Feb 5, 2023 · 48m · lennys-podcast
AI and product management | Marily Nika (Meta, Google)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this podcast episode, Meta Product Lead Marily Nika joins Lenny Rachitsky to discuss how product managers can transition into AI leadership, demystify machine learning workflows, and avoid common product traps. Nika shares actionable insights on working with research scientists, validating early prototypes, leveraging low-code tools, and upskilling for an AI-first future.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Lenny holds 31.5% of the talking time here. How this is scored →
Lenny as informed peer 3.4 Guest teaching 4.4 Guest disagreement 1.1 Lenny pushing back 1.3
speaking balance: gold is Lenny, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 14:43 Hard rejection of AI in MVPs
Marily forcefully shuts down early-stage founders attempting to build complex ML models for validation, telling them to fake functionality with Figma instead.
Hardest push from Lenny ▶ 10:28 Lenny clarifies researcher collaboration premiseLenny stops Marily to explicitly clarify whether she means collaborating with automated tools or literal PhD researchers embedded inside product squads.
Biggest teaching moment ▶ 18:52 Explaining machine learning via toddler cognitive developmentMarily breaks down complex machine learning training into an intuitive analogy about teaching a three-year-old to recognize rhinos and elephants.
Lenny holds their own ▶ 33:09 Lenny challenges the success rate and ROI of internal ML projectsLenny leverages his extensive PM background to point out that most internal ML investments fail and asks whether off-the-shelf APIs will replace in-house model training.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Lenny as informed peer | Guest teaching | Guest disagreement | Lenny pushing back | Why |
|---|---|---|---|---|---|---|
| Staying Informed on AI Developments | 3 | 4 | 1 | 1 | Lenny opens by inquiring about resources and hype versus reality in AI. Marily explains that AI will become the default in tech and dismisses fears about writing jobs being replaced. | |
| Practical ChatGPT Workflows for Product Managers | 3 | 4 | 1 | 2 | Lenny asks how Marily practically incorporates ChatGPT into her daily PM work. Marily explains concrete prompt patterns for drafting mission statements and developing user personas. | |
| The Future of PMs and Working with Research Scientists | 4 | 5 | 2 | 2 | Lenny probes Marily's statement that all PMs will be AI PMs. Marily clarifies that she means PMs will directly partner with PhD research scientists and must learn to navigate research uncertainty. | |
| Shifting Mindsets: From Generalist to AI Product Management | 4 | 5 | 2 | 1 | Lenny asks how non-technical PMs can enter the space, and Marily reframes the AI PM role as solving the right problem rather than building for the sake of shiny objects. | |
| When to Avoid AI: The Pitfalls of AI in MVPs | 5 | 6 | 2 | 2 | Lenny shares prior experiences with low-ROI ML investments and asks when to avoid AI. Marily strongly warns early-stage founders against training models for an MVP, recommending Figma prototypes instead. | |
| Demystifying Models and Model Training | 4 | 6 | 1 | 2 | Lenny asks for plain-English definitions of models and training. Marily uses the analogy of teaching a toddler to recognize animals to explain neural network pattern recognition. | |
| Impactful AI in Action: Google Glass Real-Time Translation | 3 | 5 | 1 | 1 | Lenny asks about exciting real-world applications and whether AI will replace PMs. Marily shares Google Glass real-time translation examples and argues AI frees PMs from tedious PRD writing. | |
| Sponsor Break: Pando | 1 | 1 | 0 | 0 | Contains a sponsor read by Lenny followed by a quick exchange detailing online bootcamps and coding courses for PMs wanting to learn programming. | |
| Structure and Curriculum of the AI PM Course | 3 | 5 | 1 | 1 | Lenny asks how early-career PMs can level up. Marily outlines the curriculum of her course, emphasizing navigating career evaluation differences inside research organizations. | |
| Gaining Buy-In and Bridging Academia to Production | 5 | 5 | 1 | 2 | Lenny notes that most ML investments fail and queries how to secure executive buy-in. Marily recommends de-risking experiments through analogies to prior adjacent successes and contingency rollback plans. | |
| Student Projects and Accessible No-Code AI Tools | 3 | 5 | 1 | 1 | Lenny asks for standout student projects and no-code tools. Marily details student-built medical diagnostic prototypes and explains Google Cloud AutoML using a drone wind turbine inspection case study. | |
| Productizing Course Design and Iterative Teaching | 3 | 4 | 1 | 1 | Lenny asks about course creation mechanics. Marily describes applying product management principles and user discovery to design, iterate, and update her curriculum. | |
| Lightning Round: Favorite Books, Media, and Tools | 3 | 2 | 1 | 1 | Lenny runs through rapid-fire lightning round questions on books, favorite podcasts, TV shows, and AI interview prompts, wrapping up the conversation collaboratively. |
Statements from this episode (10)
Nika: Every future technology product will be AI by default
“What I'm advocating for and what I'm telling people is that in the future, everything will be AI by default.”
Nika: All Product Managers Will Become AI PMs in the Future
“I believe that all product managers will be AI product managers in the future.”
Nika: Product managers will never need to code or train AI models
“And as a PM, you will never need to actually train or code.”
Nika: Any product with user behavior data can be improved with AI
“Basically, anything where you can get data behind the behavior of the users can be improved with AI.”
Nika: Generalist PMs ship products while AI PMs solve problems
“I usually say that a generalist PM helps their team and their company build and ship the right product, but the AI PM helps their team and company solve the right problem.”
Nika: Never train AI models for an MVP; fake functionality with prototypes
“Don't do it for your MVP. It makes zero sense. Do not waste time of data scientists that can train models with using powerful machines that are going to take weeks to train. This is because if you have an MVP and you just want to get buy-in for an idea or a fe…”
Rachitsky: Most Startups Lack the Data Needed to Train Custom Models
“My guess is most startups are going to have nowhere near enough data to build their own model and make it something really interesting.”
Nika: Commercially packaged datasets create undifferentiated AI models
“There are agencies that are selling data, packages of data that are ready, so you can get them and train your models. But the question is, if everyone takes that exact data set, then the quality that every single company is producing is going to be the exact s…”
Nika: Product managers should learn coding fundamentals to understand AI tools
“I encourage people to just take an online course, understand more, get your hands dirty, pair up with someone else that's in the same boat as you, because this is going to give you the skillset to understand how that tool that's going to help you in your day-t…”
Nika: AI research PMs launch less often and require distinct evaluation metrics
“Usually product managers get ahead The more they launch. But if you're in a research org, you're not going to launch as often. So you need to make sure to clarify with the hiring managers early on. Hey, what does progress mean? How am I going to get assessed i…”