Previously on Swarm Chasers - Digital Swarms and the Persistence Problem
Intelligent Machines (Audio)
October 8, 2026
From deleting critical files to oversharing your banking details in Slack, today's AI agents are proving both incredibly helpful and hilariously unpredictable. Hear what actually happens when AI gets a little too persistent and why transparency about their actions matters more than ever.
Speakers Leo Laporte, Paris Martineau, Jeff Jarvis, Spencer Thompson
TopicsTechnology
Leo Laporte (0:00)
It's time for Intelligent Machines, Paris Martineau is here, Jeff Jarvis is here, our guest Spencer Thompson has a company that keeps track of what AI is doing behind your back. And apparently, it's a lot more than you might imagine. We'll also talk about Meta's Muse, it's very popular, but is it safe? And the judge who was persuaded by an AI video of a victim, not to throw the book at the killer. That and a whole lot more coming up. Next, on Intelligent Machines. Podcasts you love.
This is TWiT.
This is Intelligent Machines with Paris Martineau and Jeff Jarvis, episode 891, recorded Wednesday, October 7th, 2026, previously on Swarm Chasers. It's time for Intelligent Machines, where we cover AI, robotics, and all those smart little doohickeys all around us these days. Paris Martineau is here, investigative reporter from Consumer Reports. A big fan of the Liberty. What are they? The New York Liberty?
Paris Martineau (1:12)
The New York Liberty, which are in the semi-finals. I had a really rough game one, but game two is tonight, and it's got to be better.
Leo Laporte (1:22)
All right. What time do I have to get you out of here for the game?
Paris Martineau (1:25)
I mean, the game starts at 7.30, and I don't know how early the Queer Bar will fill up with fans.
Leo Laporte (1:33)
Oh, okay. So that's a good place to see it, actually.
Paris Martineau (1:36)
It's the perfect place to see it. It's all Liberty fans.
Leo Laporte (1:40)
All Liberty fans, Liberty liberties. Well, we'll get you out of here at 4 How about that so you can run over to the bar and get a good seat?
Paris Martineau (1:46)
Great. I'm doing time differences in my head.
Leo Laporte (1:51)
Wait a minute, 7pm, your time.
Paris Martineau (1:54)
Yes.
Leo Laporte (1:54)
Yes. All right. We're going to do that. Also, Jeff Jarvis is here. He is the Emeritus Professor of Journalistic Innovation at the Craig Newmark Graduate School of Journalism.
Craig Newmark. New York. He's also the author of Hot Type, bestseller, The Magnificent Machine.
Jeff Jarvis (2:13)
And this Saturday, I will be in Haverhill Mass at giving a talk on the book. And then we go up the hill to the Museum of Printing, where you can see a live line of type actually working. Go to jeffjarvis.com, where you can buy autographed copies and you can also register for free for the event Saturday.
Leo Laporte (2:27)
If you're in New England, come, please, please, I'm going to venture kind of the pinnacle of mechanical invention.
Jeff Jarvis (2:37)
Yes.
Leo Laporte (2:37)
Because it's more complicated than a pocket watch.
SPEAKER_4 (2:42)
I mean, this thing was made by a watchmaker.
Paris Martineau (2:45)
You have to walk up a hill to go see it at Jeff's event if you go on his website.
Jeff Jarvis (2:49)
Exactly.
Paris Martineau (2:50)
It implies the largeness based on a change in geography.
Jeff Jarvis (2:56)
What's so great is we think this is a machine now. You can't see anything.
Paris Martineau (3:01)
He's holding up a laptop.
Jeff Jarvis (3:02)
If anything is moving, it's a fan to cool the damn thing off.
Leo Laporte (3:07)
No.
Jeff Jarvis (3:07)
The mechanics of the line-of-type are all out there for everyone to see and see how brilliant it is.
And Ottmar Mergenthaler invented before his death the Model 1 He died in 1899 And every model from then to the 70s is recognizable from his design.
Leo Laporte (3:24)
Wow.
It's quite an invention. I should say that probably much of the complexity that's visible in a line-of-type is invisible, but in the chips and the circuit boards. The line-of-type is in your computer.
Jeff Jarvis (3:37)
With all the cogs, it's an algorithm in steel.
Leo Laporte (3:39)
Yeah.
That's a good way to put it. Maybe even use that in the book.
Jeff Jarvis (3:43)
I didn't do it.
Leo Laporte (3:46)
We're going to talk about AI today. And I want to introduce our guest, Spencer Thompson. He's the CEO and founder of Origin Technology. Joining us from Vancouver, a beautiful British Columbia. Look at that.
Spencer Thompson (3:59)
Oh. Perfect day. That's the background. Yeah, it's beautiful today.
Thank you for having me.
Leo Laporte (4:03)
I spent I should disclaim immediately Origin Technology is a sponsor. You might have heard our ads for them. And I met Spencer when we when we first lot of times with advertisers, we like to talk to them so I can understand what they're up to. And I was so interested in what you've done here with this, uh, that I said, you know, you really ought to come on our show. Everybody's talking these days about rogue agents escaping and hacking other sites and, you know, and how persistent they are and more and more we're hearing, especially with open AI, how invisible they are. And you have, uh, origin is kind of the solution. So I want to say upfront, I've never done this. I don't think I, to my memory, I've never put a sponsor on as an interview, but this is very timely. And I think your case studies in here are also your investigations, as you call them on the website are very interesting.
179 more minutes of transcript below
Thousands of transcripts fetched by people building searchable podcast archives
Fetch the whole transcript
The demo key returns a sample episode in full, no card needed:
curl -H "x-api-key: pt_demo" \
https://spoken.md/transcripts/1000651996090Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.
From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.
Using your own key:
curl -H "x-api-key: YOUR_KEY" \
https://spoken.md/transcripts/1000793789099