EPISODE 54

Kevin and Jason break down why Anthropic is out of compute, why that's actually a strategy, and what it means for everyone using Claude right now. They dig into the Mythos model as the best marketing moment in AI, why artificial scarcity works, and why $200/month for Claude Max is the cheapest hire you'll ever make. Then they shift to AI in the enterprise — why one "AI Week" won't rewire your company, why Anthropic's one-person growth marketing team is the new bar, and the playbook for founders selling into big companies: pick a hard enough wedge, stop selling the product, and sell transformation instead.

CHAPTERS

00:00 – Cold Open: The Mythos Model Is Too Good To Release

00:35 – Welcome to Founder Mode

00:59 – Why Anthropic Is Out of Compute

02:28 – Good Customer, Bad Customer: Who's to Blame?

04:00 – $200/Month Is the Cheapest Hire You'll Ever Get

05:31 – Token Maxing and the New Scarcity

07:01 – How to Fire Bad Customers (And Why Anthropic Is Doing It Wrong)

10:50 – The Mythos Model: The Best Marketing Moment in AI

13:19 – AI in the Enterprise: Why One AI Week Isn't Enough

18:25 – The "One More Prompt" Flow State

19:57 – Selling Into Enterprise: Pick a Hard Wedge, Sell Transformation

22:48 – Closing Takeaways and Top Five

LINKS

Stay Connected with Founder Mode

Subscribe to our newsletter

Connect with Kevin

LinkedIn • X/Twitter

Connect with Jason

LinkedIn • X/Twitter

EPISODE 53

In this episode, Kevin and Jason sit down with Shain Noor, co-founder of Silvia, an AI-powered personal CFO built to help people reason through financial decisions, not just track them. Shain explains why the entire history of personal finance apps has focused on clicking and aggregating data rather than helping users actually decide what to do, and how Silvia uses Anthropic-powered agents with a verification layer to deliver trustworthy, personalized financial guidance. The conversation covers the co-pilot vs. autopilot distinction, the surprising discovery that users ask Silvia things they'd never tell their human financial advisor, how proactive alerts like the daily summary email drove retention, and why building the reasoning layer first, before adding any execution or action capabilities, is the right foundation for trust.

CHAPTERS

00:00 - The judgment-free financial advisor

02:38 - Introducing Shain Noor and Silvia

03:51 - Why finance apps have always missed the reasoning layer

05:51 - Co-pilot vs. autopilot: trust, transparency, and guardrails

08:29 - What surprised Shain: users sharing what they hide from their advisors

12:42 - Measuring retention and the proactive alerts breakthrough

17:02 - Team size, the ProCap merger, and competing with legacy finance

19:41 - The future: everyone becomes a manager of AI agents

LINKS

Connect with Shain Noor

Silvia • LinkedIn • X/Twitter

Stay Connected with Founder Mode

Subscribe to our newsletter

Connect with Kevin

LinkedIn • X/Twitter

Connect with Jason

LinkedIn • X/Twitter

EPISODE 52

Jason Shafton and Kevin Henrikson unpack where AI is genuinely useful and where it starts to create more noise than leverage, using examples from AI email triage, long chat memory drift, and agentic workflows. Kevin explains how memory can become polluted when models start treating their own prior inferences as fact, including a prompt he used to compare what an AI thought was “ground truth” against what he had actually told it. From there, the conversation shifts into a practical framework for building AI systems and human teams the same way: define the job, provide the right tools and access, layer in review and guardrails, and judge success by whether time spent together compounds into more output. They close by connecting startup hiring, high-agency operators, and founder-led culture back to the same core test they use for AI: does this person or tool create leverage, or does it create drag?

CHAPTERS

00:00 – AI memory drift and false “ground truth”

01:24 – Testing AI email triage and the risks of over-filtering

03:13 – Good AI versus bad AI in real workflows

05:31 – Why controlled memory leads to more consistent AI outputs

08:29 – How to apply AI to workflows that currently rely on humans

11:12 – Building multi-agent content systems with clear roles and QA

13:40 – Hiring high-agency people for early-stage teams

16:01 – The “pick up the shovel” standard for startup operators

22:36 – The real test for both employees and AI: leverage or drag

26:16 – Founder Mode Top 5 Takeaways

LINKS

Connect with Kevin Henrikson

LinkedIn • X/Twitter

Stay Connected with Founder Mode

Subscribe to our newsletter

Connect with Kevin

LinkedIn • X/Twitter

Connect with Jason

LinkedIn • X/Twitter

EPISODE 51

Courtney Spritzer breaks down how she built, scaled, and monetized a community-first business by starting with conversations instead of a business model, and why most founders confuse audiences with real communities. Drawing on her journey from launching a social media agency to co-founding Entreprenista, she explains how trust and engagement—not follower count—determine whether a community actually works, and how to measure that through real outcomes like connections, clients, and visibility. The conversation covers practical approaches to monetization through membership tiers, founder-led power groups, and events, as well as why IRL experiences are a powerful growth engine. Courtney also shares how she evaluates opportunities as an investor, how she maintains authenticity while scaling, and why founders should build around their strengths rather than chase trends like AI.

CHAPTERS

00:00 – Following vs. community: the core distinction

04:49 – From agency to Entreprenista: turning audience into business

08:23 – How to measure if a community is actually working

10:56 – Monetization: tiers, power groups, and testing models

13:15 – Transitioning post-exit and doubling down on community

14:22 – IRL events as a growth and engagement engine

17:50 – Investing and evaluating founder-led opportunities

20:25 – Building in the AI era vs. staying human-first

LINKS

Connect with Courtney Spritzer

Entreprenista • Entreprenista LinkedIn • X/Twitter • LinkedIn • Instagram

Stay Connected with Founder Mode

Subscribe to our newsletter

Connect with Kevin

LinkedIn • X/Twitter

Connect with Jason

LinkedIn • X/Twitter

EPISODE 50

Kevin Henrikson and Jason Shafton unpack the reality of working with AI agents, why they feel more “broken” than chatbots when they fail, and what it actually takes to make them useful in real workflows. They explore the shift from prompt-based interactions to autonomous systems with memory, triggers, and recurring tasks, and why expectations are often misaligned with how these systems behave. The conversation dives into the importance of guardrails, human-in-the-loop review, and treating AI like a junior employee rather than a perfect operator. They also cover the emerging dopamine loop of working with AI, how it’s changing the way people think and work, and why communication—not technical skill—is becoming the key differentiator in an AI-driven world.

CHAPTERS

00:00 – Why AI agents feel more broken than chatbots

02:38 – Harness engineering, workflows, and expectations

05:50 – AI agents as employees and human-in-the-loop systems

07:25 – The dopamine loop and changing how we work

13:54 – The future of work and communication as the edge

LINKS

Connect with Kevin Henrikson

Website • LinkedIn • X/Twitter

Stay Connected with Founder Mode

Stay Connected with Founder Mode

Subscribe to our newsletter

Connect with Kevin

LinkedIn • X/Twitter

Connect with Jason

LinkedIn • X/Twitter

EPISODE 49

Nima Keivan joins Kevin Henrikson and Jason Shafton to break down what it takes to move AI-built software from demos into production. Drawing on his background in robotics and autonomy, Nima explains why the real challenge is not generating code but closing the “autonomy gap” between what a system can do reliably and the messy corner cases humans still have to carry. He unpacks why durable software starts with requirements, how his team approaches PRD-driven development and scenario testing, why just-in-time mocking matters when automations touch live enterprise systems, and where natural-language software building is already working versus where full end-to-end autonomy is still not ready.

CHAPTERS

00:00 – The autonomy gap between demos and production

03:15 – What robotics teaches AI builders about reliability

07:32 – Why code generation is not the real bottleneck

13:18 – How Durable turns operator workflows into production software

26:32 – When natural language can actually replace writing code

LINKS

Connect with Nima Keivan

Durable • LinkedIn • X/Twitter

Stay Connected with Founder Mode

Stay Connected with Founder Mode

Subscribe to our newsletter

Connect with Kevin

LinkedIn • X/Twitter

Connect with Jason

LinkedIn • X/Twitter

EPISODE 48

Dennis Pilarinos joins Kevin Henrikson and Jason Shafton to unpack what AI in software development actually looks like beyond the demos, arguing that the real bottleneck is not code generation but context. Drawing on his experience building Buddybuild, working inside Apple, Amazon, and Microsoft, and now leading Unblocked, Dennis explains why source code alone is not enough for either engineers or AI agents to succeed, how historical decisions buried in Slack, Jira, and docs shape production-safe software, where AI-generated code breaks down in legacy systems, what traits matter most when hiring in an AI-native era, and why the workflows teams rely on today, including pull requests, may look very different in the near future.

CHAPTERS

00:00 – Why AI coding tools fail without context

04:03 – Building Unblocked to solve the missing layer in engineering

12:11 – Where AI-generated code breaks in real production environments

20:33 – Hiring for curiosity, ownership, and success in an AI-native world

27:18 – Founder lessons from aviation, uncertainty, and staying grounded

LINKS

Connect with Dennis Pilarinos

Unblocked • LinkedIn • X/Twitter

Stay Connected with Founder Mode

Subscribe to our newsletter

Connect with Kevin

LinkedIn • X/Twitter

Connect with Jason

LinkedIn • X/Twitter

EPISODE 47

In Episode 47 of Founder Mode, Kevin Henrikson and Jason Shafton unpack why AI agents should no longer be thought of as simple tools, but as a new kind of workforce that founders can hire, coach, evaluate, and orchestrate. They explore how the founder role is shifting from building and doing toward designing systems, managing agent workflows, and making the judgment calls that still require human trust, taste, and strategic thinking. Along the way, they discuss AI-first operating habits, what this means for hiring and team design, why small teams may become dramatically more powerful, where human oversight still matters most, and why the real moat in an AI-native world is not the agents themselves but distribution, proprietary workflows, and unique data.

CHAPTERS

00:00 – The founder’s job has changed

00:32 – AI agents as a new staffing model

06:29 – What AI-first workflows look like in practice

16:48 – Hiring and scaling in the age of agents

24:06 – The new moat: data, workflows, and distribution

LINKS

Connect with Founder Mode

Founder Mode

Stay Connected with Founder Mode

Stay Connected with Founder Mode

Subscribe to our newsletter

Connect with Kevin

LinkedIn • X/Twitter

Connect with Jason

LinkedIn • X/Twitter