Aaron Epstein | Lightcone Summary

1. Key Themes

HTML/CSS as the Strategic Foundation for Agentic Design Tools

Paper's core architectural bet is that using the browser's rendering engine (HTML and CSS) instead of a custom canvas engine makes the tool radically more compatible with AI agents. This is the central technical insight of the entire episode.

"Historically, a lot of design tools have custom rendering engines. And those are a lot harder for agents to use. And so you'll see higher token spend, slower times, lower accuracy. So what we're building with Paper is basically... using HTML and CSS as the rendering engine... agents understand HTML and CSS really well... you get less weird hallucinations." — Stephen Haney 00:01:19

Design as a Competitive Moat in the Age of AI-Generated Slop

Stephen argues that as AI lowers the barrier to shipping software, design quality becomes a more powerful differentiator, not a lesser one. Companies that settle for unedited AI output will blend into a sea of sameness.

"If you look at every great company of the last 10, 20 years, basically all of them have exceptional design. And so I think if you want to be one of those really great companies, you have to put the care in on design and make sure that you stand out." — Stephen Haney 00:33:18

"I haven't felt like the software is getting better necessarily. I haven't had a great software experience in the last six months that blew me away. And so it's important for us... I want to help designers and people with taste keep up with all of the software that's being created." — Stephen Haney 00:04:27

The New Agentic Design Stack

Paper is positioning itself as the visual communication layer within a multi-agent developer stack: Cursor/Claude Code for coding, Paper as a visual interface for agents, and GitHub/Conductor for code review and deployment. This represents a new category of tool — the visual harness for agents.

"The way we see it is kind of the new agent stack. You'll see Cursor, Claude Code, you have multiple agents. You have Paper as a visual way to communicate with the agents. And then maybe you have a GitHub or the code review. That's what we see builders using now." — Stephen Haney 00:12:41

Overnight Agentic Design Generation as an Emerging Workflow

A non-obvious behavioral pattern is emerging: designers are leaving agentic loops running overnight to generate hundreds of design variations, then curating in the morning. This is changing the nature of design from creation to curation.

"We see a lot of folks are leaving these running overnight. They're setting up loops of an agent just generating hundreds of variations. And it's actually a challenge for us to be agent scale with our servers and everything. They're doing a lot more work than I thought they were going to be doing. And it's so cool. And now they come in and they curate." — Stephen Haney 00:20:41

The Code Base as the Single Source of Truth — Eliminating Dual Design Systems

One of the most practically important themes: modern agentic teams are abandoning the dual-maintenance of a design file and a code base, instead treating the code as the only source of truth. Paper enables this by making the canvas and the repo interoperable.

"We see so many teams maintain two copies of their design system. And it's not a tractable problem to keep them in sync. It's impossible, really. And so you just have a lot of maintenance costs with keeping both updated and you just have accuracy problems. So what we see from the newer companies, the more agentic builders, is they are treating the code base as the source of truth." — Stephen Haney 00:16:04

Building Marketing Assets as Products Instead of Buying Ads

Paper explicitly uses its shader library, its new Paper Mono font, and other creative tools as marketing vehicles rather than spending on paid advertising. This content-as-product approach builds authentic community and brand equity simultaneously.

"The way we think about the shaders is we spend our — this is our marketing budget. We don't think of this as product development. We think of it instead of buying ads, let's make this useful library for the world and spend that effort and that money on building the shader library. And we get a lot of attention. It works really well. Similarly, we develop a font. Paper Mono is a new font that we're launching." — Stephen Haney 00:10:11

Values-First Community Building Before Product Completeness

Paper reached 25,000 Twitter followers before having a fully featured product by authentically communicating shared values with its target audience — designers who love their craft. This is a replicable go-to-market principle.

"We didn't have a product for the first year and we still got to about 25,000 Twitter followers just by talking about our values. I'd go on podcasts and be like, I really like design. Do you like design? Being values aligned with your market, they will forgive so much about your product missing things. They'll wait for them to come out. They'll root for you." — Stephen Haney 00:47:37

Human Judgment Remains Irreplaceable for Organizational Design Decisions

Stephen draws a clear and important distinction: AI can improve tactical execution (font weights, spacing, contrast) but cannot replicate the organizational, stakeholder, and strategic dimensions of design work, which are inherently human.

"A lot of design is actually a very human task of decision-making in an organization. It's showing people comps and people can be like, oh, not like that. The role of design in an org is decisions and stakeholders and bringing in requirements and problem space. And those are not things that agents have any skill at, really." — Stephen Haney 00:41:53

2. Contrarian Perspectives

Reading Every Single Line of Code at a 12-Person AI-Native Company Is Not Slow — It's the Moat

While the broader industry narrative says vibe-coding and shipping fast without reviewing AI-generated code is the winning approach, Paper reads every line multiple times. Stephen argues this deliberate precision is why people choose Paper over feature-rich competitors.

"We have competitors that have every single feature, but they don't have the attention or people don't care. And for whatever reason, people want Paper to be better. They don't want the competitors that have every single feature to be better. And I think that's because we care. And we're putting in that precision, that care about these things. So even if we're moving a little slower, people still care more about it." — Stephen Haney 00:46:26

You Cannot Vibe-Code a Serious Design Tool

Against the cultural moment of AI-generated everything, Stephen flatly asserts that some categories of software still require extremely careful human engineering — and that AI coding agents are not close to being able to build tools like Figma unassisted.

"You can't vibe code a design tool at the level of a Figma today. Maybe you can eventually. And so it takes a lot of really precise, really careful work. It's very multidimensional in terms of the systems. And so we just have to be really careful. And we want to make sure our experience — designers need 120 FPS." — Stephen Haney 00:44:03

There Is Less Change Coming to Design Than People Think

While the industry hypes AI as transforming everything, Stephen argues that most of what design actually does inside organizations — problem space exploration, competitive analysis, executive stakeholder communication — is fundamentally not changing.

"I think a lot of design isn't changing. In orgs, it's like the function of design is still really important as a way to explore problem space, competitive analysis, executive stakeholders, things like this. We're not seeing that change at all. And sometimes I think there's less change coming than people think, actually, because I think that a lot of design is a very human problem." — Stephen Haney 00:53:40

AI-Generated Design Defaults Are Actively Harmful to Startups

The standard advice to founders is to ship fast and use whatever tools work. Stephen argues that accepting unedited AI design output actively undermines brand credibility, trust, hiring, and long-term company building.

"I don't think it helps you stand out from the crowd. I think it makes you look like one of a million other projects and it makes you look less intentional... it even impacts who you hire and who's attracted to your company for recruiting and the talent you can attract." — Stephen Haney 00:33:18 / 00:37:27

Models Have Aesthetically Distinctive Bad Habits That Require Deliberate Human Correction

Rather than AI taste improving on its own, Stephen argues that the most impactful intervention is a small set of specific, teachable rules (fewer font sizes, lower font weights, less purple, no unnecessary icons) that humans must actively impose on model output.

"Models love bold. They love bold styles and really try to pull your font weights back. It's so tempting to bold things. Just try to pull them back as light as you can get and it'll just magically look better designed. Another thing is too many sizes. Models love having five, six, seven, eight different font sizes. Try to work with three." — Stephen Haney 00:25:09

3. Companies Identified

Paper

AI-native design tool using HTML/CSS as its rendering engine, with an MCP server for desktop agentic workflows, shader library, image generation, and Chrome extension for importing live sites into the canvas. Why mentioned: It is the subject company, described as the fastest-growing new entrant in the design tool market since Figma, with a cult-like early adopter community and measurable Ramp subscription data showing it is on par with Sketch.

"Paper is right there with Sketch. And so we're the first company since Figma's come out to make a dent in that market at all. And it's growing very quickly, which is really encouraging." — Stephen Haney 00:48:34

Figma

The dominant professional design tool, owned by a large enterprise user base. Why mentioned: Benchmark competitor; cited as the tool Paper is positioning against, with the observation that as Figma has moved upmarket and into enterprise, it has created space for a new tool focused on the individual designer.

"They have a huge company now and they're doing very well with enterprise companies. And we were thinking, well, maybe there's time for a new tool to focus on the designer." — Stephen Haney 00:02:30

Conductor

An agentic coding tool that integrates push/merge code workflows directly inside the app. Why mentioned: Called out as the primary paired tool with Paper in agentic design-to-code workflows; described enthusiastically by both Stephen and Aaron as a best-in-class product gaining rapid community adoption.

"Conductor and Paper are this pairing right now. A lot of people in the community are using both together. I love Conductor. It's a great, great product." — Stephen Haney 00:13:22

WorkOS

Auth and identity platform for developers. Why mentioned: Acquired Paper's predecessor company (Modulz / Radix UI); Paper now uses WorkOS for its own authentication, and it was described as an "amazing auth and identity company."

"We sold this company to WorkOS, which is an amazing auth and identity company. We use WorkOS for Paper." — Stephen Haney 00:50:14

Ramp

Corporate spend management platform. Why mentioned: Ramp's AI-enabled subscription data from company credit cards was cited as third-party evidence of Paper's market penetration, showing it is already competitive with Sketch in terms of paid subscriptions.

"Ramp has been releasing the data of actual Ramp subscriptions on their credit cards. And if you ask it about design tools, it's Figma, obviously. Sketch is there. And then Paper is right there with Sketch." — Stephen Haney 00:48:34

Cursor

AI-powered coding IDE. Why mentioned: Described as one of Paper's primary agentic harness tools for production coding, praised specifically for being "not too lazy" and for spinning up sub-agents for complex tasks.

"I find Cursor is a really good harness. It's not too lazy. It'll often spin up sub-agents to get jobs done." — Stephen Haney 00:23:13

Anthropic / Claude Code

AI coding agent and underlying model provider. Why mentioned: Claude Code's emergence in December 2025 was cited as the pivotal market moment that convinced Paper to build a desktop app and MCP server; Anthropic's models noted for best visual taste.

"Last December of 2025, the Claude Code explosion happened. And I didn't think designers were going to be in terminals. I never thought that would be a thing we'd see. But we clearly did. And so that was the moment where we were like, we got to get onto the desktop." — Stephen Haney 00:12:11

Fable (model)

An AI model referenced for its visual design capability. Why mentioned: Described by Stephen as the current best-in-class model for visual taste and design thoughtfulness, though noted to be very slow.

"Fable is a wonderful... I think that's probably the best visual taste right now. It feels like it actually put some thought into its decisions compared to previous models." — Stephen Haney 00:22:45 / 00:41:28

BugBot

Automated PR code review tool. Why mentioned: Paper uses it for pull request reviews; described as useful for catching bugs even though it is "a little too pedantic sometimes."

"We use BugBot for our PR reviews, which has been really helpful. It's a little too pedantic sometimes, but it still catches bugs, which is good." — Stephen Haney 00:46:26

Vercel

Cloud platform for frontend deployment. Why mentioned: Cited as the destination where design prototypes are currently scattered and disorganized, motivating Paper's upcoming prototype-organization feature.

"People are building all these prototypes all over the place. And they're ending up on Vercel or their HTML files. And people don't know how to leave feedback onto them." — Stephen Haney 00:54:40

Linear

Developer productivity / project management tool known for its distinctive design aesthetic. Why mentioned: Cited as the company whose design style was so influential it became encoded into AI models as a default — the origin of the "purple gradient" AI design tell.

"Linear made this style great. And then I think it got encoded into the models for the next..." — Stephen Haney 00:36:50

Radix UI (Modulz / Radix)

Open-source React component library, built as a side project of Paper founder's prior company Modulz. Why mentioned: Became the default React component library; demonstrates Stephen's track record of building tools developers and designers adopt at scale.

"Radix became this default React component library... it has amazing product market fit. We just didn't see a way to build a venture scale business around it. And so we thought, let's just make it free and give it away." — Stephen Haney 00:50:14

Quarter (poster studio)

A design team making highly regarded editorial posters, referenced for their earnings-report poster series. Why mentioned: Cited as a real-world customer who migrated their poster production to Paper to achieve a distinctive retro visual style efficiently.

"Quarter, they do I think it's like earnings reports. And they've transitioned a lot of these to be built in Paper now because they can get this cool retro vibe really fast." — Stephen Haney 00:07:12

Legion Health

Online psychiatry and medication management platform (submitted by a user for design review). Why mentioned: Used as a live design critique example; identified as likely a YC company that would benefit from adding the YC badge as social proof and tightening design to increase brand trust in a high-stakes healthcare context.

"This is an area where you have to have trust, right? The design elements — the alignment of things, the contrast choices — they feel maybe vibe coded. And then your mind goes to what else don't they spend attention on?" — Stephen Haney 00:17:57

Cytex

B2B field infrastructure operations platform targeting construction/heavy equipment sectors. Why mentioned: Live design critique example; used to demonstrate how applying the three-font-size / no-bold rule dramatically improves perceived professionalism and trustworthiness of an AI-generated site.

"All we did was pull back the weight of the font and we consolidated how many sizes there are. To me, it looks much more designed and more intentional." — Stephen Haney 00:32:11

Mereda

Global payments wallet enabling QR code-based local payments across Asia and Latin America. Why mentioned: Live design critique example; praised for its exciting international energy and unique QR code integration, with feedback to move that QR code into the hero and make the hero work harder.

"I like the excitement. I like the energy. The vibe is if we can make it a little more trustworthy... I feel like this hero could be working a lot harder for them." — Stephen Haney 00:39:34 / 00:40:08

4. People Identified

Stephen Haney

Founder and CEO of Paper; previously co-founded Modulz (which built Radix UI and was acquired by WorkOS). Lifelong designer and builder of creative tools. Why mentioned: Guest and subject of the episode; demonstrated deep founder-market fit, a replicable community-building playbook, and specific design principles that are immediately actionable.

"This is where I had the strongest founder market fit because I love building tools for creative professionals. It's something I'm going to wake up every day and like, I'm not going to have a bad day building creative tooling." — Stephen Haney 00:02:07

Agu (Paper's Brand Designer)

Lead brand designer at Paper; described as primarily a designer who codes a little but is not a programmer by trade. Why mentioned: Cited as the canonical example of a designer using agentic workflows to ship an entire animated marketing site end-to-end in one week — something that would have been impossible to do on that timeline previously.

"He made our entire website in about a week before this last launch, including shipping it, including all of the animations. And he designed everything in Paper. No one else looked at this. So Agu did this end to end in a week." — Stephen Haney 00:45:00

Ev (YC's Lead Designer)

Lead designer at Y Combinator. Why mentioned: Named as a power user of Paper at YC, creating custom animated shader-based acceptance tickets for Startup School, and was in New York City when she encountered someone wearing a Paper Mono shirt in the wild — illustrating the depth of community Paper has built.

"Ev, who leads design here at YC, has actually created custom animated shaders with the YC branding, with people's names on them, tickets that are animated when people get accepted and send it to them." — Aaron Epstein 00:08:18

Wes (referenced by last name only)

Designer/developer commentator active on social media (likely Wes Bos, given the context). Why mentioned: His tweet identifying the "four horsemen of the apocalypse" of AI design tells — all-caps kickers with extra letter spacing, number badges with icons, the purple gradient, and dark/light mode toggle — was cited as a precise cultural reference.

"Wes had a great tweet recently showing the four worst men of the apocalypse, which is... the all uppercase little kickers. The badges with the extra little icons. And of course the purple gradient." — Stephen Haney 00:36:37

5. Operating Insights

The "Bottleneck First" Product Prioritization Framework

Stephen articulates a simple but powerful operating rule: identify the single constraint limiting the business right now, and focus all resources on removing it — not on a sophisticated roadmap or brilliant strategy.

"Bottlenecks first. Huge, huge philosophy I really believe in. For us right now, the bottleneck is early adopters are using Paper. The people that are more, maybe not an early adopter type are very used to what they're doing now. And they're missing features like comments and components. And so now our bottleneck is actually features again. And so now it's like, well, that's what we got to go do next." — Stephen Haney 00:51:11

Mandate Daily User Conversations in Year One — Then Embed It Culturally

Stephen ran a personal rule of talking to one designer every single day for the entire first year, creating a constraint that forced ongoing market immersion. This was then operationalized into an onboarding expectation for every new hire.

"I had a rule. The first year of the company, I had to talk to a designer every single day. And I did it all the way through the year... When people onboard, it's like, yeah, you have to be in the Discord talking to people or in the Slack. It's really, really important." — Stephen Haney 00:51:40 / 00:52:46

Encode Senior Designer Judgment Directly Into Agent System Prompts

Rather than accepting model defaults, Paper systematically distilled the expertise of senior designers into model instructions to prevent the most common AI design mistakes. This is a replicable template for any team deploying agents in creative or high-quality domains.

"In Paper, we actually give instructions to the model not to do these common mistakes. And sometimes people go like, what's your secret sauce? And it's just like, honestly, all we're doing is these basic rules of typography and contrast. And we had the more senior designers on the team distill their knowledge into the instructions for the model." — Stephen Haney 00:31:32

Use Agentic Comment-Based Workflows to Scale Design Review

Rather than iterating through a prompt loop, leaving comments on canvas frames and then instructing the agent to "address all comments" allows a designer to batch and queue work asynchronously — unlocking overnight generation cycles and freeing daytime focus.

"I love leaving comments. I think it's just the best way to do things. You can imagine having a big design and leaving all of your feedback and then just setting your agent on it, letting it spin. We see a lot of folks are leaving these running overnight." — Stephen Haney 00:21:11 / 00:31:04

Stay Small and Elite Deliberately — Fewer People, Less Communication Overhead, Better Decisions

At 12 people, Paper consciously avoids hiring to keep communication costs low and quality control high, relying on agents to multiply individual output rather than headcount to multiply throughput.

"A small elite team. 12 people. Keep the talent bar extremely high. I think when you're building companies, that's how you should try to do it these days and use the agents to accelerate the people. But if you have less people, you have less communication burden. You have less of these other costs to go through." — Stephen Haney 00:44:33

6. Overlooked Insights

The Prototype Organization Problem Is a Massive, Unsolved Coordination Gap

Mentioned in a single paragraph near the end of the conversation, but enormously significant: as vibe-coding proliferates, teams are generating enormous numbers of prototypes scattered across Vercel deployments and HTML files with no shared home, no feedback mechanism, and no way to track which version anyone is discussing. Paper is quietly solving this as its next big feature — and whoever owns prototype organization may own a critical layer of the new development workflow.

"People are building all these prototypes all over the place. And they're ending up on Vercel or their HTML files. And people don't know how to leave feedback onto them, how to even find which one someone's talking about. So what we're going to do is we're going to put, basically, you can iframe things onto the Paper canvas as a place to keep your prototypes kind of organized. And then people can leave comments as they're used to doing on the prototypes." — Stephen Haney 00:54:40

This is the seed of a Loom-for-prototypes or a Notion-for-deployed-previews — a coordination layer that will become more critical as AI accelerates the proliferation of built artifacts that teams cannot manage.

Ramp's AI-Searchable Subscription Data Is Now a Real-Time Market Intelligence Layer

Mentioned almost as a throwaway data point to substantiate Paper's growth, but the underlying mechanism is significant for investors: Ramp now offers an AI interface to its anonymized corporate credit card subscription data, making it possible to query real-time software adoption trends across thousands of companies. This is a qualitatively new kind of market signal that did not previously exist in accessible form.

"Ramp has been releasing the data of actual Ramp subscriptions on their credit cards. And if they released an AI you can chat with about the Ramp data. And if you ask it about design tools, it's Figma, obviously. Sketch is there. And then Paper is right there with Sketch." — Stephen Haney 00:48:34

For investors and operators, Ramp's chat interface over subscription data is now a viable due diligence tool for tracking software market share in near-real-time — something that previously required expensive third-party data providers or guesswork from app stores.