Free Form AI

Being technically right doesn't mean anyone understands you. In Episode 53 of Free Form AI, Michael Berk and Ben Wilson break down how engineers can communicate complex technical ideas to nontechnical stakeholders, explain uncertainty, and make sure their message actually lands. We discuss how to present data without misleading people, why trusted experts need to question their own assumptions, an...

AI makes it dramatically faster to build software. It also makes it easier to build the wrong thing. In this episode, we break down the product development lifecycle from defining the right problem to designing, testing, and driving adoption, including why teams should actively try to invalidate their own ideas, where Gen AI creates false confidence, and what actually limits product development wh...

Four years at Databricks produced eight career lessons. Now Ben gets to challenge them. We dig into why humility makes you more effective, how “make them tell you no” can create career opportunities, why understanding the problem matters more as AI gets better at implementation, and how systems, visibility, curiosity, and personal brand shape long-term career growth. 00:00 — Why you’re less import...

Fifty episodes into Free Form AI, we’re looking back at what actually changed our minds. We break down five lessons on why predicting AI is nearly impossible, how sharing unfinished ideas makes them better, why unexpected connections and analogies are powerful tools for learning, and how much control you actually have over your career and workplace culture. 00:00 — Why predicting AI is nearly impo...

Writing a technical book sounds great until you learn what it actually takes. In this episode, Ben breaks down his experience writing Machine Learning Engineering in Action , from pitching and rewriting chapters to technical reviews, royalties, and spending up to 60 hours a week on the process. We also get into whether publishing a tech book actually helps your career, why writing forces you to un...

The best career advice isn't about climbing the ladder faster. It's about building habits that compound over decades. In this episode, Michael reflects on four years at Databricks, sharing eight lessons on career growth, systems thinking, mentorship, decision-making, and personal brand. Whether you're an engineer, technical leader, or just starting out, these are the principles that have had the b...

Modern LLMs are already incredibly capable, yet most enterprise AI projects still fail. In this episode, we explore why context, not model intelligence, is the real bottleneck, how traditional RAG systems fall short, and what it takes to build AI that actually understands your business. We also share lessons from building internal AI systems at Databricks and discuss why better knowledge represent...

The most impactful careers aren't built by following a plan, they're built by following curiosity. In this episode, we explore why mastery comes from understanding systems instead of memorizing facts, how curiosity compounds into influence over time, and why helping others develop intuition may be the highest-leverage work an engineer can do. We also discuss legacy, mentoring, and what actually dr...

As AI makes information retrieval nearly free, the value of memorization continues to decline. In this episode, we explore why systems thinking, curiosity, and deep intuition are becoming the most important skills in the AI era. We also break down the concept of the Agora, how stories transfer knowledge more effectively than facts, and why learning to ask better questions may matter more than lear...

Everyone debates when AGI will arrive. Fewer people ask what happens next. In this episode, we break down the difference between AGI and ASI, why humans instinctively personify AI systems, and the societal challenges that emerge when intelligence becomes abundant. We also discuss AI companionship, regulation, creativity, and what remains uniquely human in an AI-driven world. 00:00 — What AGI actua...

LLMs are optimized to sound convincing—not to know when they’re wrong. In this episode, Deanna Emery breaks down why hallucinations are fundamentally tied to how language models work, why confidence is often disconnected from correctness, and how better evaluation strategies can make AI systems more reliable in production. We also get into uncertainty, semantic reasoning, and what humans still do...

Building a great AI team takes more than hiring smart people. In this episode, Brooke Wenig breaks down how Databricks built the AI FDE organization, why culture compounds faster than technical skill, and what separates high-trust engineering teams from teams that slowly degrade over time. We also get into mentoring, hiring in the age of AI coding tools, and why software engineering fundamentals m...

Most AI startups optimize speed, automation, or revenue. Sol built one to stop people from emotionally destroying each other. After a brutal divorce and years trapped inside high-conflict co-parenting, he realized the real problem wasn’t logistics, it was emotional escalation through constant communication. BestInterest uses AI to filter manipulative, hostile, and triggering messages before they r...

AI alignment breaks the moment we assume intelligence automatically produces morality. Dr. Peter R. Solomon argues the real danger isn’t sentient AI becoming evil, it’s AI inheriting no emotional history, no family structure, and no reason to value human survival. The conversation moves from CRISPR in high schools to AI-generated writing, autonomous agents, synthetic memory, and why “guardrails” f...

Most AI systems fail before they scale because teams build the “final architecture” too early. Research spikes exist to expose what you don’t understand, not to prove you’re right. The fastest path to a working system is starting with something intentionally simple, validating invariants, and throwing away what doesn’t hold. 00:00 What a research spike actually is 04:15 The real problem: context o...

LLMs don’t fail loudly, they drift into undefined behavior and take your system with them. The only way to build stable AI systems is to enforce contracts at every boundary, especially when dealing with non-deterministic outputs. Modern Python tools like Pydantic, enums, and structured interfaces aren’t optional, they’re how you turn probabilistic generation into reliable software. 00:00 Why LLMs...

AI has made it easier than ever to build fast, but also easier to fake progress. In this episode, we break down the difference between people who optimize for real value versus those who optimize for visibility, how this shows up in AI workflows, and why long-term thinking is the only way to build systems that actually matter. 00:00 — Builders vs posers in tech 05:00 — Why visibility can distort i...

AI coding agents can generate code faster than ever, but they also introduce new risks that most teams don’t understand yet. In this episode, we break down how software development is shifting from writing code to reviewing and directing AI systems, why poorly guided agents create fragile systems, and what skills actually matter as AI takes over execution. 00:00 — AI writing code vs humans reviewi...

The conversation delves into the use of AI in the legal industry, focusing on the Succession platform and its approach to will creation. It explores the collaboration between technical and non-technical stakeholders, the efficiency of billable practices, and the fundamentally human aspects of law, particularly in in-person negotiations. Takeaways AI for sensitive topics Collaboration between techn...

The conversation delves into the importance of data science in AI, the challenges with existing evaluation frameworks, the significance of looking at data, types of tests and prioritization, the evolution of evaluation solutions, overcoming psychological blocks, summary statistics and data exploration, and the future of evaluation and outsourcing. Takeaways Evaluating AI products requires a deep u...

The conversation covers topics related to software development, testing, and team culture. It emphasizes the importance of system-driven solutions, effective testing strategies, and fostering a positive and collaborative work environment. The discussion also delves into the impact of individual behavior on team dynamics and the significance of leading by example. Takeaways System-driven solutions...

The conversation delves into the philosophy of collaboration, the absurdity of technical work, the ethical imperative of mentorship, the role of a technical role model, and embracing the absurdity of the technical world. Key takeaways include the emphasis on collaboration over individual excellence and the importance of mentorship and knowledge transfer. The conversation delves into the power of k...

The conversation delves into the concepts of hygiene vs. motivation factors in the workplace and explores the methods of creating motivation within a team. It also touches on the role of a manager in maintaining team efficiency and morale. The conversation delves into the importance of mentorship and the builder ethic, emphasizing the value of humility and team cohesion. It also explores the conce...

The conversation delves into the scale of finance and the concept of 'enough' in relation to wealth. It explores the value of money, safe withdrawal rates, financial mistakes, and the transition to wealth management as a career path. The conversation delves into the use of AI-based coding assistants, the importance of guardrails and testing focus, AI oversight and refactoring, workflow and quality...

The conversation delves into Victor Dibia's career journey, global experiences, transition to a PhD, and strategic career planning. It also explores his focus on AI tooling and frameworks, as well as the evolution of Autogen and the Microsoft Agent Framework. The conversation delves into the actor-first paradigm in multi-agent systems and the concept of ensembling in machine learning. It explores...