Wednesday, June 3 was the fourth meeting of the Dallas AI Club. Near the end of the night, one of the latecomers walked in, listened to where the conversation had gone, and dropped a line that oddly summarized much of the evening: never doubt the power of organized idiots.
"Never doubt the power of organized idiots."
I smiled as I thought about the line as a marriage of George Carlin’s never underestimate the power of stupid people in large groups and Margaret Mead’s famous line about small groups changing the world. Perhaps this is the new frame for the DAIC?
(Also: A member warned me that it is too soon to use acronyms for the group. I post this first one in loving defiance.)
The line landed in our room because it named what we had all been feeling. Two members had already confessed, openly and without any prompting, that they felt completely stupid in this room compared to some of the others — not because they actually were, but because their knowledge of technology was like a child’s compared to others in the room. Three of the members were 30+ year technologists whose careers began in Silicon Valley in the 90s; these other two had spent those same years working in the nonprofit sector, not in rooms full of AI agents, APIs, and hardware comparisons.
As one remarked, the distance between can build this and can barely keep up with the announcements is wide enough that it can feel like a different species across the table from you. It named what the room actually was: not a brain trust, not a council of experts, but a stubborn little gathering of (slightly) organized idiots who keep showing up.
And the lift the line carries: never doubt them.
Most of the meeting before that line was a working training session led by one of the more technical members of the room. We have started inviting members to give walkthroughs of how they actually use the tools, and this was the first one. He didn’t just describe his stack — he showed it, screen by screen. And his numbers were beyond impressive: in fact, over the past months, he has built out a suite of automated agents that have applied on his behalf to over 1,200 jobs! The agents track each application, summarize his inbox, draft tailored LinkedIn messages to specific people in his network to request referrals, and, once an interview actually happens, record it, feed the audio into Claude, Gemini, and ChatGPT in parallel, and ask each model to give him a ruthless critique of his answers so he can do better the next round. For one application at a major media company, the agents ingested his entire multi-decade career history and produced a custom infographic, a tailored cover letter, and a personalized video pitch. The whole machine ran without much of his hands on the keyboard.
And that was just the job search. He also walked us through NotebookLM as a deep-research engine, showing how it could synthesize a prospective employer’s entire public footprint into a usable dossier, generate a podcast-quality conversation between two AI voices to give him something to listen to on the company, and turn a messy pile of notes into a structured brief to prep for his interviews.
The capability on display was, frankly, a little dizzying. Several of us were taking notes on what to try later. A few were going further than that, logging in and opening accounts in real-time without waiting for the session to end. One member got so taken with what he was seeing that he created his NotebookLM account on the spot, and before the walkthrough was finished he had started not one but five separate projects in it (interestingly, he was one of the two who had previously described himself as being unable to speak their language; how quickly we can catch each other up!).
And then, as we have been doing since the first meeting, we vocalized the underlying question: as these things scale, what is left for a human being to actually do?
There was also an elephant in the room that had to be named: Why has the person giving the walkthrough not been hired after seven months of agentically applying to over 1,200 jobs?
He is eminently qualified — over thirty years of results, a recognizable stack of employers, and the technical chops to have built the entire agentic apparatus that has been doing the applying. And yet … he is now well into the latter half of his career.
And he is not the only person in the room actively hunting either; at least two of us are, and the conversation turned briefly to ageism, which is real and not subtle. Another member said he had seen a post on LinkedIn just days before in which a company was bragging openly that it was only hiring programmers under thirty — partly because, if you read between the lines, they could be paid less. The Harris segment we would watch later in the night was about labor replacement in the macro; this is what labor replacement already looks like in the room — experience valued less than profitability.
And then the other side of the room spoke up.
Several members admitted, without preamble, that they had been struggling. One had not opened his computer all week, despite a long list of things he had meant to build. Another said he had mostly been using AI as a stand-in for a therapist, because he is going through a hard stretch and is having trouble getting out of his house. A third described the feeling that the more the tools can do, the harder it gets to decide which thing to start. Opportunity paralysis — a phrase the room first put a name to in our second meeting and has come back to in every gathering since — has a real cost, and that cost was sitting in the room with us in the form of people who genuinely want to be building and cannot quite make themselves do it.
Nobody tried to talk anyone out of it. There were no productivity hacks offered, no have you tried this app. The room just held it. And it was right at that moment — the stack on one side, the can’t-open-the-laptop on the other — that the late-arriver’s line cut through. Never doubt the power of organized idiots. What he meant was that showing up matters more than knowing what you are doing, and that the group itself, with all its messiness, is the unit that actually carries any of us through this.
At the top of each hour, we held the memento mori meditation as we always do. Knowing we will die, how do we want to live the next hour? The question lands a little differently each week depending on what the night has stirred up. On Wednesday it landed inside the gap — the gap between agents that can apply to 1,200 jobs in someone’s name and the actual question of what someone wants his life to be for.
In the latter half of the meeting, we watched a CNN segment that had aired the day before. The headline was Anthropic filing for an IPO at a valuation that may exceed a trillion dollars — one of three AI-company listings expected this year that will, between them, reshape what public capital looks like. The segment was not just about money. It included an extended interview with technology ethicist Tristan Harris, who has been making a specific and very careful argument for the last several months that the rest of us cannot quite shake.
Harris’s argument, in plain terms: the AI companies have taken on so much debt and infrastructure cost that the only economic outcome that justifies their valuations is the replacement of most human labor. A near-trillion-dollar promise can only be paid back by capturing the global labor economy, which Harris pegs at around fifty trillion dollars. The reason these companies are not designing AI as a tool to augment human work but as a system designed to replace it is that augmenting human work does not, mathematically, return the investment. He calls the trajectory we are on “an anti-human future.” He is careful to say it is not inevitable; he is also careful to say that, given the incentives currently in place, it is the path the financing is paying for.
The room sat with this differently than other warnings have landed. Partly because the person making the case is not a doomer — Harris has been one of the more grounded public voices on these questions for years now, and he is not pleading apocalypse. He is naming a structural fact about the money. And partly because we had just spent an hour looking at exactly the kind of capability that, when you scale it across an economy, is what he is warning about. The agents that can apply to 1,200 jobs are the same agents that, scaled, can replace 1,200 hiring teams. The capability we were celebrating in one breath is, in the next breath, the thing he is describing.
The conversation that followed was the same fork the room has now reached in three out of four meetings:
-
Chase the wave: build the agents, build the businesses, ride the trillion-dollar IPOs to whatever security can be had on the way down.
-
Or step out of the water: find a piece of land, learn to grow food, form an intentional community, become as un-replaceable-by-an-agent as it is still possible to be.
Most of us are walking around with both impulses live at the same time, and the contradiction does not get easier with discussion.
One member pushed back on the binary the room had set up. Do we actually have to choose, he asked — why can’t we build the business and also remain deeply human? Another answered, slowly, that of course we can; but while AI is now powerful enough to let us do anything, it does not give us the time or the energy to do everything. We still have to choose one thing — one direction — to point our finite attention at. The choice is not between living a human life and building with the tools. The choice is in which of the thousand possible somethings we will actually invest our limited time, energy, and lives.
And then the meditation chimes dinged, and we all held our silence behind closed eyes as we pondered that inquiry.
What we landed on, in the end, was not really either side of the original fork. We landed on the room itself — on organized idiots — and the slow, unfashionable idea that the answer might not be optimization in either direction but the much messier work of being in a room with other people walking each other home in the dark and the fog.
We do not have the answers. We feel stupid half the time. We get overwhelmed. We sometimes cannot open our laptops for a week; and when we do, it might only be to chat with an agent about our depression.
But we keep coming back anyway. That is not a strategy; it is a purpose … to be here for one another.
The capability on display in this room is genuinely staggering. The struggle is just as real. And the only thing I have seen, over four meetings now, that touches either is the room itself — the chair, and the person in the chair, and the willingness of every person in the chair to admit out loud how they are actually feeling in the face of all of this.
The organized idiots meet again next Wednesday.
If you are not yet in the room and want to be, you can request an invitation.
— Jeremy