5 guests 5 episodes 3,241 words

The Taste Question: Can You Train a Sixth Sense, or Does It Arrive Fully Formed?

Is taste a skill you build through reps and exposure, or is it founder-level magic that can't be manufactured?

Guillermo Rauch built Vercel on the premise that taste is a muscle you can measure, and made "exposure hours" an internal operating principle. Stewart Butterfield traces the word to its etymology — chefs improve, so can product people — and made a Vancouver umbrella story the centerpiece of Slack's new-hire ritual. Krithika Shankarraman, the first marketer at both Stripe and OpenAI, thinks taste will be the primary differentiator in an AI-saturated market. Then there is Peter Deng, who has managed hundreds of PMs across Instagram, Uber, Facebook, and OpenAI, who watched Joanne Jang work and concluded she was the only person he had ever met with her particular combination of technical depth and product taste — so rare that he built a new role around her.

Dylan Field sits in the middle, describing his own "sixth sense" as a hypothesis generator fed by obsessive ingestion. On paper, he agrees with the teachable camp. In practice, he operates at a volume of signal almost no non-founder can replicate. The real question is not whether taste can be learned. It is whether the mechanism people propose for learning it generalizes — or quietly requires founder-level surface area.

Is taste a skill you build through reps and exposure — something any motivated product person can compound through deliberate practice — or is it founder-level magic that clusters in a handful of people and has to be structurally protected when you find it, because you can't manufacture it on demand?

The 3 Positions

1 / 3

Taste Is Trainable, and Almost Nobody Bothers

Stewart Butterfield · Guillermo Rauch · Krithika Shankarraman

Stewart Butterfield — Founder of Flickr and Slack, two of the rare B2B products widely cited for craft-forward polish, and one of the few enterprise founders who publicly writes about taste as a discipline.

Butterfield starts from the word itself. "Taste" is literally what you develop by putting food in your mouth over time, and chefs measurably improve with training. He concedes the innate edge, but refuses to let the concession collapse into fatalism.

"So one is can you learn to develop it? I think so because the word literally comes from experiencing food and putting stuff in your mouth. And can people become better chefs with training? Yes, absolutely. Undoubtedly, some people have a natural advantage and are born with this ability to make discernments that are difficult for other people to make and stuff like that. But you can definitely practice and you can definitely get better."
Stewart Butterfield
▶ 00:15:49

The more interesting move is the economic one. If taste is trainable but most teams don't invest, any team that does gains a defensible commercial moat. He invokes Bezos to make the point explicit.

"The second thing I'd say is you can create a real advantage for yourself, for your product, for your company by leaning into it because most people don't have good taste and don't invest. You're probably familiar with, again, Jeff Bezos line, your margin is my opportunity and pretty obvious what he meant by that."
Stewart Butterfield
▶ 00:16:30

At Slack, Butterfield operationalized this through the Vancouver umbrella story — noticing how few pedestrians bothered to tilt their umbrella to avoid poking passersby, and turning that small act of consideration into shorthand for the kind of attention Slack wanted its people to pay.

"We noticed how few people would move their umbrella out of the way... something like one-third of the people would do it. And we had this conversation about it... It actually made it part of the new hire welcome."
Stewart Butterfield
▶ 00:16:54

The implicit theory: taste is a discipline of noticing, and noticing is teachable at the team level. Slack's obsessive polish on onboarding, empty states, and loading messages is the compounding output of trained attention in an industry that mostly optimized for "hyperrealistic work-like activities" instead.

Guillermo Rauch — Founder and CEO of Vercel and v0, whose default outputs are widely cited as the most aesthetically coherent of any AI builder.

Rauch is the archive's most explicit anti-innatist. He names the "born with taste" framing and rejects it, and unlike Butterfield he does not even offer the concession about natural advantage. What distinguishes his view is that he refuses to let taste stay aspirational — he turns it into an operating principle you can actually run.

"Taste, sometimes I think we think of as this inaccessible thing that, 'Oh, that person was born with taste.' I see it as a skill that it can develop. I think is extremely important to try lots of products. We have one of our internal operating principles as increasing exposure hours."
Guillermo Rauch
▶ 00:00:57

The mechanism is exposure hours: a KPI-able amount of time people spend watching users interact with their product and consuming great work. The premise is that taste is pattern recognition, pattern recognition compounds with reps, and the highest-leverage thing a company can do is remove friction between its people and real user behavior.

"Try to quantify how much time you expose yourself to watching how people use your products, even to watch how people use other products, and you'll develop that muscle. Taste, sometimes I think we think of as this inaccessible thing that, 'Oh, that person was born with taste.' I see it as a skill that it can develop."
Guillermo Rauch
▶ 01:09:07

"You need to get yourself out there. I think it's very important to try lots of products."
Guillermo Rauch
▶ 01:08:03

There is a second-order implication in v0's product strategy. If taste is exposure, a model trained on high-quality reference design has effectively accumulated exposure hours on the user's behalf — and tools that compress the canon can hand taste to builders who would never have acquired it through traditional apprenticeship. Rauch is not just defending a theory of taste; he is monetizing it.

Krithika Shankarraman — Partner at Thrive Capital, first marketer at both Stripe and OpenAI.

Krithika takes the teachable side but adds a stakes-raising argument: the era of cheap AI generation makes taste more valuable, not less. If every category is about to be flooded with drivel, the only companies that will distinguish themselves are the ones that show their craft.

"If anything, taste is going to become a distinguishing factor in the age of AI because there's going to be so much drivel that is generated by AI, can be generated by AI, that power is at anyone's fingertips. But truly, the companies that are going to distinguish themselves are the ones that show their craft."
Krithika Shankarraman
▶ 00:55:58

Asked how she built her own taste, Krithika refuses the "I was born this way" framing and names practices: pottery, voracious reading, wide exposure to other people's work, and a Thrive Slack channel called "share" where people post things they've seen that resonated, independent of deal flow.

"At Thrive, we have this share channel, which is just sharing things that we're seeing out into the world. It's not particularly deal flow news or competitive news or anything like that, but it is things that we have seen that resonated with us for whatever reason."
Krithika Shankarraman
▶ 00:57:10

She explicitly endorses Rauch's exposure-hours frame as the cleanest articulation of what she is already trying to do.

"There's a concept that I love that recently I learned from Guillermo at Vercel. He calls it exposure hours. That's when I asked him how to build taste, and that's kind of a value they have at their company is just increase your exposure hours to great stuff, because that is how you build taste. I love that. It's such a simple actionable thing you can do."
Lenny Rachitsky
▶ 00:56:50

The quiet worry underneath her optimism is that the next generation of marketers may never develop taste at all, because they'll outsource their inputs to ChatGPT before the muscle forms. Taste is learnable, but only if you actually do the learning.

When This Applies

when your category rewards opinionated execution over metric-driven iteration. Linear's method and Instagram's Systrom-and-Krieger origin both show that no amount of team exposure hours substitutes for a small group with durable conviction.

Taste Is Empirically Scarce — Codify It When You Find It

Peter Deng

Peter Deng — Former VP of Consumer at OpenAI, with prior product leadership at Instagram, Uber, and Facebook. Deng has worked with hundreds of PMs across four of the most consequential consumer products of the last fifteen years, which gives his "very rare" claims unusual weight.

Deng is the strongest voice on the innatist side, but his argument is not that taste is mystical. It is that taste is empirically scarce, in specific combinations, in a way that forces organizational accommodations most teams are unwilling to make. His strongest signal is his reaction to Joanne Jang at OpenAI.

"She is the only person that I've worked with that has as much technical depth as she does have product taste. And I just want to pause there. It's just truly special... She has this taste and those two things are very rare to find together."
Peter Deng
▶ 01:20:56

The revealing move is what Deng did next. If taste were straightforwardly teachable, the rational response would be to turn Jang's approach into a curriculum. Instead Deng treated her as a unique resource to be structurally protected — he had her write her own job description and built the model-designer role around her.

"It felt like we had to find a way to craft this. And sure enough, I was like, 'Hey, can you just write up a job description of what is this thing? Because there's something magical here, but I don't fully understand it.'"
Peter Deng
▶ 01:21:44

Deng makes the same move at the company level. Instagram's breakthrough, he argues, was not technology any competitor couldn't have built — it was the specific taste and conviction of two specific founders.

"It wasn't anything that any other company couldn't have done, but it was that product taste that Kevin and Mike had and conviction that there's a certain sort of vibe, if you will, that people wanted, and building that and iterating, and look at it now, it's a core part of our lives."
Peter Deng
▶ 00:25:10

Deng's view is not that taste is genetic. It is that taste clusters in specific people, in rare combinations, and that the correct organizational response is to design around the people who already have it rather than pretending you can manufacture it evenly.

When This Applies

when raising the baseline of attention across non-founders. Slack's umbrella story and Thrive's share channel work because they give people a shared vocabulary for small acts of consideration.

Looks Innate, Behaves Like Exposure — The Founder Loophole

Dylan Field

Dylan Field — Co-founder and CEO of Figma, whose personal reputation among his team is built on product intuition and a "sixth sense" for what will work.

Field occupies the most interesting position because he gets described in explicitly innatist language by his colleagues, and then refuses it. When Lenny stages the compliment, Field's answer is a surprisingly mechanical reframe.

"What I've heard from a number of your colleagues is that one of your superpowers is intuition and product taste. And someone said that you have the sixth sense for what's going to work, when you're designing Figma and you're making decisions in the product."
Lenny Rachitsky
▶ 00:10:45

"Here's my framework for it. I think intuition is like a hypothesis generator and you're constantly generating these hypotheses and others are generating hypotheses as well. And you then take these hypotheses and you put them forward and you debate them and you try to find data to support them or negate them. And then you winnow it down into what is our working hypothesis?"
Dylan Field
▶ 00:11:19

Intuition, in Field's account, is a fast pattern matcher whose outputs still have to be validated downstream. But the way he feeds the pattern matcher starts to sound less democratizing and more like something only a very small number of people could actually run.

"I definitely look everywhere trying to constantly ingest information about Figma, and it's not just Twitter/X, whatever that's called now, but anywhere on the internet, support channels, et cetera. And I'm always trying to understand. I also ask a lot of questions and I try to get to root problems and understand where people are coming from and what are they actually trying to solve."
Dylan Field
▶ 00:12:00

Field reads every Figma mention and shares them to an internal channel. He monitors support. He interrogates users past what they say they want to what they're actually trying to solve. On paper this is exposure hours on steroids — the Rauch mechanism at its logical extreme. In practice, it requires founder-level time, attention, and authority. His view is best read as the reconciliation the rest of the debate is missing: taste looks innate from outside, behaves like exposure from inside, and only really scales to people in founder-adjacent positions.

When This Applies

when leading a team that has never been deliberate about craft inputs. Rauch at Vercel and Krithika at Thrive both saw lift from formalizing what most teams leave to individual habit.

Evidence from the Archive

Slack

Slack made a Vancouver umbrella observation — that only a third of people tilt their umbrella for others — into a new-hire welcome ritual

Stewart Butterfield argues taste is literally trainable (the word comes from food) and most people don't invest, so leaning into it is a structural advantage

Stewart Butterfield · Founder of Slack and Flickr ▶ 00:15:49

Vercel

Vercel turned taste into a KPI: 'increasing exposure hours' is a formal internal operating principle, measuring time employees spend watching users

Guillermo Rauch directly rejects the 'born with taste' framing; v0's aesthetic baseline is itself evidence that exposure-trained taste can ship at model scale

Guillermo Rauch · Founder and CEO of Vercel / v0 ▶ 01:09:07

Thrive Capital

Thrive runs an internal 'share' Slack channel — things the team has seen in the world that resonated, independent of deal flow — as a live exposure-hours exercise

Krithika Shankarraman, first marketer at Stripe and OpenAI, argues AI-saturated markets make taste more valuable, and that you build it through idiosyncratic inputs

Krithika Shankarraman · Partner at Thrive Capital; formerly first marketer at Stripe and OpenAI ▶ 00:55:58

Figma

Dylan Field reads every Figma mention on the internet and posts them to an internal Slack channel — his 'sixth sense' is really exposure hours on steroids

Field's colleagues describe his taste as a sixth sense, but he reframes intuition as a hypothesis-generation engine trained by obsessive information ingestion

Dylan Field · Co-founder and CEO of Figma ▶ 00:11:19

OpenAI

Peter Deng invented a whole new 'model designer' role at OpenAI around Joanne Jang because the combination of technical depth and taste was too rare to train

After managing PMs at Instagram, Uber, Facebook, and OpenAI, Deng's population-level claim is that the taste+technical combo is empirically scarce enough to require structural accommodation

Peter Deng · Former VP of Consumer at OpenAI; Instagram, Uber, Facebook alumnus

The Synthesis

The surface disagreement is between teachable (Butterfield, Rauch, Krithika) and innate (Deng). But once you hold Field next to Deng, both camps are describing the same mechanism and drawing opposite organizational conclusions from it.

The Founder Loophole

Is exposure really teachable at normal doses?

Attention vs. Judgment

Which half of taste are you actually training?

Codify Instead of Teach

What do you do when you find a Joanne Jang?

Ira Glass Gap

Did your taste arrive before your ability?

AI-Era Stakes

Will the next PM cohort ever develop taste at all?

Rauch's exposure-hours theory and Field's obsessive Figma ingestion describe the same mechanism. The difference is dose. Field ingests at a rate only a founder can sustain because he sits at the center of a flood of user signal. The mechanism is real; the dose required to develop Jobs-level taste may only be physically achievable from a very small number of seats.

Taste is two stacked things: attention (the trainable discipline of noticing) and judgment (knowing which of your hypotheses is right). Butterfield's umbrella story and Krithika's share channel raise team attention. Nobody in the archive has a clean story for how judgment improves beyond 'ingest more and hope.' The floor rises faster than the ceiling.

Peter Deng's revealing move wasn't to turn Joanne's approach into a curriculum — it was to have her write her own job description and build the model-designer role around her. When taste clusters in rare combinations, the right organizational response is structural protection, not manufactured replication. Design around the scarce person; don't pretend you can make more.

The Ira Glass framing fits the evidence better than either pure camp: taste-as-discernment arrives early and looks innate; taste-as-execution trains over years. That's why Slack's polish and v0's house style come from people who clearly had strong judgment before they started practicing — the reps closed the gap, they didn't create the discernment.

Krithika's quiet worry: the next generation of marketers may outsource their inputs to ChatGPT before the muscle forms. Taste is learnable only if you actually do the learning. In an AI-saturated market where drivel is free, exposure hours become both more valuable and more endangered at the same time.

Which Approach Fits You?

Answer 3 questions about your situation. We'll match you to the right approach.

Question 1

What does taste development look like on your team today?

Question 2

What is your highest-leverage intervention right now?

Question 3

What kind of output do you most need more of?

Notable Absences

The Bottom Line

The non-obvious conclusion is that the teachable camp's prescriptions are all correct — exposure hours, attention rituals, share channels, root-cause interrogation — but the return is smaller and more uneven than their advocates suggest. You can raise the floor of a team's attention dramatically. You probably cannot raise the ceiling of their judgment by the same amount. Butterfield and Deng are both right in ways that only appear contradictory if you treat taste as indivisible: invest in the trainable half because most teams don't, and codify structural roles around the people who already have the scarce half.

Lenny's newsletter archive sharpens this. The Ira Glass "gap" framing from his essential-reading roundup describes taste as the thing that got you into the game in the first place — present before any reps — and the years of work as the grind required to close the distance between your taste and your ability. That fits the evidence better than either pure camp: taste-as-discernment arrives early and looks innate; taste-as-execution trains over years. Karri Saarinen's Linear method, featured in Lenny's inspiration roundup and documented in "How Linear builds product," sidesteps the teachable question entirely by building the company around a small group who already have strong opinions and letting the product inherit them — a concrete operational answer to Deng's scarcity argument.

Sources

Stewart Butterfield — "Slack founder: Mental models for building products people love ft. Stewart Butterfield" — Lenny's Podcast, November 20, 2025
Guillermo Rauch — "Everyone’s an engineer now: Inside v0’s mission to create a hundred million builders | Guillermo Rauch (founder and CEO of Vercel, creators of v0 and Next.js)" — Lenny's Podcast, April 13, 2025
Krithika Shankarraman — "Growth tactics from OpenAI and Stripe’s first marketer | Krithika Shankarraman" — Lenny's Podcast, May 25, 2025
Dylan Field — "Figma’s CEO: Why AI makes design, craft, and quality the new moat for startups | Dylan Field" — Lenny's Podcast, June 30, 2024