Dan Klein, professor at UC Berkeley and CTO at Scaled Cognition, explains that AI systems generate answers based on patterns in language rather than verified knowledge. This is what allows them to be flexible and useful across many tasks, but it also means they can produce confident answers even when they are not fully accurate.

He introduces the “jagged frontier,” where AI performs very well in some areas and less reliably in others. Because responses are fluent and easy to read, it is not always obvious where those limits are, which makes it important to stay engaged when using these systems.

The conversation also explores hallucinations as a natural part of generative systems. In some cases, this is what makes them valuable, especially for creative or open ended tasks, while in other cases reliability becomes more important.

Finally, Dan highlights that working effectively with AI is a skill. As more people start using these systems in their daily work, knowing how to guide them, check outputs, and use them in the right contexts becomes increasingly important. He also shares how his team at Scaled Cognition is approaching this challenge, focusing on building more deterministic systems designed to provide stronger guarantees and reliability, particularly for high stakes, policy driven use cases.

Key Takeaways:

• AI is designed to sound right, not to know it’s right
  Models generate fluent answers without knowing whether they are correct, which means users need to actively evaluate outputs
• You have to learn where AI works and where it doesn’t
  Capabilities are uneven, and understanding those limits is key to using AI effectively
• Working with AI shifts your role from creator to editor
  Instead of starting from scratch, you are reviewing, refining, and validating what the model produces
• Most people are using AI without knowing how to manage it
  Skills like delegation, verification, and judgment are becoming essential, but are not widely taught

Dan's LinkedIn: linkedin/dan-klein/
Scaled Cognition Website: scaledcognition.com
Scaled Cognition LinkedIn: linkedin/company/scaledcognition/
Scaled Cognition X: x.com/ScaledCognition

00:00 Intro: Fluency vs Truth
00:34 Meet Dan Klein
02:53 Why Fluency Misleads
05:11 How LLMs Guess
07:30 What Is Hallucination
08:54 Deception and Alignment
11:22 Why Agents Break
12:48 Chaining and Determinism
16:01 When Hallucination Helps
22:33 Beyond Scale for Reliability
30:40 Synthetic Data Training
31:10 Enterprise Agent Use Cases
33:44 Healthcare Risks
39:13 Enterprise Literacy Gap
41:27 Delegation and AI Management
54:37 The Debrief

📜 Read the transcript for this episode

For more prompts, tips, and AI tools. Check out our website: https://www.beyondtheprompt.ai/ or follow Jeremy or Henrik on Linkedin:

Henrik: https://www.linkedin.com/in/werdelin
Jeremy: https://www.linkedin.com/in/jeremyutley

Show edited by Emma Cecilie Jensen.

Melissa Cheals leads Smartly, a payroll and people management platform serving 24,000 small and medium businesses in New Zealand. In this conversation, she shares how AI is reshaping product development, leadership, and how organizations operate.

A key moment comes when her team estimates new features will take 12 months and $1M to build. Instead of accepting it, Melissa pushes back, using AI to better understand her team’s perspective and communicate the need for change more effectively. This becomes a broader shift in how she approaches leadership, using AI to think more clearly and navigate conversations with less friction.

The discussion expands into strategy. Companies now face a fundamental choice: become AI-native or continue building on existing systems. As AI adoption increases, it also exposes silos and bottlenecks. Melissa shares why cross-functional collaboration—and leaders actively engaging with AI themselves—is critical to navigating this shift.

Key Takeaways:

• Becoming AI-native is a defining decision
  It’s not just a technology shift. Leaders need to decide whether to rebuild around AI or continue layering it onto existing systems, and that choice shapes how the company operates.
• AI shifts us from scarcity to abundance
  Many organizations still think in terms of limited time and resources, but AI changes what’s possible and forces leaders to rethink how big they can think and what they can achieve.
• AI is a leadership amplifier
  Beyond productivity, AI helps leaders think more clearly, reframe conversations, and communicate change in a way that is both effective and respectful.
• Leaders can’t delegate AI
  Without hands-on experience, it becomes difficult to challenge assumptions, guide teams, or make informed decisions about what’s possible.

Smartly: smartly.co.nz
LinkedIn Melissa: linkedin.com/melissa-cheals
LinkedIn Smartly: linkedin.com/company/smartlynz/

00:00 Intro: Challenging AI Assumptions
00:28 Meet Melissa Cheals
01:17 The Spark For Change
02:36 Vision And Early Signals
03:48 Hiring For Transformation
06:12 Unlocking Data With AI
08:27 Breaking Silos Across Teams
10:39 Why Leaders Must Learn AI
13:42 Leading With AI And Clarity
17:05 The AI-Native Decision
21:45 Thinking Bigger With AI
25:23 Less Meetings More Writing
26:33 The Self-Disruption Imperative
29:11 Breaking Silos With Value Streams
31:28 Managing Fear And Change
32:50 Learning And Shipping Faster
34:58 Debrief

📜 Read the transcript for this episode: ai-native-or-not-the-defining-choice-for-companies-right-now-with-melissa-cheals-ceo-of-smartly/transcript

For more prompts, tips, and AI tools. Check out our website: https://www.beyondtheprompt.ai/ or follow Jeremy or Henrik on Linkedin:

Henrik: https://www.linkedin.com/in/werdelin
Jeremy: https://www.linkedin.com/in/jeremyutley

Show edited by Emma Cecilie Jensen.

Greg Shove describes a growing gap between individual and organizational AI adoption. A small group of employees are already using AI effectively, while most companies are still early. AI is generating real productivity gains, but those gains are not being captured at the company level. Instead, they are absorbed by individuals who use AI to work faster, often without changing team outputs or structures — raising a central question: if AI creates time, where does that time go?

The conversation explores why enterprise AI adoption remains uneven. Many organizations lack a clear point of view on AI, and workflows take time to adapt, making it difficult to turn individual gains into coordinated results. At the same time, AI is breaking capability boundaries, allowing people to take on work across roles while companies remain structured around existing ways of operating.

From a leadership perspective, Greg emphasizes that the challenge is not just efficiency. AI creates capacity, but without clear direction on how to use it, that capacity disappears. Leaders must decide how to reinvest the time AI creates if they want to capture real business value.

Key Takeaways:

• AI’s ROI is leaking, not missing
  Companies are generating value from AI, but it’s being captured by employees rather than the organization.
• A small group drives most of the impact
  Roughly 10–15% of employees adopt AI early and use it effectively, creating an uneven distribution of gains.
• AI is breaking capability boundaries
  Individuals can now take on work across roles, but organizations are still structured around fixed responsibilities.
• Most companies lack a clear point of view on AI
  Without direction from leadership, adoption becomes fragmented and employees are left to figure it out themselves.
• Leaders must decide what to do with the time AI creates
  Efficiency gains alone don’t create value. Organizations need to define new, higher-value work or the gains disappear.

Greg's LinkedIn: linkedin/gregshove
Section LinkedIn: linkedin/company/sectionai
Section AI: sectionai.com
Prof AI: prof.ai

00:00 Intro: Entering the Era of AI Chaos
00:31 Meet Greg Shove
01:32 Enterprise AI Is a C Minus
01:51 AI’s ROI Is “Leaking” to Employees
03:04 When Individuals Outrun the Organization
05:44 When AI Breaks Workflows
06:47 Disposable Software and New Ways of Building
09:10 Cut vs Create
12:01 Using the Calendar as a Lever
16:24 Why Enterprises Don’t Move
17:32 When Customers Force Change
21:31 AI Breaks Capability Boundaries
25:44 The Productivity Firehose
27:49 Who Actually Captures the Value
28:45 Why Everyone Needs Good AI
32:00 Adoption Beats Buying More Tools
40:17 Teaching the 90 Percent
43:48 Where Humans Still Matter
48:09 The Debrief

📜 Read the transcript for this episode: greg-shove-on-why-most-companies-are-not-seeing-roi-on-ai-yet/transcript

For more prompts, tips, and AI tools. Check out our website: https://www.beyondtheprompt.ai/ or follow Jeremy or Henrik on Linkedin:

Henrik: https://www.linkedin.com/in/werdelin
Jeremy: https://www.linkedin.com/in/jeremyutley

Show edited by Emma Cecilie Jensen.

Leidy Klotz has spent years studying a simple but overlooked phenomenon: when we try to improve something, our first instinct is to add rather than remove. He shares the Lego bridge experiment that sparked his research and explains how this additive bias scales from small design decisions to entire organizations. Over time, companies accumulate reporting lines, meetings, software, and policies without questioning what no longer serves them.

Henrik and Jeremy explore how AI tools intensify this pattern. When generating ideas, launching projects, writing code, or producing content becomes effortless, the temptation to add grows stronger. The cost of producing information drops, but the cost of consuming it rises. Without guardrails, organizations risk what Leidy calls “organizational indigestion.”

The discussion moves from insight to implementation. Leidy outlines practical ways to counteract additive bias, including stop-doing lists, default kill dates on projects, and designing environments that make subtraction visible and acceptable. In a world of accelerating AI output, leaders must intentionally decide what to remove, what to protect, and what truly matters.

Key Takeaways:

• We default to adding, not subtracting
  When faced with a problem, our instinct is to introduce something new. Subtraction rarely occurs to us, even when removing something would improve clarity and performance.
• Generative AI amplifies additive bias
  AI makes producing content, code, and ideas easier than ever. Without constraints, this frictionless creation can accelerate complexity instead of progress.
• More organizations die from indigestion than starvation
  Over time, companies accumulate tools, processes, and policies that quietly slow them down. The real risk is often not too few ideas, but too many unexamined additions.
• Architecture beats willpower
  Rather than relying on discipline alone, leaders can design systems that encourage subtraction. Stop-doing lists and default expiration dates make removal expected instead of exceptional.
• Protect what matters before adding more
  Before introducing new tools, workflows, or AI systems, leaders must define what is already working and worth protecting. Subtraction requires clarity about what should stay, not just what should go.

Subtract: amazon/Subtract-Untapped-Science-Leidy-Klotz
In a Good Place: amazon/Good-Place-Spaces-Where-Thrive/
Leidy's Speaking: https://leidyklotz.com/
Clip from Bear: Subtract - this is how you do better

00:00 Intro: Our Instinct to Add
00:28 Meet Leidy Klotz
01:15 The Subtract Idea
02:56 Organizations Get Bloated
03:49 Scandinavian Design Mindset
04:32 New Book: In a Good Place
05:59 AI Abundance and Indigestion
08:12 Curate Context, Not More
11:38 Cues and Stop-Doing Lists
15:00 Default Debt and Kill Dates
17:10 Odysseus Contracts and Biases
21:28 Reengage the Physical World
29:17 Bike Shedding and Priorities
36:10 Making Is Thinking
49:16 The Debrief

📜 Read the transcript for this episode: how-to-subtract-the-most-underrated-skill-of-the-ai-era-with-leidy-klotz/transcript

For more prompts, tips, and AI tools. Check out our website: https://www.beyondtheprompt.ai/ or follow Jeremy or Henrik on Linkedin:

Henrik: https://www.linkedin.com/in/werdelin
Jeremy: https://www.linkedin.com/in/jeremyutley

Show edited by Emma Cecilie Jensen.

Fathom was built on the assumption that transcription would become commoditized and generative models would steadily improve. Rather than training proprietary models, Richard focused on building the infrastructure around them and waiting for model capabilities to reach the right threshold.

In this conversation, he explains why AI has made effort and impact harder to predict, and why that shifts product development from roadmap execution toward experimentation. He describes separating an exploratory AI team from core engineering, structuring that team to prototype and write specs, and expecting a meaningful portion of experiments not to work.
Richard introduces his Jenga model for AI development, testing different models and use cases to find where resistance is lowest. He also discusses the operational realities of rapid model updates, hallucination rates, and what he calls the LLM treadmill.

The discussion explores qualitative QA, organizational design, buy versus build decisions, and why leadership taste plays an increasingly important role as AI lowers the barrier to generating outputs.

Key takeaways:

• Estimating effort and impact is becoming harder
  As model capabilities improve quickly, features that require months today may take far less time in the near future. This makes traditional planning assumptions less stable.
• Product development increasingly resembles R&D
  With shifting capabilities and uncertain outcomes, teams must experiment, prototype, and iterate rather than rely solely on long term roadmaps.
• Organizational structure must reflect experimentation
  Separating exploratory AI work from core engineering can allow faster iteration while maintaining stability elsewhere.
• Rapid model updates create operational pressure
  Frequent improvements and changing performance levels can require teams to revisit and adjust features more often than in traditional software cycles.
• Qualitative judgment plays a larger role
  As AI lowers the cost of generating outputs, evaluating quality and deciding what to ship becomes increasingly important.

Fathom: fathom.ai
Fathom LinkedIn: linkedin/company/fathom-video/
Richard's LinkedIn: linkedin/in/rrwhite/

00:00 Intro: Why AI Breaks Roadmaps
00:19 Meet Richard White (Fathom AI)
02:16 From Roadmaps to R&D
04:49 Designing AI Teams for Speed
07:11 The Jenga Model
09:56 Failing 50% & AI Team Psychology
13:40 LLMs as Interns & Anti-Planning
21:01 QA, Data Pain & Developing Taste
24:59 Executive Taste & Culture Rules
27:20 Reacting to AI Waves
28:50 Fathom’s 4-Step Product Plan
30:47 What New Models Unlock
32:13 From Scribe to Second Brain
40:32 Build vs Buy in AI
45:32 The Debrief

📜 Read the transcript for this episode: from-roadmaps-to-rd-how-ai-is-changing-product-development-with-richard-white-founder-of-fathom-ai/transcript

For more prompts, tips, and AI tools. Check out our website: https://www.beyondtheprompt.ai/ or follow Jeremy or Henrik on Linkedin:

Henrik: https://www.linkedin.com/in/werdelin
Jeremy: https://www.linkedin.com/in/jeremyutley

Show edited by Emma Cecilie Jensen.

In this episode, Bryan McCann joins Henrik and Jeremy to explore how search is evolving from simple queries into more conversational and agent-driven systems, and why prompting is likely a temporary skill. Bryan shares how his definition of productivity changed as an AI researcher, moving away from doing the work himself and toward designing plans and experiments that machines could run continuously.

The conversation expands to leadership and organizational design. Bryan explains why helping others learn how to work with AI became his highest-leverage activity, and offers a simple rule of thumb: try to get AI to do the task first, and treat anything it can’t do as an interesting research problem. Henrik and Jeremy connect this to Bryan’s view that organizations may increasingly resemble neural networks, with information flowing more freely and decisions less tied to rigid hierarchies.

Key Takeaways:

• Productivity can be measured by machine output, not human effort
  Bryan explains how “keeping the GPUs full” became his primary measure of productivity.
• Prompting is useful, but likely temporary
  The episode discusses why future systems may rely less on explicit prompts and more on inferred context.
• Try AI first, then learn from what it can’t do
  Tasks AI struggles with can reveal meaningful research opportunities.
• Leadership is about scaling others
  Bryan shares how his focus shifted from scaling himself to helping his team increase impact.
• Organizations may benefit from neural-network-like design
  Better information flow and fewer bottlenecks can improve decision-making.

YOU: You.com
Bryan's website: bryanmccann.org
LinkedIn: linkedin/company/youdotcom/

00:00 Intro: Keeping the GPUs Full
00:22 Meet Bryan McCann: CTO & co-founder of You.com
00:43 Why Search Is Breaking - and Why It Becomes a Skill
01:41 From Search to Agents
03:18 The Case for Proactive, Context-Aware AI
04:30 We Don’t Need New Hardware - We Need Trust
05:43 The Trust Problem of Always-On Listening
07:57 Trust as the Real Bottleneck (Not AI Capability)
09:52 Delivering Immediate Value to Earn Trust
12:13 Business Models and Escaping the Attention Economy
17:27 What “Agents” Really Mean - and Why the Term Will Fade
20:37 Productivity, Parkinson’s Law, and Keeping the Machines Running
23:52 Scaling Yourself vs. Scaling Your Team
29:57 Building Culture: Automate, Throw Away, Rebuild
35:46 Designing Organizations Like Neural Networks
45:02 Recruiting for Initiative in an AI-Native Organization
49:18 The debrief

📜 Read the transcript for this episode: podcast.beyondtheprompt.ai/heres-how-to-know-if-youre-getting-the-most-out-of-ai-with-bryan-mccann-cto-of-youcom/transcript

For more prompts, tips, and AI tools. Check out our website: https://www.beyondtheprompt.ai/ or follow Jeremy or Henrik on Linkedin:

Henrik: https://www.linkedin.com/in/werdelin
Jeremy: https://www.linkedin.com/in/jeremyutley

Show edited by Emma Cecilie Jensen.

In this episode, Humza Teherany breaks down how he bridges deep technical fluency with strategic leadership at MLSE, home to the Raptors, Maple Leafs, and more. He shares how a vacation turned into an AI reawakening and how that hands-on immersion led to a fundamental shift in how his organization builds and experiments.

Humza walks through MLSE’s build in a day practice, their internal AI platform, and why speed to prototype now unlocks more than just efficiency. It changes who gets to shape the future. He, Jeremy, and Henrik explore the limits of traditional enterprise AI rollouts and how to build spaces for superusers that enable company-wide transformation. The conversation covers how technical literacy impacts credibility, why idea execution is the new differentiator, and how Humza’s five-year-old inspired a bedtime story app powered by AI.

Whether you're a CTO, a founder, or just figuring out where to start, Humza makes a compelling case. The best leaders don’t delegate this moment. They build.

Key Takeaways

• Leaders should not delegate the AI moment
  Humza, Henrik, and Jeremy agree that this is a moment for leaders to be hands-on. The ones who build and explore the tools themselves are the ones unlocking real impact.
• Technical fluency builds credibility and better decisions
  Humza’s return to his technical roots has changed how he leads. Understanding how AI works helps leaders earn trust and make smarter, faster choices.
• Speed enables inclusion
  MLSE’s build in a day model allows more people to contribute ideas and see them turned into real prototypes. Moving fast isn’t just efficient - it changes who gets to participate.
• Empower your superusers first
  Rather than starting with enterprise-wide training, Humza focuses on enabling the small group already eager to build. That early energy helps drive broader culture change.

MLSE: mlse.com
LinkedIn: Humza Teherany - LinkedIn

00:00 Intro: Humza Teherany and MLSE
00:27 The Role of C-Suite Leaders in AI
01:08 Reconnecting with Technical Skills
02:08 Diving Deep into AI Tools
03:03 The Importance of Hands-On Learning
04:25 Progression from Consumer to Technical AI Tools
07:28 Building a Business Case for AI
10:03 Creating a Culture of Innovation
14:00 Implementing AI in Business Operations
21:05 Challenges and Strategies in AI Adoption
26:17 Organizational Structure for AI Success
32:02 The Importance of Reviewing and Planning Code
33:01 The Future of Solo Developers and New Technologists
34:58 Reimagining Company Structures with AI
38:55 Key Skills for Future Technology Leaders
41:19 Personal AI Experiments and Innovations
46:52 Encouraging Creativity in Children with AI
49:11 The Debrief

📜 Read the transcript for this episode: building-an-enterprise-ai-innovation-lab-a-master-class-with-humza-teherany-chief-strategy-officer-of-maple-leaf-sports-and-entertainment/transcript

For more prompts, tips, and AI tools. Check out our website: https://www.beyondtheprompt.ai/ or follow Jeremy or Henrik on Linkedin:

Henrik: https://www.linkedin.com/in/werdelin
Jeremy: https://www.linkedin.com/in/jeremyutley

Show edited by Emma Cecilie Jensen.

Mikkel B. Rasmussen brings a rare lens to the AI conversation. As an applied anthropologist, he has spent decades helping companies like LEGO uncover what is really going on beneath the surface.

In this episode, he shares how deep insight often begins with being wrong, why surprise is the clearest sign you have found something meaningful, and how the pain of not knowing is essential to breakthrough thinking. He also explains how AI is transforming his own research, from pattern recognition to video ethnography, and introduces a provocative idea: Anthropology Without Anthropologists.

Jeremy and Henrik reflect on what it means to teach AI how to surprise us, how synthetic data might reshape experimentation, and why better insights begin with better questions.

Key Takeaways

• Insight starts with being wrong
  Mikkel defines insight as the gap between how we think the world works and how it actually is. Anthropology helps uncover these mismatches, and that is where real breakthroughs begin.
• Pain is part of the process
  Mikkel and Jeremy both reflect on the emotional struggle that precedes insight. The doubt, sleepless nights, and questioning whether the work will ever come together is not failure. It is a necessary stage of discovery.
• Surprise is a signal
  The moment of surprise, when a new pattern emerges or an assumption is shattered, is at the core of applied anthropology. For Mikkel, it is the clearest sign that you have found something real.
• AI can accelerate experimentation
  Mikkel shares how AI is already helping his team analyze patterns, run faster experiments, and even conduct interviews that outperform humans in some cases. The goal is not to replace people but to push the limits of what is possible.

HARL: humanactivitylab.com

00:00 Intro: Why This Conversation Matters
00:25 Meet Mikkel: Founder of Human Activity Laboratory
01:14 Understanding Anthropology and AI
03:32 Applied Anthropology: Tools and Techniques
04:56 The Role of Narratives in AI
07:06 The Importance of Sensory and Social Dimensions
13:06 Case Study: LEGO and the Anthropology of Play
21:07 The Role of Surprise in Anthropology
27:51 AI and Human Synergy
31:26 Exploring AI's Limitations and Potential
32:46 Anthropology Without Anthropologists
34:17 AI's Role in Generating Insights
37:23 Human Bias in AI-Generated Ideas
42:05 Synthetic Data and Its Applications
47:34 The Future of AI in Anthropology
49:25 The Debrief

📜 Read the transcript for this episode: why-ai-gets-people-wrong-the-real-source-of-insight-with-anthropologist-mikkel-b-rasmussen/transcript

For more prompts, tips, and AI tools. Check out our website: https://www.beyondtheprompt.ai/ or follow Jeremy or Henrik on Linkedin:

Henrik: https://www.linkedin.com/in/werdelin
Jeremy: https://www.linkedin.com/in/jeremyutley

Show edited by Emma Cecilie Jensen.