics on Medium at this link, including in the publications Ellemeno and Rome. I can also be found on social media under my name as above.
Life on the Wicked Stage: Act 3 