AI raises everyone to an 8 out of 10. Jared Tredly's read is that the premium then moves to whoever pushes that 8 to a 10, because taste is the part AI hasn't replicated. That same floor-raising is dissolving role boundaries and shrinking teams across every function, not just design.

Jared, Design Recruiter at Shopify, tells Keith how he runs a talent pipeline on a 12-month content loop instead of cold calls, why living prototypes are replacing static specs as the source of truth, and what signals top talent before a resume does.

Topics discussed:

• AI raises the floor, so top-end taste commands the premium

• The 12-month content loop that nurtures passive candidates

• Extend-yourself AI skills that compress team headcount

• Living prototypes replacing static specs as source of truth

• Work-in-public as the new top-of-funnel hiring signal

• Building room for curiosity where the best ideas win

• Creatively directing and QC-ing multiple agents by 2030

Canva runs account-based marketing across a thousand accounts, producing custom copy, images, and a personalized landing page for each, with the company name and a tagline in every asset. Emma Robinson, Head of B2B Marketing at Canva, tells Keith how product usage signals, CRM, and sales-call data feed the creative to make that scale work.

She argues leads are giving way to domain signals as the real marker of quality, and that the hard part of AI adoption is never the build. It's the final mile of getting a workflow to run consistently, and getting a whole team to actually use what one person made.

Topics discussed:

• Replacing lead-based qualification with domain signals

• Running personalized ABM across a thousand accounts

• Embedding company name and tagline in every asset

• Consolidating onto one platform to reduce chain breakage

• Separating creative direction from creative production roles

• Why AI offers optionality but not creative judgment

• AI Discovery Week where every leader builds an agent

• The final-mile problem of consistency and enablement

A year ago, TZP Group had zero AI licenses. Within days of Claude's Excel plugin launching, employees who had resisted every tool started demanding access, and that consumer pull became the firm's adoption engine.

Jarrad Berman, Operating Partner and Head of AI at TZP Group, tells Keith how he blocked Google to force his own habit change, built a skill program mapped to 18 core firm functions, and ran a CEO hackathon that had one founder writing 200,000 lines of code six weeks later.

Topics discussed:

• Blocking Google to force a personal AI habit change

• Building a firm-wide skill library from core functions

• Turning a leadership summit into a hands-on AI hackathon

• Deploying AI champions to find and clear bottlenecks

• Vibe designing products to compress development timelines

• Staying tool-agnostic using the Netscape browser-wars lesson

• Rebuilding financial models with adjustable toggles in Claude

• Training an internal server on years of deal data

Slava Khristich has run a software firm for 13 years, and he argues the buy-versus-build math that ruled enterprise software for decades is now inverting. After hitting a functionality wall with HubSpot, he took screenshots and had a working CRM running in his own environment a week later.

As CTO and Founder of TATEEDA, Slava tells Keith why his team treats a second AI model as an adversarial reviewer of the first, and why that still doesn't remove the human gate. His core argument: when you hand-write code, you remember where every function lives, so you can find the bug. AI generates so much code so fast that this mental map disappears, which is why architectural knowledge, not typing speed, becomes the constraint. That single idea explains why he refuses to hire junior developers, why he thinks QA people are getting more valuable, and why he still won't let agents run unattended on complex work.

Topics discussed:

• Using a competing AI model as an adversarial code reviewer

• Why AI code volume destroys a developer's mental debugging map

• Architectural knowledge as the real constraint, not typing speed

• Why junior developers become a liability in AI-assisted teams

• Rebuilding a commercial CRM from screenshots in one week

• Where generic SaaS loses the buy-versus-build war

• The forward-deployed model: senior engineers who talk to clients directly

• Why QA roles gain value as code output accelerates

What happens when you spend three decades perfecting cancer diagnostics without ever taking a dollar of venture money? At Invivoscribe, it means building what Peter Thiel calls a complex coordinated business, the same category he places Apple in, one too slow to mature for VC funding but built to compound for decades.

Jeffrey Miller, Founder, CSO & CEO, breaks down how his team uses unsupervised learning trained only on normal samples to detect cancer without human bias, how a new sample-prep platform cuts a 50-touchpoint workflow down to nine and saves 63% on cost, and why the FDA's shift toward surrogate endpoints could turn decade-long drug approvals into one or two years.

Topics discussed:

• Using unsupervised learning trained only on normal samples

• Cutting sample-prep touchpoints from 50 to nine, saving 63% cost

• Detecting emerging clonal cell populations before they escape treatment

• Growing to 700 clinical labs through KOL publications, not marketing

• Peter Thiel's complex coordinated business framework applied to diagnostics

• FDA's shift toward surrogate endpoints accelerating drug approval timelines

• Reducing hematopathologist review time from 40 minutes to three

• Applying the HIV monitoring model to managing cancer long-term

Steve Weiss built MuteSix into one of the most competitive performance marketing agencies in the country and sold to Dentsu for a little north of $100M. In this episode, Steve joins Grayson Lafrenz, Founder and CEO of Cadre AI, to get into what agency owners consistently get wrong about AI adoption, why most companies can't point to AI actually showing up in their P&L, and where the real operating leverage will come from as teams get smaller and output expectations go up.

Steve's starting point for AI is not a company-wide rollout. It's identifying who your two or three biggest key man risks are, then building the AI system around how those specific people think, make decisions, and operate. He used his CSO Moody as the example: before trying to automate the company, you create an operating system around that one person. He also walks through the conditions that actually determine M&A readiness, and how he thought about valuation in a way most founders skip. Then both Steve and Grayson get into why competitive pressure, not internal enthusiasm, is what will finally push businesses to treat AI as a must-have.

Topics discussed:

• Building AI systems around key individuals before any company-wide deployment

• Using behavioral signals like CMO disengagement to predict client churn early

• M&A readiness conditions: EBITDA floor, churn, and buyer type by timeline

• PE versus strategic: how your desired runway determines which path makes sense

• Avoiding the "CRM Ponzi scheme" trap where new clients just replace churned ones

• POD structure and LMS as the infrastructure that makes hiring and scaling repeatable

• Why SaaS UIs disappear and agents take over buying decisions, with a 20-30% premium for those who opt out

• Why public speaking and interpersonal communication beat coding as the highest-leverage skill to develop now

Neal Bloom spent a year on the sidelines from 2023 to 2024, passing on nearly every software deal because he couldn't figure out how to underwrite pure SaaS to 100x anymore. That pause pulled him toward hardware, and what he found changed how he thinks about where returns actually come from now. As Managing Partner at Rising Tide Partners, and with roughly 70-80 investments behind him across angel bets and Interlock Capital, he has a sharp and specific point of view on what's worth backing and what isn't.

In this episode, Neal walks Keith through his actual diligence process: running Granola AI on every conversation, building longitudinal founder profiles over multiple meetings before any pitch happens, and assembling three-person quasi diligence teams before he'll seriously look at a deal. The frameworks are specific and replicable.

Topics Discussed:

• Sitting out 2023-2024 on software and what that silence led to

• Why every company building its own software kills the pure SaaS return profile

• Hardware's new appeal: easier manufacturing software lowering the build barrier

• Meeting founders in their "negative one year" phase and tracking them over time

• Turning recorded conversations into AI-synthesized founder memos

• Three-legged stool diligence: voice of customer, go-to-market operator, technical builder

• Using diligence gaps as a recruiting tool to expand the LP network

• Training a real-time matching algorithm to connect the right people at the right moment

Zeke Blattler, Co-Founder & CEO of Los Cuernos Wine, is running a two-person company gunning for a $100-300M exit — and his argument for why that's achievable without a large team is one of the more clear-eyed takes on company building you'll hear. With $90 billion in revenue locked inside glass bottles and an aging consumer base, he saw a structural window most incumbents are too slow to act on. The strategy he and co-founder Cory Assink built to get there is specific, counterintuitive, and worth studying regardless of what industry you're in.

Zeke walks Keith through a full first-principles rebuild of the wine supply chain — every distribution, packaging, and placement decision made by working backward from the consumer, not forward from industry convention. He's not pitching wine. He's showing how a two-person operation with fractional resources and a borrowed distribution network is out-maneuvering legacy players sitting on billions in entrenched infrastructure.

Topics Discussed:

• Rebuilding an entire supply chain from scratch by starting at the shelf and working backward

• Targeting beer distributors specifically because they're weakest in wine and strongest in the venues — restaurants, hotels, golf courses — where canned wine actually wins

• How two real customer moments (a half-open cork, a forced full-bottle purchase) reveal why wine loses the sale before it's even poured

• Surviving a 60% revenue collapse from peak summer to Q4 and what it actually takes to forecast a can-first business

• Running 423 accounts with two full-time employees by structuring a 1-to-50 headcount ratio through distributor networks

• How incumbent wine companies are using the same regulatory infrastructure built after Prohibition to slow down new entrants

• The shift in company valuation signals from headcount growth to profitable, talent-dense efficiency — and why AI accelerates it

• Targeting $24M run rate as the entry point to own the premium position in a $10-15B emerging category before competition catches up

Zeke Blattler, Co-Founder & CEO of Los Cuernos Wine, is running a two-person company gunning for a $100-300M exit — and his argument for why that's achievable without a large team is one of the more clear-eyed takes on company building you'll hear. With $90 billion in revenue locked inside glass bottles and an aging consumer base, he saw a structural window most incumbents are too slow to act on. The strategy he and co-founder Cory Assink built to get there is specific, counterintuitive, and worth studying regardless of what industry you're in.

Zeke walks Keith through a full first-principles rebuild of the wine supply chain — every distribution, packaging, and placement decision made by working backward from the consumer, not forward from industry convention. He's not pitching wine. He's showing how a two-person operation with fractional resources and a borrowed distribution network is out-maneuvering legacy players sitting on billions in entrenched infrastructure.

Topics Discussed:

• Rebuilding an entire supply chain from scratch by starting at the shelf and working backward

• Targeting beer distributors specifically because they're weakest in wine and strongest in the venues — restaurants, hotels, golf courses — where canned wine actually wins

• How two real customer moments (a half-open cork, a forced full-bottle purchase) reveal why wine loses the sale before it's even poured

• Surviving a 60% revenue collapse from peak summer to Q4 and what it actually takes to forecast a can-first business

• Running 423 accounts with two full-time employees by structuring a 1-to-50 headcount ratio through distributor networks

• How incumbent wine companies are using the same regulatory infrastructure built after Prohibition to slow down new entrants

• The shift in company valuation signals from headcount growth to profitable, talent-dense efficiency, and why AI accelerates it

• Targeting $24M run rate as the entry point to own the premium position in a $10-15B emerging category before competition catches up

Most boards pressure founders to reduce direct reports as they scale. Elia Wallen took the opposite approach: he manages 15 direct reports while growing Engine past 1,000 employees to a $2.1B valuation, and he eliminated the Chief Product Officer role entirely. His logic: additional management layers slow decision-making when speed matters most. When you build talent density and restructure one-on-ones around problem-solving instead of status updates, you can operate with higher spans of control.

That philosophy was nearly tested to breaking point. Engine came within weeks of running out of cash before their Series A with Telescope Partners closed. At the time, they had 150-180 employees across two entities. That near-death experience shaped Wallen's rejection of rigid planning. He refuses to create five-year plans, preferring to pivot quickly when new information emerges. Case in point: he publicly declared Engine would never do flights and cars, then reversed the decision weeks later. On AI, they've now automated over 50% of support cases while actually improving customer satisfaction scores—not by replacing people, but by redeploying them to higher-value work as the company doubles annually.

Topics discussed:

• Managing 15 direct reports by focusing one-on-ones on unblocking problems and identifying blind spots

• Operating without a Chief Product Officer to eliminate layers between product teams and customers

• Nearly running out of cash weeks before Series A closed despite previous successful exits

• Understanding valuation multiples: 30x in 2021 versus 9-10x in 2024 makes growth more substantial than surface numbers suggest

• Prioritizing will over experience when hiring to find people still hungry to prove themselves

• Automating 50%+ of support cases with AI while improving satisfaction by overcoming legacy system skepticism

• Deploying AI internally first for faster iteration with immediate feedback before customer-facing rollout

• Building personalization and voice/chat interfaces for travel instead of traditional search experiences

• Establishing talent density before removing rules to enable Netflix-style freedom and responsibility

• Actively deleting processes and bureaucracy to keep the organization young despite reaching scale