Schoening — AI era skills: Why cultivating agency matters more than job titles (Lenny’s Podcast)
Max Schoening is head of product at Notion, where he’s been especially effective at getting designers and PMs to ship code, prototype in the terminal, and launch extremely successful AI products. He was previously a PM at Google, ran design at Heroku, was VP of Design (and a part-time engineer) at GitHub, and is a two-time founder. He’s one of the most AI-forward product leaders out there and one of the deepest thinkers on how AI changes how we build and use software.
— channel description, Lenny’s Podcast
TL;DR
A ~87:22 Lenny’s Podcast interview (published 2026-05-02; auto-generated English captions, 1684 segments). Host Lenny Rachitsky; guest Max Schoening, head of product at Notion (prior career: PM at Google → ran design at Heroku → VP of Design + part-time engineer at GitHub → two-time founder). The episode anchors on agency as the durable AI-era skill — Schoening’s central thesis is that capability (being able to make things in code) is now democratised by AI, so the differentiator is agency (the disposition to make things, the recognition that the world is malleable and you can change it).
The substantive contributions are five.
1. Designers + PMs shipping code at Notion (~1:55–8:24). The Notion product org has been operating with designers writing/prototyping code and PMs shipping working software for a sustained period. Schoening pushes back on framing this as role-blurring and frames it instead as role-evolution: “do you drive Notion like it’s stolen?… you can still contribute to the company in a way that you feel agency and you’re not sort of just like, what’s your role?”
2. Agency-as-the-differentiator thesis (~10:32–11:49) — the episode’s titular claim:
“I think before it was very easy to always say ‘well I will never be able to do this because, insert skill issue,’ and I think we’re realizing that even if you have the skills at your fingertips because now I don’t know an AGI-adjacent model helps you, the thing that matters is agency. And I don’t think agency is very evenly distributed in the world, and I think people who have true agency and they understand that the world around them is malleable will do great, and the folks who stick to ‘what does it mean to be a PM, what does it mean to be a designer, what’s my job as an engineer’ — I think that will be much harder. And yeah, cultivate agency, I think that’s the thing.”
This is the wiki’s clearest single articulation of agency as a durable skill at the founder/product-leader altitude — distinct from the taste / judgement framings the wiki carries (Karpathy, Hu, Mittal, Bender) by being explicitly about the disposition to act rather than the discernment of quality.
3. Two Notion worked examples of high agency (~11:49–13:52):
- Brian Leven — Notion designer who “blurs engineering and design, but he also is probably our number one recruiter — hey, this is what the org needs, I’m going to go out and talk to people and find someone.” The agency vignette: contributing to org outcomes outside the day-to-day role.
- Eric Lou — Notion PM who asked Schoening, “hey look, at some point in the future if you started a startup would you hire me?” Schoening: “well not in the first 10 — I don’t need a product manager.” Eric: “oh okay, I’m going to work on the skills so that you would hire me in the first five.” He then spent more time in Figma instead of writing PRDs, and now “it’s just why do I have to do the Figma thing, can’t I just build the prototype and at least show you what I think and do the thinking in there?”
The Eric-Lou story is the wiki’s clearest narrative arc of a PM-evolving-into-builder-via-agency — the practitioner-altitude instantiation of Nika’s prescription from five months earlier.
4. What we might lose as roles merge (~13:52–15:56) — the physical-metaphor framing. Schoening pushes back on the simple role-merging is great framing:
“If we’re not careful, we will lose specialists. The way I would describe this — sometimes I like to think about software in terms of physical metaphors. If you and I were to build a hardware startup, well, we would make the first enclosures and prototypes with 3D printing. And you would see all the layer lines. It would be very obvious to you that this is not a thing that you should just give to people to pay for. And then there’s a long windy road all the way to the end where at some point if you’re very lucky you get to manufacture that product for 100 million people. So then the engineering is actually ‘how do I optimize the factory so that we have enough yield and so that we have enough precision.’ That to me is very absent right now from most of the discourse in software, which is it’s all about how many tokens can we spend and how many features can we ship. I’m like, OK, but where’s the engineering part — and the engineering part is the ‘you make sure that this thing works for 100 million people for a billion people.‘”
This is the wiki’s first articulation of the prototype-engineering distinction at AI-era scale: AI-generated code is the 3D-printed-prototype-with-visible-layer-lines; the engineering work is the optimise-the-factory-for-100M-users discipline that the agentic-engineering discourse currently underweights. On the design side equivalently: “anyone can now very quickly take a design system off the shelf and build a very usable user interface… but where is the delight in craft?”
5. Malleable software (~17:42–20:43) — the long-arc design philosophy Schoening has been advocating since before the AI revolution.
“Malleable software is the idea that software works closer to the interest of the people that use it than the interest of the corporation that makes it. Imagine you lived in an environment where you do not get to rearrange your living room and the kitchen has to be exactly set up the way that someone else decided. We would not take that. But that is kind of the world that we have in software right now where we have this world of apps and apps are like — every layer is glued together, the user interface, the data ownership and so on… The moment you’re like ‘OK this is a really great app but I just want to change a little bit’ that is usually not possible.”
Schoening names Ink & Switch as the research-tradition anchor for the malleable-software pattern (citing Jeffrey Litt as a daily collaborator). The wiki’s first explicit lineage anchor for malleable software as a pre-AI tradition that the current AI tooling is operationalising.
Additional substantive contributions from the back half (sampled, not read in full):
- “First 10% of every project is now free” (the chapter title suggests this is the central later-stage argument).
- “Tiny core” theory of great products (iPhone multitouch / GitHub PR / Notion blocks / Dropbox menu-bar icon) — products with a single tiny dense atomic unit that everything else composes around.
- SaaSpocalypse debate (~24:00) — Schoening’s hot take that the all SaaS dies because everyone builds their own with AI framing is overstated.
- Notion AI (~48:38) — Schoening’s account of why Notion’s AI launch was successful.
- Building taste through iterations (~56:40) — agency + iterations + AI tooling = taste compounds.
- Jobs-to-be-done framework (~65:06) — still load-bearing in product design even with AI tooling.
- Hot take on universal basic income (~67:28) and What Max would do with AGI (~69:26) — extended forward-looking musings; not central to the episode’s titular thesis.
Caveats. Long-form podcast on a vendor-adjacent channel (Notion is a paying sponsor pattern for product-leader podcasts; Schoening is the head of product at Notion plugging Notion’s success). Treat the Notion AI was a successful launch claims as motivated. The agency as the differentiator thesis is a strong-opinion-loosely-held founder/product-leader claim, not an empirical finding — the wiki anchors it as a vantage rather than as evidence. The Eric Lou wants to be hired in the first five vignette is a narratively-compressed retelling and should not be over-interpreted as a literal HR practice.
The wiki source page focuses on the first ~22 minutes (chapters 1–10) where the agency thesis is fully developed; the back half (~22:00–87:22) is summarised by chapter title without close reading — future ingests that need the SaaSpocalypse or tiny core claims should return to the raw transcript.
Why this matters in the corpus
This source is the wiki’s clearest single articulation of agency-as-the-durable-AI-era-skill at the practitioner-altitude — distinct from the taste / judgement framings that occupy the same conceptual space:
| Framing | Source | What it means |
|---|---|---|
| Taste | Karpathy; Hu | Discernment of quality output — the eval-of-the-eval |
| Judgement | Forsgren (“delegate tasks, not judgement”) | Knowing what to delegate vs what to keep |
| Agency (this source) | Schoening | The disposition to act / make / change things — prior to taste or judgement |
Schoening’s agency sits upstream of taste and judgement — you can have taste without doing anything with it, and you can have judgement without exercising it; agency is the disposition to act on either. The wiki’s durable-skills page now has its first explicit agency-anchor and can frame agency as the load-bearing precursor to taste-and-judgement rather than as a synonym for them.
The Eric-Lou and Brian-Leven vignettes are the wiki’s most concrete operational worked examples of what agency looks like inside a mature product org. Together with [[2026-05-19-mittal-yhangry-private-chef-all-in-on-ai-agents|Mittal’s I fired my tech lead because he didn’t know what skills was]] vignette, the wiki now carries a paired what-thriving-looks-like (Schoening) ↔ what-firing-looks-like (Mittal) anchor on the tool-fluency-as-org-design-criterion thesis at small-to-medium operating scale.
The malleable software lineage citation (Ink & Switch + Jeffrey Litt) opens an ingest target for the pre-AI design tradition that the current AI tooling is operationalising — a useful counterweight to the AI tooling is unprecedented framings the wiki currently carries.
What was actually ingested
The full ~87:22 transcript was read closely for chapters 1–10 (intro through malleable software); chapters 11–28 (back half, including SaaSpocalypse / Notion AI / token spend / AI stack / UBI / AGI hot takes) were skimmed via the chapter outline + description but not read line-by-line. The agency thesis is fully developed in the first 22 minutes; the back half adds breadth but not central conceptual weight.
The transcript renders Brian Leven correctly; the Ninos employee-nickname (“once you’re boil — I would say one example would be someone like Brian Leven”) — Schoening starts to say a hazing-style word and self-corrects to “Ninos are notion employees, sorry” — is captured.
Linked entities and concepts
Entities promoted by this source:
- Lenny’s Podcast — already entity. Bumps source-count.
Dangling — single-source mention, deferred:
- Notion — already first-mentioned across the wiki as a product reference; this is the first wiki ingest with substantive Notion-as-org content. Promote on second source.
- Max Schoening — head of product at Notion; first wiki mention.
- Brian Leven — Notion designer-engineer-recruiter; named in the agency vignette.
- Eric Lou — Notion PM who shifted from PRDs → Figma → prototype-in-code.
- Lenny Rachitsky — Lenny’s Podcast host; first wiki mention by name.
- Ink & Switch — research lab on malleable software; first wiki mention.
- Jeffrey Litt — Notion-collaborator on malleable software; works with Ink & Switch; first wiki mention.
- Dieter Rams — design philosopher; chapter title mentions him (~20:43); not read in detail.
Concept pages touched:
- durable-skills — major addition: agency as the load-bearing durable AI-era skill, distinct from taste and judgement. The Eric-Lou and Brian-Leven vignettes as the wiki’s clearest practitioner-altitude worked examples of agency-in-action.
- ai-employment-effects — adds the role-merging concern (we will lose specialists) and the prototype-vs-engineering distinction at AI-era scale — the 3D-printed-prototype-vs-100M-user-factory metaphor as a load-bearing framing for what gets lost when everyone becomes a builder.
- agentic-engineering — adds Schoening’s agency / disposition to make framing as the practitioner-side counterpart to Karpathy’s agentic engineering is a discipline framing. Karpathy names the discipline; Schoening names the disposition that produces practitioners of the discipline.
- enterprise-ai-adoption — adds the malleable software tradition as a long-arc design philosophy that the current AI tooling is operationalising. The Ink & Switch lineage is a non-AI-native intellectual tradition that the wiki has not previously anchored.
- vibe-coding — adds the “first 10% of every project is now free” chapter-title claim (not read in full but flagged for future ingest); the designers prototyping in the terminal worked example at Notion’s product-org altitude.
Source quality
- Channel: Lenny’s Podcast — product-management podcast; long-form interview format; high audience overlap with the wiki’s existing product-leader-focused sources (Nika, Singhal).
- Format: ~87-minute conversational interview; chapter-marked; ASR transcript only (no manual captions).
- Empirical anchors: Schoening’s career trajectory (Google PM → Heroku design → GitHub VP Design → 2× founder → Notion head of product) is verifiable from public LinkedIn/personal-website data. Notion AI’s launch success is not quantified in the episode.
- Bias / motive: Schoening is the head of product at Notion appearing on a product-management podcast. Treat Notion’s product org runs this way claims as accurate-by-self-report but motivated; treat the agency framework as Schoening’s earnest synthesis-of-his-own-experience rather than as an empirical finding.
- Transcript provenance: youtube-transcript-skill (Playwright path); ASR-only English captions used. Standard rolling-caption dedupe applied.