Ships abandoned at San Francisco during the 1949 California Gold Rush.
There have been five tech booms since Gordon Moore coined his eponymous Law in 1965:
- Mainframes (IBM, the BUNCH and minicomputers (1965-1982)
- PCs (1978-1985)
- Internet (aka "dot com") (1995-2000)
- Smartphones (2007-present)
- AI (2023-present)
In my long experience, the AI boom is the biggest of all. Tom Friedman calls it "the second big bang." Watch that. I think he's right.
Business-wise, we get a lot of sky-pie, naturally. This, for example, from Mordor Intelligence:
Meanwhile, AI conferences are thick on the ground:
I have a feeling that's one reason why I've been hearing that non-AI conferences are seeing less attendance. AI is a gold rush, like the one that left ships abandoned in San Francisco's harbors in 1849 (the image above).
So I'm here to call your attention to three events next week that will all deal with AI, but in a deeply pragmatic way.
Chief among those is IIW, the Internet Identity Workshop, which has been going since 2005 and is about whatever participants want to work on. It's an open space unconference, with no booths, speeches, or panels. It's about what people want to work on and move forward. Wikipedia describes it pretty well. No doubt lots of the breakouts will be on AI, especially the personal kind. (Expect at least one from Liam Broza and Companion Intelligence. If you're going to be there, I'll add you and your work to this post.)
We have two IIWs every year, and the next one will be our 43rd. It happens next week, Tuesday to Thursday, 3 to 5 November, at the Computer History Museum in Silicon Valley.
The day before, Monday the 2nd, is VRM Day, where we'll be focusing on MyTerms, personal AI, and much more. Love to see you there too.
And the day after, Friday the 6th, is AIW, the Agentic Internet Workshop. Same format as IIW, but just one day and about agentic whatever (literally: you bring the topic, just like at IIW).
So register and mark your calendars.
ChatGPT's response to "I need an image of the World Wide Web as a fossil, to use under the headline 'The Fossil Present.'"
The Web isn't dying, but it is being removed from our experience. As a place to seek knowledge and connections, it is being replaced for many (including me) by AI chatbots that harvest findings from it, rewrite and summarize what they find, and—as gravy on that steak—point to sources it finds there.
To some degree, this disincentivizes publishing on the Web. Chatbots remove readers and viewers, discourage advertisers and subscriptions, and tell writers that the Web is less worth their effort than it was in the past.
Also, as the sum of information harvested to train and inform chatbots approaches a virtual Everything, it almost seems as though the collective provisioning of human knowledge is complete—or near enough.
Of course we'll continue to publish, because we are curious, thinky, and chatty animals. But the land rush that gave us forty zillion websites (seriously, about 210 million, according to Netcraft) actually ended around 2011 (Netcraft again), and has pretty much plateau'd since then:
Source: Netcraft April 2026
If the blue or red line starts to drop, I am sure that will be because agent-to-agent connections between buyers and sellers push aside the old roles of websites for publishing goods, and for buying them. Available inventories will be published in other media, using AI agents, while customers will shop and buy through agents of their own, which will peruse those same media, and make smart deals. We might still call this new matrix The Web, but its nature will be different.
We always mischaracterized the Web with mixed real estate and publishing metaphors: pages we'd author and post in sites we'd construct or build at locations on domains that we'd visit and browse, as if these were the solid and durable occupants of a world wide library.
But in fact the Web is a whiteboard. And it's much easier on the Web to erase stuff than to produce and keep stuff up.
While I love the Internet Archive and consider it the most saving grace ever invented, it's a cemetery with glass-top graves. What's there is done. It's of near-infinite value, but not new. And new is what keeps the Web alive.
I don't know if chatbots are to the Web what the Chicxulub event was to dinosaurs. But I fear that it might be. I just hope the resemblance is a small one.
Stormy Whether
GDPR services is a market we shouldn't need at all, but it's growing like a thundercloud. (Source: Mordor Intelligence)
Toward Agentic VRM
vCon is "A new IETF open standard defining how conversation data — audio, video, text, and metadata — is structured, signed, and exchanged across any system. The foundation of VCONify." More:
A vCon (Virtualized Conversation) is a standardized JSON container for all data related to a real-time human conversation. Think of it as the universal format for conversations — the way PDF standardized documents or MIME standardized email.
One vCon carries everything: the audio recording, video stream, text transcript, participant identity, timestamps, derived analytics, and a cryptographic signature that proves the record hasn't been altered.
It is platform-agnostic, vendor-neutral, and AI-ready by design.
A Brake on the Breaks
ReverseAlignment says,
Most of the effort and capital in AI goes to two tasks: making systems more capable, and aligning those systems with human values.
Both matter. But even a perfectly aligned system will underdeliver, or do real harm, if it is released into economies, governments, and information systems unprepared to receive it.
We call that neglected work reverse alignment: the deliberate redesign of institutions, norms, skills, and governance so that societies can absorb AI safely and share its gains widely.
They have a coalition.
Frontiers of Importance
[Agents of Chaos](Agents of Chaos) is an academic study abstracted this way:
We report an exploratory red-teaming study of autonomous language-model-powered agents deployed in a live laboratory environment with persistent memory, email accounts, Discord access, file systems, and shell execution. Over a two-week period, twenty AI researchers interacted with the agents under benign and adversarial conditions. Focusing on failures emerging from the integration of language models with autonomy, tool use, and multi-party communication, we document eleven representative case studies. Observed behaviors include unauthorized compliance with non-owners, disclosure of sensitive information, execution of destructive system-level actions, denial-of-service conditions, uncontrolled resource consumption, identity spoofing vulnerabilities, cross-agent propagation of unsafe practices, and partial system takeover. In several cases, agents reported task completion while the underlying system state contradicted those reports. We also report on some of the failed attempts. Our findings establish the existence of security-, privacy-, and governance-relevant vulnerabilities in realistic deployment settings. These behaviors raise unresolved questions regarding accountability, delegated authority, and responsibility for downstream harms, and warrant urgent attention from legal scholars, policymakers, and researchers across disciplines. This report serves as an initial empirical contribution to that broader conversation.
The list of co-authors is half as long as that paragraph.
So I'll wait
I want a Mac Studio for M5-grade personal AI and 16 TB of storage for my massive photo file collection. Wired calls it "only a preview of what's to come." Only $18,299.00.
So I'm not
The Observer says, "AI giants are trying to sell us the apocalypse. Don't buy it." Subhead: "Corporate spiel about the dangers of the technology is paradoxically just a way of marketing it as a transformative power of world-changing significance."
First, and most importantly, don't miss Peter Kaminsky's talk today at 4 PM Eastern.
Required Viewing
I EXPOSED what big tech knows about you by becoming… a data broker, by Chris the Producer, on YouTube. It's brilliant. More about him here. And here. And here. Hat's off to Don Marti for the find. Bonus link.
Because data edges (yours, mine) will do more for each and all of us than data centers alone
The AI Data Center Bet Is Collapsing - And Asia Is Quietly Building the Lifeboat, by GD Naidu. Actually, it's two kinds of data centers, one built under failing regulatory capture (the US model) and the other built under central political control (the Asian model). And the latter isn't really a lifeboat (or choice of lifeboats) for the former.
Soon as we skirted that ridge, the skies opened. Then it got colder. A lot colder.
Riding wide
On the road for a while. Above is a storm coming over Elk Country in north-central Pennsylvania. After that, Elk Country, Ohio, and back to Indiana, by RAV4.
Didn't know there was one, but yeah
Chick-fil-A takes a stand against fast-food drive-thru AI trend.
Maybe
This is the flyer for the next talk in our speaker series at IU. It's on Tuesday, and a can't-miss event. I explain more below.
I know a lot of people who know a lot about AI. Or seem to. Or say they do. But Pete Kaminski is the real deal. More importantly, he knows more about putting AI to work for you than anyone else I know. If you want to know how to get personal—not personalized—AI working for you, Pete's your guy.
His talk is on this coming Tuesday, 20 October, in the Student Building, room 150, at Indiana University. And on Zoom. Go here to register and tune in. (See the Join Event button.) It's participatory, so be ready to ask questions.
The Dune fans among us should grok the headline, but in case not, mentats are superintelligent humans. These were handy to have if AIs are outlawed, as they were in the Dune universe. For AI knowhow, Pete is my mentat, even if nobody in reality fits the description.
Big thanks to the Ostrom Workshop and the Hamilton Lugar School at IU for their help and support in this, our fifth year of the series—and for all their earlier efforts as well.
See ya'll there!
Nikki D & Sisters of Thunder at the 2026 Lotus Festival
The idea was to get my photos from the latest Lotus Festival up right after the three-day event ended on Sunday the 4th. I had too many other things to do, so a week went by. But at least my shots of the last two performances are up. This one is for Nikki D & Sisters of Thunder, who describe their show as "sacred steel gospel and blues." Fun group. Singing and musicianship are first-rate. Great personalities. Very charismatic. Dancing is required.
This one is for Gadan, which does "Irish trad and bluegrass with the soul of Celtic and Appalachian roots music - all spiced with unmistakable Italian flair!" Here's that photo album:
Gadan at the Lotus Festival
Another great band. Loads of fun. Great music. Can't sit to them either.
Here's another thing. All the photos in both albums are Creative Commons licensed CC-BY-4.0: there for anyone to use, requiring only photo credit.
I invite both bands, the Lotus Festival folks, and any member of Bloomington's news commons to have at them as well.
Here's more about Lotus. It's an amazing event. If you like great music from everywhere, you'll love the show. It'll be back next October.
I'm in Chicago for the Marathon and our son's birthday (he's running tomorrow). Also hanging with relatives and enjoying the city, which always feels to me like a giant Boston. (Color me wrong, it's okay.)
Meaning it was fully yours
Steven Levy: I'm Still Convinced the iPod Was the Perfect Gadget—Apple's original music player got many things right. Chief among them: It didn't connect to the internet.
AI isn't just a "surf's up" moment. It's a tsunami. You have to surf that tsunami, just to keep from drowning.
Using your time. Your calendar is your surfboard.
What's your best beach for riding that calendar?
Well, right now there are countless conferences with "AI" or "Agentic" in their titles. Yes, they're beaches. You can ride there.
But where will you get the most done? Where can you write the agenda, and work with other people dealing with the same problems or pursuing the same opportunities? Real people doing real work. People who can help.
Really, do you want to hear yet another CEO blather about what their VCs and marketing people want them to say?
Or do you want to get work done?
IIW was made for getting things done. It has been a wide beach for surfing big tech waves since 2005, twice a year at the Computer History Museum in Silicon Valley. The next one is our 43rd. It's on November 3rd to 6th, less than a month from now.
We immodestly believe that IIW is the most leveraged conference in the world. Two reasons:
- Our focus (without limits) is on identity, with lots of real-world results. OAuth, OpenID, passkeys, self-sovereign identity (SSI), decentralized identifiers (DIDs) UMA, SSI, VRM, and Personal AI—to name a few—would not be the same without countless forward-moving conversations and work at IIWs.
- What we don't have. No keynotes. No panels. No agenda other than what participants choose.
And that's you. IIW is what you make of it.
So come join us. Register here.
You'll leave wet but a lot more fit.
See you there.
