Skip to main content
My First Million

This guy cured his dog’s cancer with ChatGPT + 4 other crazy AI stories

54 min episode · 2 min read

Episode

54 min

Read time

2 min

Topics

Startups, Leadership, Design & UX

AI-Generated Summary

Key Takeaways

  • AI-powered self-treatment: An Australian entrepreneur used ChatGPT to identify a treatment pathway, AlphaFold to model tumor protein structures, and Grok to design a custom single-shot cancer vaccine for his dog. The regulatory approval process proved harder than the science itself. This framework — sequence, model, design, manufacture — is now accessible to determined non-experts.
  • Anthropic's 10x growth trajectory: Anthropic generated $6B in a single month, surpassing the full annual revenue of Snowflake and Databricks. The company has grown 10x every year for three to four consecutive years, moving from $0 to $100M to $1B to $10B. CEO Dario Amodei warns that conservative capital deployment is still necessary despite explosive growth.
  • AI transformation consulting as a business model: Anyone can build a viable business by auditing companies and implementing AI workflows for non-technical operators. The playbook mirrors the social media marketing agency (SMMA) model: pick a niche like dentists, demonstrate ROI, then replicate across 50 clients. No upfront capital is required, and demand is immediate across every industry.
  • Niantic's data pivot: After Pokemon Go declined, Niantic retained all real-world street-level imagery captured by hundreds of millions of players globally. That dataset is now licensed to AI companies building autonomous delivery robots that need real-world terrain recognition. Unique, niche datasets — not generic data — are the actual scarce resource, analogous to a specific grade of oil.
  • Service businesses repriced as software: AI is pushing service company gross margins from roughly 40% toward 75% by reducing headcount requirements. Private equity firms are actively shifting acquisition budgets away from SaaS toward service businesses, valuing them at software-style multiples. The underlying thesis, called "service as software," was articulated roughly two years ago and is now playing out in deal flow.

What It Covers

Sam Parr and Shaan Puri react to five real-world AI stories — from an Australian entrepreneur sequencing his dog's tumor DNA to cure its cancer, to Anthropic hitting $6B in monthly revenue, to Andrej Karpathy mapping AI disruption risk across all 143 million U.S. jobs — exploring how AI reshapes business, medicine, and work.

Key Questions Answered

  • AI-powered self-treatment: An Australian entrepreneur used ChatGPT to identify a treatment pathway, AlphaFold to model tumor protein structures, and Grok to design a custom single-shot cancer vaccine for his dog. The regulatory approval process proved harder than the science itself. This framework — sequence, model, design, manufacture — is now accessible to determined non-experts.
  • Anthropic's 10x growth trajectory: Anthropic generated $6B in a single month, surpassing the full annual revenue of Snowflake and Databricks. The company has grown 10x every year for three to four consecutive years, moving from $0 to $100M to $1B to $10B. CEO Dario Amodei warns that conservative capital deployment is still necessary despite explosive growth.
  • AI transformation consulting as a business model: Anyone can build a viable business by auditing companies and implementing AI workflows for non-technical operators. The playbook mirrors the social media marketing agency (SMMA) model: pick a niche like dentists, demonstrate ROI, then replicate across 50 clients. No upfront capital is required, and demand is immediate across every industry.
  • Niantic's data pivot: After Pokemon Go declined, Niantic retained all real-world street-level imagery captured by hundreds of millions of players globally. That dataset is now licensed to AI companies building autonomous delivery robots that need real-world terrain recognition. Unique, niche datasets — not generic data — are the actual scarce resource, analogous to a specific grade of oil.
  • Service businesses repriced as software: AI is pushing service company gross margins from roughly 40% toward 75% by reducing headcount requirements. Private equity firms are actively shifting acquisition budgets away from SaaS toward service businesses, valuing them at software-style multiples. The underlying thesis, called "service as software," was articulated roughly two years ago and is now playing out in deal flow.

Notable Moment

Karpathy built an AI research agent explicitly programmed to skip human feedback entirely — it generates hypotheses, conducts research, evaluates its own outputs, and loops autonomously overnight. The implication is that the next major AI capability leap may itself be produced by AI running without human direction.

Know someone who'd find this useful?

Episode Transcript

Alright. There's a lot of stuff going on, specifically AI stuff. It feels like every day, there's a new crazy story about AI. So I wanted to do a little bit with you, Sam. Crazy AI stories, you know, one crazier than the next. Well, except they get crazier every single story. I was talking to a friend, and he was like, I was trying to talk to my boss, about something. And, apparently, the boss had made, like, four or five agents that he calls employees. And the boss replied to the feedback of, like, you know, like, we need to create x, y, and z. The boss said, can you talk to my chief of staff about that? And, and so this person had to go and, like, message the chief of staff and was, like, messaging the chief of staff. And it like, he they weren't getting the answer that they wanted. And they go to the the boss again and go, hey. Your chief of staff sucks. Like, what's what's their deal? And the boss said, well, can you tell her boss, my head of operations? And so she, like, she or he literally had to go to, like, the boss of the agent to, like, say, like, can we get this person on a pip? And this whole boss, this boss of this, like, 50 person organization was, like, referring to these people as people and, like, making his real life employees interact with them. Have you done anything that extreme with AI and agents? No. But I find the whole, hey. I'm gonna have this entire, like, autonomous company of agents and then me on top thing to be so funny because it's like, who says you get to be on top? Why would why would you be why would you be the the all knowing if if AI is so good that it could do all these jobs, you think it just can't do yours? Like, why is yours so special? And it reminds me of a what's that, like, poem where it's like, first they came for the the socialist, and I did not speak out. And then they came for the for the Jews, and I did not speak out. And then they came for this. And when they came for me, there was no one left to speak out. That's that's basically what I think is happening with AI. I, Andrew Wilkinson came out with this personality test, and I took it. And it was really cool. And then listen to this. My wife and I did it together because there's, like, a thing where you could share it with your spouse. And then there's one where you could share it with your business partner. So I saw tell my business partner to do it, And he mistakenly shared the wife version that he he the and it said that, like, it talked about, like, our sexual like, are we …

Get the full transcript (11,518 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all My First Million transcripts →

You just read a 3-minute summary of a 51-minute episode.

Get My First Million summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

Books, tools, and gear mentioned in this episode

SignalCast may earn commission on purchases via these links. As an Amazon Associate, SignalCast earns from qualifying purchases.

Tools

  • by DeepMind

    An Australian entrepreneur used ChatGPT to identify a treatment pathway, AlphaFold to model tumor protein structures, and Grok to design a custom single-shot cancer vaccine for his dog.
  • An Australian entrepreneur used ChatGPT to identify a treatment pathway, AlphaFold to model tumor protein structures, and Grok to design a custom single-shot cancer vaccine for his dog.
  • by Beehiiv

    💼 SPONSORS [Beehiiv]
  • by HubSpot

    💼 SPONSORS [HubSpot]
  • by OpenAI

    An Australian entrepreneur used ChatGPT to identify a treatment pathway, AlphaFold to model tumor protein structures, and Grok to design a custom single-shot cancer vaccine for his dog.

Products

  • by Niantic

    After Pokemon Go declined, Niantic retained all real-world street-level imagery captured by hundreds of millions of players globally.

company

  • Anthropic generated $6B in a single month, surpassing the full annual revenue of Snowflake and Databricks.
  • Anthropic generated $6B in a single month, surpassing the full annual revenue of Snowflake and Databricks.
  • After Pokemon Go declined, Niantic retained all real-world street-level imagery captured by hundreds of millions of players globally.
  • Anthropic generated $6B in a single month, surpassing the full annual revenue of Snowflake and Databricks.

More from My First Million

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best Startup Podcasts (2026) — ranked and reviewed with AI summaries.

Read this week's Startups & Product Podcast Insights — cross-podcast analysis updated weekly.

You're clearly into My First Million.

Every Monday, we deliver AI summaries of the latest episodes from My First Million and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime