May 29, 2026 · 50m · more-or-less
Normal People Don't Want AI Agents (And Silicon Valley Can't Accept It) · More or Less Podcast
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the More or Less podcast, hosts Dave Morin, Britt Morin, and Jessica Lessin are joined by guest Amir Efrati of The Information to debate the realistic ROI of corporate AI, consumer adoption barriers, national security implications, and the cultural tension between human connection and Silicon Valley's efficiency obsession.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 70.2% of the talking time here. How this is scored →
The hosts as informed peer 5.3 Guest teaching 3.0 Guest disagreement 3.2 The hosts pushing back 4.5
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 27:22 Britt defends agent utility against Jessica's skepticism
Britt directly refutes Jessica's assertion that ordinary people do not care about AI agents, passionately highlighting how moms would readily pay twenty dollars a month to reduce parental mental load.
Hardest push from the hosts ▶ 28:40 Jessica rejects the Silicon Valley efficiency dogmaJessica forcefully challenges Britt and Dave's premise, stating that technologists mistakenly assume every human wants to march toward efficiency and that superior free technology does not guarantee adoption.
Biggest teaching moment ▶ 19:31 Amir clarifies Uber executive statements and challenges cost thesisAmir corrects Jessica regarding the difference between Uber's CTO and COO comments and sharply asks Dave to reconcile his claims about cost barriers with Anthropic's massive commercial surge.
The host holds their own ▶ 20:15 Dave explains agent token multiplier economicsDave demonstrates deep enterprise expertise by explaining how enterprise agent harnesses drive a three-thousand-percent surge in token usage, reconciling massive vendor revenue with intense client cost fatigue.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| AI Voice Whispering & Prediction Banter | 3 | 2 | 2 | 4 | The conversation opens with casual intro banter and teasing. Dave pushes back playfully on Britt's claim of having predicted whispering interfaces, calling it a classic VC prediction that lacks technical understanding. | |
| AI Executive Order & The Pope's AI Encyclical | 5 | 4 | 1 | 2 | Jessica and Dave discuss the pending AI Executive Order and the Pope's encyclical. Amir brings reporting details regarding Anthropic's national security ties and internal dynamics with the DoD. | |
| Debate on Anthropic's Vatican Event & AI Marketing | 6 | 3 | 3 | 5 | Dave critiques the Vatican event as a pure marketing stunt by Anthropic, while Amir defends the Pope's rationale in having a credible AI founder present. Dave and Jessica also drill down into model binaries and White House oversight. | |
| AI Backlash, Data Center Opposition, and Foreign Influence | 6 | 3 | 2 | 4 | Dave presents his theory connecting student protests and Utah data center opposition to foreign influence operations, citing Kevin O'Leary's investigation. Amir contextualizes how the federal administration avoids heavy-handed moves that might disrupt the AI infrastructure economy. | |
| Federal Reserve Data on AI Adoption & Token Pricing | 7 | 4 | 3 | 5 | Britt cites Federal Reserve statistics showing plateauing AI adoption. Dave gives in-depth enterprise commentary on token economics and OpenClaw metrics, while Amir presses Dave on how to reconcile pricing concerns with Anthropic's meteoric revenue growth. | |
| Debate: Will Normal People Adopt AI Agents? | 7 | 3 | 6 | 7 | Jessica forcefully argues against the Silicon Valley assumption that normal people want to run personal AI agents. Dave, Britt, and Amir push back immediately, with Dave comparing her stance to rejecting personal computers in 1984. | |
| Practical AI Use Cases vs. Silicon Valley UX Deficit | 6 | 4 | 6 | 7 | Britt details practical consumer agent use cases like parenting schedules and family finance dashboards. Jessica repeatedly rejects the idea that ordinary users want or need this, leading Dave to diagnose Silicon Valley's deep UX and product design deficit. | |
| Productivity Maxing vs. Human-Centric Technology Design | 6 | 2 | 3 | 5 | Jessica criticizes the current wave of software for focusing solely on productivity maxing rather than real human problems. Dave agrees that tech has over-indexed on optimization, advocating for AI that simplifies computing and reduces wasted time. | |
| Market Penetration, Discovery, and Real-World Case Studies | 5 | 4 | 4 | 5 | Amir points out that AI is at tiny market penetration and faces a massive discovery hurdle rather than a demand ceiling. Dave shares an anecdote about a non-tech real estate professional resolving 20 years of legal files, which Jessica dismisses as an outlier corporate edge case. | |
| Defining AI Agents & Wager Banter | 4 | 2 | 4 | 5 | Britt attempts to bet Jessica on AI adoption figures, but Jessica refuses to participate in what she calls a moving target definition of an agent. Dave re-asserts that agentic interfaces will represent a multi-decade paradigm shift. | |
| AI Entertainment Evolution vs. Original Storytelling | 3 | 2 | 1 | 1 | The panel pivots to offline trips and cultural recommendations, discussing Suno, 11 Labs, Mark Rober's Crunch Labs, and Jessica's tennis-ball cannon shot. The dynamic is lighthearted and conversational. |
Statements from this episode (25)
Startup developers are increasingly whispering to AI tools in the office
“She highlighted the trend of developers whispering to their AI. So if you go into an office of a startup, you shall see many a whispering.”
US national security establishment relies heavily on Anthropic despite DoD tensions
“While you have the fight between Anthropic and DoD over these various guardrails that Anthropic was insisting on, it's still the case that much of the national security establishment relies on and loves Anthropic models, and they want to keep using them, and t…”
Anthropic exploited the Catholic Church for marketing with Pope appearance
“It's actually the story because the Pope shouldn't have had anybody on stage with him, let alone, you know, one of many of these different researchers providers. And so I really think that the church was taken advantage of, and it became a marketing moment for…”
US government will inevitably secure early access to review frontier models
“Where we seem to be going is the government eventually, one way or the other, having much deeper participation in the, I don't want to call it vetting, but certainly getting an early Access early look into what is being developed.”
AI reverse-engineering will make all software effectively open source
“I'm hearing that they can reverse engineer any software binary. And so like software is no longer open Closed source. All software will be open source going forward because you can just rip open any binary and reconstruct the software.”
Trump administration's lack of meaningful AI regulation will not last
“We're basically back to where we were, which is very little drumbeat of meaningful regulation, which I just don't think will hold.”
Forensic experts proved China is funding opposition to Utah data center
“And then he hired forensic experts to go and factually prove that it is China actually funding all of this opposition in Utah. And he actually mapped it a lot. You can go look on his Twitter and Instagram and he shares all of the data. And apparently has now s…”
Anthropic's high token pricing is driven by severe server capacity limits
“Anthropic is also up against scaling problems where they are unable to scale their downtime is It's really, really bad. You can just go look at their website and look, they've got demand through the roof, product market fit through the roof. Right. And so the …”
Trump administration is unlikely to enact economy-damaging AI regulations
“Well, the Trump administration is unlikely to do something so heavy handed that it takes the economy.”
AI and data center investments are literally sustaining the US economy
“If Anthropic were to fail, that would have a massive cascading effect on the entire economy and AI and data centers in general are what is literally keeping this economy going.”
Anthropic possesses an unreleased frontier AI model called Mythos
“We have the whole scenario now where there is a model that exists called Mythos. Nobody really has it except for Anthropic, and who knows if the actual real model as opposed to distilled versions of it that they give access to will ever kind of come out, and m…”
Federal Reserve data shows US AI adoption growth flattened recently
“Did you guys see that the fed introduced their latest data about AI adoption growth and it's actually been flat or linear, nearly flat for the last three months.”
High frontier AI model prices are driving the plateau in adoption
“And to me, the only explanation has to be price, right? Like it's just too expensive to use the frontier models.”
Founders are swapping frontier models for open-source to maintain margins
“We're also seeing founders that are running massive agentic stacks, and they're swapping out for open source models, like left, right, and sideways, because it's just like unsustainable for their margins. Like they just cannot pay these luxury prices.”
Uber COO sees little ROI after CTO exhausted annual AI budget
“So the CTO almost, almost bragged about blowing through the AI budget in the first few months of the year. Then the COO over the weekend in a podcast said that I think that this led to a bunch of internal meetings and that he, this is Andrew McDonald, did not …”
Anthropic has completely overtaken OpenAI in revenue
“So I'm just curious how you reconcile what you were saying about the costs and everything with just the unprecedented growth that Anthropic has seen and the like flippening with open AI where it has just completely overtaken open AI from a revenue perspective.”
Deploying AI agents in an enterprise can spike token consumption 3,000%
“We see this at OpenClaw, is that when you deploy agents into an enterprise, your token consumption goes up sometimes on the order of 3000%.”
Most consumers won't need a $20 per month personal AI agent
“And I actually think that it's very possible we live in a world where over the next 10 years, most people don't need to be paying 20 dollars a month for a personal agent that can do all these amazing things.”
Moms will always pay $20 monthly if AI reduces mental load
“Moms will always pay 20 dollars a month if like literally two percent of their mental load gets shaved off.”
Only single-digit millions currently use coding AI tools like Claude Code
“Nobody uses, you know, Claude code and codex in the grand scheme of the internet. There's like single digit millions people using it. They're incredibly technical.”
Current AI tool adoption is at roughly one percent of market potential
“We're like at one percent penetration of what is the market in terms of what you can currently do with these tools. Only one percent of the people are actually doing it.”
Assuming everyone wants constant productivity could be AI's ultimate undoing
“My just fundamental point is like, we cannot assume that the goal is more productivity for everyone. There are fields and people where it is, but this technology is being designed by people who think that, and I think that could be its undoing.”
Truly offline experiences will become extremely valuable amid infinite AI content
“Now that you can generate anything you want on the internet for free and that like the internet's just like full of infinity content, I do think actually these offline, truly offline experiences like are going to become extremely valuable.”