There you have it! If you've ever wondered how one person runs the equivalent of a 150-person co…
September 22, 2026 · 0 likes · 0 comments
AI Cybersecurity Workforce
There you have it! If you've ever wondered how one person runs the equivalent of a 150-person company without the actual people, this is the show to catch. In about six hours I'm going live — 1:30pm ET, today, right here on LinkedIn.
https://lnkd.in/e6j6-_xH
I'm walking you through exactly how I use AI and a fleet of agents to run my entire operation. Research. Code. Security. Content.
Watch it live. It's the closest look you'll get at where all of this is actually headed.
And on cue, here's the news that makes the timing worse.
Figure AI just sent a humanoid robot into 30 homes it had never entered. Not a lab. Not a demo stage taped off for the cameras. Thirty real houses in the Bay Area, none of them in its training data.
It made beds. It folded towels. It tidied up toys.
Success rate: 56%. Across 420 trials, it finished 237 of them.
No data collected from those homes. A fixed system, same brain in every house. This is a machine walking into a stranger's home cold and getting more than half the job right on the first try. A year ago the same approach scored 9%. That's a 6x jump. In one model generation.
56% is not "good enough to fire the cleaning crew." But that is not the point, and it never is with this stuff. The point is the slope of the line. 9 to 56 in one release. You do not need a PhD to draw where the next two releases land.
Now stack the two halves of the picture on top of each other.
On the white-collar side, the CEO of Anthropic — the company whose AI I use every single day — has already said out loud that AI could eliminate half of all entry-level white-collar jobs within five years. Half. The lawyers, the analysts, the coders, the coordinators. The desk.
And now the other half. The physical half. The part everyone comforted themselves with — "sure, but a robot can't walk into my house and actually do the work." That comfort just took a 56% hit.
I wrote about exactly this line in REPLACEMENT. In the chapter on what gets replaced first, I said the physical labor survives and the cognitive labor gets eaten. That the wood still needs to be cut even when the AI does the design.
I need to be honest about where that leaves us. When the desk jobs go to the models and the hands-on jobs go to the robots, you don't have a "disruption." You have a squeeze from both ends at the same time. White collar from above, blue collar from below. That is not a soft landing. Handled the way we are handling it right now — pretending, delaying, hoping it stalls — it heads toward total disaster.
I'm not saying this to scare anyone. I'm saying it because it's true, and because I have three daughters and I think about what work even looks like for them.
Source: https://lnkd.in/eMBiC8bx
What's the last job you're certain a machine can't touch — and how sure are you, really?
https://lnkd.in/e6j6-_xH
I'm walking you through exactly how I use AI and a fleet of agents to run my entire operation. Research. Code. Security. Content.
Watch it live. It's the closest look you'll get at where all of this is actually headed.
And on cue, here's the news that makes the timing worse.
Figure AI just sent a humanoid robot into 30 homes it had never entered. Not a lab. Not a demo stage taped off for the cameras. Thirty real houses in the Bay Area, none of them in its training data.
It made beds. It folded towels. It tidied up toys.
Success rate: 56%. Across 420 trials, it finished 237 of them.
No data collected from those homes. A fixed system, same brain in every house. This is a machine walking into a stranger's home cold and getting more than half the job right on the first try. A year ago the same approach scored 9%. That's a 6x jump. In one model generation.
56% is not "good enough to fire the cleaning crew." But that is not the point, and it never is with this stuff. The point is the slope of the line. 9 to 56 in one release. You do not need a PhD to draw where the next two releases land.
Now stack the two halves of the picture on top of each other.
On the white-collar side, the CEO of Anthropic — the company whose AI I use every single day — has already said out loud that AI could eliminate half of all entry-level white-collar jobs within five years. Half. The lawyers, the analysts, the coders, the coordinators. The desk.
And now the other half. The physical half. The part everyone comforted themselves with — "sure, but a robot can't walk into my house and actually do the work." That comfort just took a 56% hit.
I wrote about exactly this line in REPLACEMENT. In the chapter on what gets replaced first, I said the physical labor survives and the cognitive labor gets eaten. That the wood still needs to be cut even when the AI does the design.
I need to be honest about where that leaves us. When the desk jobs go to the models and the hands-on jobs go to the robots, you don't have a "disruption." You have a squeeze from both ends at the same time. White collar from above, blue collar from below. That is not a soft landing. Handled the way we are handling it right now — pretending, delaying, hoping it stalls — it heads toward total disaster.
I'm not saying this to scare anyone. I'm saying it because it's true, and because I have three daughters and I think about what work even looks like for them.
Source: https://lnkd.in/eMBiC8bx
What's the last job you're certain a machine can't touch — and how sure are you, really?