Oct 19, 2026 · 2h 24m
Theranos debate meets AI’s test for real-world value
"You Can See Everything" Sparks Debate, 17% GDP Growth Please, Call of Duty Campaign Review | Katie Drummond, Bing Gordon, Daksh Gupta, Jean-Denis Greze, Quinten Farmer
The episode connects founder mythology, documentary storytelling, and AI’s productivity claims to a shared question: what survives scrutiny when ambition meets reality.
- 1The Theranos film complicates Elizabeth Holmes without erasing the scientific and institutional failures behind the company.
- 2AI infrastructure and coding-agent spending are accelerating, but visible gains in products and software remain difficult to measure.
- 3Bing Gordon and AI founders argue that trustworthy products still require iteration, human judgment, privacy, and meaningful memory.
Don't miss
Katie Drummond explains how a second viewing and additional reporting changed her interpretation of the Theranos film, holding sympathy and condemnation together.
The brief
Katie Drummond revisits Nathan Fielder’s Theranos film, weighing its empathy for Elizabeth Holmes against the company’s scientific failures and the media culture that amplified her.
The hosts widen the argument from Theranos to AI, questioning whether expanding data centers and coding-agent budgets are producing measurable changes in software and the broader economy.
Bing Gordon frames product design as a discipline of iteration, testing, and humility, while arguing that AI can expand game possibilities only when humans impose direction and coherent systems.
Daksh Gupta, Jean-Denis Greze, and Quinten Farmer describe different answers to AI’s trust problem: validate generated code, protect personal boundaries, and make memory useful without making assistants intrusive.
The episode’s throughline is a warning against charismatic shortcuts: whether in Theranos, software, games, or personal AI, impressive promises still need evidence and judgment.
What was said on this episode
43 statements · 27 positive · 12 negative · 2 mixed · 2 neutral
Elizabeth Holmes was a poor founder, independent of Theranos’s fraud.
“it seems like she's just a bad founder. Like put aside the fraud. Like it just seems like, like total slop in terms of the execution of that business and gives me absolutely no faith in a Theranos 2.0.”
Listen at 2:55
Over 60 medical-device companies sold for more than $100 million since Theranos began.
“there have been 60 plus companies since Theranos started that sold for over a hundred million dollars in that category.”
Listen at 3:41
FDA approval for medical devices can take a decade.
“It takes a decade for the FDA to actually approve it.”
Listen at 4:43
You Can See Everything was an immediate box-office and cultural phenomenon.
“it looked like we were experiencing like an absolute box office smash.”
Listen at 10:53
Online comedy ability does not indicate strong founder ability.
“Comedy on the timeline does not mean you're a great founder. Often they're not correlated at all.”
Listen at 13:41
Roche’s Cobas Vivo DX MRSA test received FDA authorization in 2019.
“in 2019, it got authorized for. The Cobas Vivo DX MRSA test”
Listen at 19:27
Theranos founders did not sell secondary shares despite an $11 billion valuation.
“No secondary selling. $11 billion.”
Listen at 22:02
Higher rates and worsening markets could delay or stall some data-center projects.
“Maybe in months, if things continue to get bad and the rates go up, some projects will be delayed or stalled.”
Listen at 25:38
Personal agentic AI will generate high demand for compute.
“the agentic stuff, the personal agentic stuff is going to be high demand for compute.”
Listen at 25:56
LLM interfaces provide a superior consumer experience for browsing the open web.
“I love interacting with the open web through an LLM. It's amazing. No cookie pop-ups, no ads. It's truly like the best consumer experience.”
Listen at 26:49
Enterprise and B2B AI spending must increase tenfold to support AI infrastructure.
“It has to 10x.”
Listen at 30:09
John Coogan wants annual GDP growth to reach 200 percent.
“I'd put it at 200%. GDP growth. I want to be tripling every year.”
Listen at 36:35
The AI-generated product economy is currently thriving.
“the ai slop economy is thriving right now”
Listen at 39:20
John Coogan positively evaluates the new Call of Duty campaign.
“I've been thoroughly enjoying it. I haven't played it fully, but it's amazing.”
Listen at 41:51
Sympathy for Elizabeth Holmes can coexist with judging her a terrible fraud.
“You can feel really sad in that moment and still think this person is a terrible fraud. Both of those things can be true at the same time.”
Listen at 57:25
Katie Drummond predicts You Can See Everything will become a mainstream classic.
“I think this will be a classic. I don't know about the cult part. I think this will be a mainstream classic.”
Listen at 1:00:22
Elizabeth Holmes exaggerated startup claims more extremely than was common in her era.
“I think she did it to the nth degree and she did it in an era where that was so much more commonplace.”
Listen at 1:01:10
Elizabeth Holmes was treated differently because she was a woman.
“I do think that Elizabeth Holmes was in some ways treated differently because she was a woman.”
Listen at 1:04:04
Katie Drummond considers You Can See Everything an excellent film.
“I think it's a great film. I think it's so good.”
Listen at 1:05:53
Bing Gordon considers Isaacson’s Elon Musk biography the best technology biography.
“That's probably the best one. It's probably the best one because he was allowed inside access.”
Listen at 1:08:59
Mobile and social platforms added one billion monthly players who do not identify as gamers.
“we also got a billion new people playing games every month that still don't call themselves gamers.”
Listen at 1:17:05
An overconfident, wrong board member can seriously damage a company.
“A single board member like that can wreck a company.”
Listen at 1:29:15
Board Game Lab is developing an AI board-game generator.
“I think the AI, I'm working with two guys who are doing an AI board game generator, and one of them has been a successful board game entrepreneur. This is called Board Game Lab.”
Listen at 1:32:21
AI will eliminate much of the tedious work involved in game development.
“I think we're going to see a lot of the grunt work that makes people not like working on games go away.”
Listen at 1:33:17
Game designers should select useful AI hallucinations and integrate them into game systems.
“you don't mind saying, hey, AI, hallucinate your butt off. And let me take the good stuff and then let me put some game systems in it.”
Listen at 1:38:18
Agentic game-development projects take longer than expected to reach good products.
“from shitty first draft to good product, it's taken a year longer than anybody expects.”
Listen at 1:40:41
Gamers prioritize social meaning, mastery, and then productivity.
“the Maslow's hierarchy of needs for gamers starts with social meaning, you know, be seen, matter. Number two is self-improvement, mastery being the flow. So what you're talking about to me sounds like the addiction of leveling up.”
Listen at 1:43:28
Greptile aims to automate validation ensuring code changes are safe to merge.
“We are trying to automate code validation, which to us means that we want to make it so that you can review and test any code change and for sure know it is safe to merge.”
Listen at 1:47:43
Companies will want independent central validators for code generated by multiple agents.
“they'd want an independent and central code validator for all the code to go through.”
Listen at 1:50:13
Human tolerance for faulty software will not increase as code production accelerates.
“the human tolerance for faulty software is not going to go up.”
Listen at 1:51:52
Coding-agent adoption has produced modest software quality and quantity gains at existing companies.
“there just hasn't been that much of an increase in the quality nor the quantity of software at the same companies.”
Listen at 1:54:28
Most people will access digital information through assistants rather than apps.
“the way most humans will be interacting with digital information is not going to be through apps in the future. It's going to be through assistants”
Listen at 1:57:52
People should use dedicated work assistants with strong data privacy and security controls.
“it makes sense for people to have a work assistant. that is really focused on work tasks and has that entire data, privacy, security, collaboration story handled perfectly.”
Listen at 1:58:27
Personal assistants must earn user trust before becoming a universal information interface.
“you have to earn the trust to become the one glass pane for people's universe.”
Listen at 2:03:46
Agents can create value by identifying opportunities for users to save money.
“when the agent finds an opportunity for you to save money.”
Listen at 2:05:25
High-quality proactive agent recommendations require substantial upfront computation.
“you have to pre-spend to make those kinds of recommendations work is really, really hard.”
Listen at 2:08:06
Tolan users engage more and value the product more after model improvements.
“Every time we ship a new model to users or upgrade some part of the background agent that we have, people engage more, people value it more.”
Listen at 2:12:20
Technologists often mistake newly available capabilities for standalone products.
“we mistake new capabilities for new products.”
Listen at 2:13:35
Muse will succeed strongly among Facebook’s small-business users.
“I think Muse is just going to crush it with the however many millions of small business owners are on Facebook.”
Listen at 2:14:07
Tolan’s proactive outreach should follow topics users first raise.
“It should always be downstream of something that you have first brought to your Tolan”
Listen at 2:15:43
Tolan’s cute, safe branding reduces concerns about AI.
“starting from this place of, hey, this is cute, safe, really helps with that as a concern.”
Listen at 2:18:30
Tolan allows parents to monitor children’s use of the assistant.
“the parent can monitor. their use of the tollen”
Listen at 2:21:09
Tech entrepreneurs often lack expertise in distributing physical products to parents.
“the channels through which you get parents to buy like a physical good are just not things that many tech entrepreneurs are familiar with.”
Listen at 2:21:55
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.