70 people · 61 organisations · 43 relationships

Who to talk to

Everyone here is real, named from a public source, and reachable through a channel this page names. Roles change and some records are marked low confidence — check before you write. The reasoning behind the ordering is in the companion notes.

The 15 to call first

Ranked by what a single conversation would change, not by how impressive the name is.

Researchers · public

Aspen Hopkins

Researcher — MIT CSAIL

Lead author of Contra Labs' flagship creative-preference paper and an MIT CSAIL researcher — the academic credibility of the leading competitor rests substantially on her name. Whether she is staff, contractor or collaborator is the most useful unknown in the entire map.

Ask: What the disagreement structure in designer preference data actually looked like, and whether she would collaborate with, or advise, a second entrant.

Reach: MIT CSAIL personal page and arXiv corresponding-author address on 2606.30561; academic collaboration is a legitimate opening.

Tool companies · public

Danny Wu

Head of AI Products — Canva

Canva is the only creative-tool company with a live AI Quality Evaluator posting and a Research Lead - Evaluations function, and he owns AI product quality for 240M+ users. He is the best-qualified single prospect in the buy-side map.

Ask: Whether Canva's evaluation roles are meant to build a permanent internal panel or to specify work it would rather outsource, and what design dimensions it needs scored.

Reach: Regular podcast guest (NVIDIA AI Podcast, Startup Project) and a Canva Create keynote presence — a genuinely public figure.

Practitioners · public

Ed Newton-Rex

Founder & CEO, Fairly Trained — Fairly Trained

He resigned from Stability over training data ethics and now certifies companies on it. He is the most credible single person who could either certify the client's supply terms or dismantle them publicly — and certification is a genuine competitive moat if won.

Ask: Whether Fairly Trained would ever certify an expert-data company on its creator terms rather than its corpus, and what the criteria would be.

Reach: Writes publicly at ed.newtonrex.com; speaks at academic and industry seminars; very responsive on X.

Practitioners · public

Elizabeth Goodspeed

Designer and writer; US editor-at-large, It's Nice That — It's Nice That

She has written the definitive practitioner essays on taste and on why AI cannot supply it, in the publication designers actually read. Her judgement of this company will substantially set how the profession receives it.

Ask: What would make selling aesthetic judgement to AI labs read as dignified professional work rather than as selling out — and what would make it read as the latter.

Reach: Publishes a regular column at It's Nice That and a personal newsletter; speaks at DesignThinkers and similar conferences.

Community · public

Felix Lee

Co-founder & CEO — ADPList

ADPList is a global network of senior designers with publicly bookable mentor profiles — the lowest-friction, highest-trust recruiting surface in existence for exactly this supply, and it is not a portfolio site so quality signal is stronger.

Ask: Whether ADPList would partner on a vetted evaluation panel, and what its mentors would consider fair pay for judgement work.

Reach: He maintains a public ADPList mentor profile that anyone can book — a literal open door.

Competitors · cold

Grace Li

Co-founder & CEO — Intelligence (Design Arena)

Claims $60M ARR selling human design-evaluation data on a ten-person team, and can sell geographic and temporal taste drift that a 30-designer panel structurally cannot produce. Either the strongest comparable in the category or its biggest overstatement — the client needs to know which.

Ask: What a frontier-lab contract for arena data is actually shaped like — seats, consumption, or dataset licence — and whether labs distinguish expert votes from anonymous ones in pricing.

Reach: Active press cycle since Aug 2026 (TechCrunch, Yahoo Finance) and a YC S25 network; approach founder-to-founder or via a shared YC contact.

Competitors · public

Hamidah Oderinwale

Member of Technical Staff — Taste Labs

She writes Taste Labs' public research agenda, including the Requests for Research that telegraph the company's roadmap into edit-sequence and design-trajectory capture. She is the technical mind of the nearest competitor and is publicly soliciting outside researchers.

Ask: Which of the nine open research problems they consider hardest, and whether design-intent inference from edit sequences is a shipped capability or still a wish.

Reach: The Prototype fellowship Airtable form and the research contact link on tastelabs.com are both open, public inbound channels she operates.

Researchers · public

Jordan Taylor

PhD researcher — Carnegie Mellon University

Lead author of the FAccT '26 audit showing LAION-Aesthetics filters by gender and reinforces Western and Japanese realist bias. That paper is simultaneously the best argument for the client's product and the framework by which the client's own panel will be attacked.

Ask: How to build a taste panel that survives the same audit — what pluralistic aesthetic evaluation would actually require.

Reach: arXiv corresponding-author channel on 2601.09896; FAccT 2026 is a public venue.

Practitioners · cold

Karla Ortiz

Concept artist and illustrator; named plaintiff, Andersen v. Stability AI — Independent practice

The most prominent artist litigant against generative AI training and the most effective organiser of creative opposition. She will be among the first to publicly frame a taste-data company as laundering exploitation, and knowing her exact objections is worth more than avoiding them.

Ask: What distinction, if any, she draws between selling artwork for training and selling judgement about outputs — and what terms would make the second acceptable.

Reach: Speaks publicly and often on AI and artists' rights; her public advocacy channels and interviews are open.

Tool companies · public

Noah Levin

VP of Design — Figma

Figma builds design quality entirely in-house, and he is publicly on record wanting more than AI-generated eye candy. Even if Figma never buys, his endorsement is the strongest possible signal to the designer supply that this work is legitimate.

Ask: What Figma would need to see to trust an external design-quality benchmark, and whether Config would host a session on evaluating generated design.

Reach: Figma Config speaker; publishes on the Figma blog; active on design podcasts.

Competitors · cold

Purvanshi Mehta

Co-founder, Lica World; now leading design research at Gamma — Gamma

Corresponding author of the TASTE paper and the person who supplied Contra Labs' reward-modelling capability. She has just moved inside a product company, which means she is simultaneously the most knowledgeable person about this exact product and no longer a competitor to it.

Ask: What signal in designer ratings turned out to be learnable, where existing preference models failed on design, and whether Gamma would buy taste data rather than build it.

Reach: Corresponding author on arXiv 2605.20731; also quoted in TechCrunch on the acquisition, so publicly reachable via press channels.

Model labs · cold

Robin Rombach

Co-founder & CEO — Black Forest Labs

Lead author of the Stable Diffusion work and now CEO of the lab whose own job spec says aesthetic judgement is an RL target requiring preference data. BFL is the most explicit stated demand in the entire buyer map.

Ask: Whether BFL builds its aesthetic preference data in-house in Freiburg or would buy a designer panel, and what dimensions of aesthetic quality it needs labelled.

Reach: Academic and ML conference circuit (CVPR/NeurIPS); BFL's Ashby board signals the post-training team he is staffing.

Investors · public

Sarah Catanzaro

General Partner — Amplify Partners

She co-led Taste Labs' seed and wrote the public thesis for the category, and she is also an investor in Runway. She is the single best-informed outside observer of this exact market and holds a genuine cross-cluster introduction path from investor to lab.

Ask: What she rejected before backing Taste Labs, how she thinks the category defends against arenas, and what a differentiated second entrant would have to look like.

Reach: Publishes under her own byline at Amplify and speaks frequently at data and ML events; publicly reachable.

Practitioners · cold

Scott Belsky

Partner, A24; founder of Behance; former Chief Strategy Officer, Adobe — A24

The only person who has founded the largest creative portfolio network, run strategy at Adobe through the Firefly contributor-compensation decisions, and now sits at A24 — which has a reported $75M AI venture with Google DeepMind. He is the highest-leverage single introduction in this entire map.

Ask: How Adobe actually decided what to pay contributors for training data, and whether A24's DeepMind venture needs professional creative evaluation.

Reach: Writes a public newsletter and speaks widely; Cornell Tech Council and A24 Labs are public affiliations.

Model labs · public

Souki Mansoor

Lead, Global Sora Artist Program — OpenAI

The single most relevant person at OpenAI: she runs the artist programme that turned a boycott into a negotiation, built over ~10 months of relationship-building with a $3M budget. She already does the client's hardest job — persuading working creatives to engage with a lab.

Ask: How artists were recruited and compensated, whether the programme produced any evaluation signal that fed back into training, and whether OpenAI would rather buy that relationship than build it.

Reach: Publicly quoted in Artnet on Sora Selects; speaks about the programme at art and film venues; OpenAI Forum sessions are public.

Competitors

The people already building this. Talk to them last, or never — but know them.

Name	Role	Why they matter	How to reach	Pri
Grace Li	Co-founder & CEO Intelligence (Design Arena)	Claims $60M ARR selling human design-evaluation data on a ten-person team, and can sell geographic and temporal taste drift that a 30-designer panel structurally cannot produce. Either the strongest comparable in the category or its biggest overstatement — the client needs to know which.	Active press cycle since Aug 2026 (TechCrunch, Yahoo Finance) and a YC S25 network; approach founder-to-founder or via a shared YC contact.	1
Hamidah Oderinwale	Member of Technical Staff Taste Labs	She writes Taste Labs' public research agenda, including the Requests for Research that telegraph the company's roadmap into edit-sequence and design-trajectory capture. She is the technical mind of the nearest competitor and is publicly soliciting outside researchers.	The Prototype fellowship Airtable form and the research contact link on tastelabs.com are both open, public inbound channels she operates.	1
Purvanshi Mehta	Co-founder, Lica World; now leading design research at Gamma Gamma	Corresponding author of the TASTE paper and the person who supplied Contra Labs' reward-modelling capability. She has just moved inside a product company, which means she is simultaneously the most knowledgeable person about this exact product and no longer a competitor to it.	Corresponding author on arXiv 2605.20731; also quoted in TechCrunch on the acquisition, so publicly reachable via press channels.	1
Alexandria Minetti	Researcher, Contra.Work Inc. Contra Labs (Contra.Work Inc.)	Named on both the Contra Labs paper and the Contra x Lica TASTE paper — one of very few people who has actually run designer-annotation studies at commercial scale and published the method.	Corresponding-author channels on the two arXiv papers she co-wrote.	2
Anastasios N. Angelopoulos	Co-founder & CEO Arena (LMArena / Arena Intelligence)	Runs the $100M-annualised arena that eats this market from above by collecting preference at zero marginal cost. Any pitch to a lab has to answer 'why not just use Arena', and he defines that answer.	Academic and Berkeley channels; he speaks at ML evaluation venues and publishes on statistical evaluation.	2
Ben Huffman	Co-founder & CEO Contra Labs (Contra.Work Inc.)	Runs Contra Labs, the incumbent in this exact lane, and has the largest claimed creative supply pool. His ML partner (Lica) was just acquired out from under him, so his capability stack has a visible hole right now.	No founder podcast or long-form interview exists for the Labs thesis; the only route is Contra's own site and the Ashby careers board, or a mutual through Unusual Ventures.	2
Brendan Foody	Co-founder & Co-CEO Mercor	Mercor already lists senior design experts at $150-250/hr — the generalist is not absent from creative work, it is outbidding the specialists. He sets the price the client's supply will compare against.	Prolific public speaker and podcast guest on the expert-data market; Mercor's research page is a public front door.	2
Kamryn Ohly	Co-founder & CTO Intelligence (Design Arena)	Built the arena infrastructure and the Bradley-Terry leaderboard stack across 40+ design categories — the technical answer to 'why not just run an arena'.	YC S25 alumni channels and the company's public engineering hiring page.	2
Priyaa Kalyanaraman	Co-founder, Lica World; now at Gamma Gamma	The other Lica founder, ex-Microsoft, Snap and Waymo design-AI. She has raised, pivoted twice and exited a design-data company on $4M — the cheapest available education in what this market will and will not pay for.	Quoted in TechCrunch; South Park Commons network; Gamma's public channels.	2
Spencer Mateega	Co-founder AfterQuery	Co-founder of AfterQuery, which bought the design-evaluation category cheaply by publishing UI-Bench with a 194-person expert panel and open-sourcing the framework. That is the marketing play the client should either copy or pre-empt.	Public on X and through YC W25 networks; the UI-Bench leaderboard site carries a contact route.	2
Thais Castello Branco	Founder & CEO Taste Labs	Runs the best-funded direct competitor and is the only founder in the category who has sold to frontier labs before (she led growth at Exa). Her trust-propagation recruiting model is the single most copyable idea in the field.	She podcasts publicly (AI Engineer) and posts the company thesis on X as @thaiscbranco_; approach as a peer founder, not a customer.	2
Edwin Chen	Founder & CEO Surge AI	Built the largest privately held RLHF data business by insisting on quality of annotator over volume — the same argument the client is making, already proven at $1B+ scale in text.	Writes publicly on annotation quality; Surge's blog and his public essays are the opening.	3

Model labs

Whoever runs human data or a creative programme inside a frontier or media lab.

Name	Role	Why they matter	How to reach	Pri
Robin Rombach	Co-founder & CEO Black Forest Labs	Lead author of the Stable Diffusion work and now CEO of the lab whose own job spec says aesthetic judgement is an RL target requiring preference data. BFL is the most explicit stated demand in the entire buyer map.	Academic and ML conference circuit (CVPR/NeurIPS); BFL's Ashby board signals the post-training team he is staffing.	1
Souki Mansoor	Lead, Global Sora Artist Program OpenAI	The single most relevant person at OpenAI: she runs the artist programme that turned a boycott into a negotiation, built over ~10 months of relationship-building with a $3M budget. She already does the client's hardest job — persuading working creatives to engage with a lab.	Publicly quoted in Artnet on Sora Selects; speaks about the programme at art and film venues; OpenAI Forum sessions are public.	1
Alejandro Matamala	Co-founder & Chief Design Officer Runway	A Chief Design Officer at a frontier video lab is a rare thing and exactly the right internal champion for an argument about aesthetic judgement as a training input.	Design conference circuit; Runway's public creative programming.	2
Amit Jain	Co-founder & CEO Luma AI	Luma is explicitly hiring to convert human evaluation frameworks into automated ones — the clearest published statement of the automation risk to this whole business. Understanding his reasoning is a direct test of the client's thesis.	Podcast and conference circuit; Luma's Hollywood lab is a public presence.	2
Anastasis Germanidis	Co-founder & CTO Runway	Owns Runway's model training; the person who would sign off on buying external preference data.	ML conference circuit; Runway research posts.	2
Cristóbal Valenzuela	Co-founder & CEO Runway	Runway has the deepest working relationships with professional filmmakers of any lab (AI Film Festival, Hundred Film Fund) and Amplify names Runway as a company feeling the taste problem — while also being an Amplify portfolio company alongside Taste Labs.	Frequent podcast guest (Acquired ACQ2 and others); Runway AI Film Festival is a public convening.	2
Mohammad Norouzi	Founder & CEO Ideogram	Ideogram is the image lab closest to graphic design — typography and text rendering are its whole differentiator, which makes designer judgement more obviously valuable there than anywhere else.	Speaks at conferences (Web Summit Vancouver); ex-Google Brain academic channels.	2

Tool companies

The app layer. The one confirmed human-data budget in the market sits here.

Name	Role	Why they matter	How to reach	Pri
Danny Wu	Head of AI Products Canva	Canva is the only creative-tool company with a live AI Quality Evaluator posting and a Research Lead - Evaluations function, and he owns AI product quality for 240M+ users. He is the best-qualified single prospect in the buy-side map.	Regular podcast guest (NVIDIA AI Podcast, Startup Project) and a Canva Create keynote presence — a genuinely public figure.	1
Noah Levin	VP of Design Figma	Figma builds design quality entirely in-house, and he is publicly on record wanting more than AI-generated eye candy. Even if Figma never buys, his endorsement is the strongest possible signal to the designer supply that this work is legitimate.	Figma Config speaker; publishes on the Figma blog; active on design podcasts.	1
Alexandru Costin	Vice President, Generative AI and Sensei Adobe	Owns Firefly and Adobe's ethically-sourced training data programme — the only company that has actually paid creative contributors for training data at scale, and therefore has both the budget precedent and the reputational scar tissue.	Publishes under his own byline on the Adobe blog; speaks at Adobe MAX and Summit.	2
Amjad Masad	Co-founder & CEO Replit	Replit's generated UIs sit near the bottom of UI-Bench, so it has a measured, public design-quality problem. He was also an angel investor in Lica, which means he already believed design data was worth backing.	Extremely active on X and podcasts; a genuinely open public channel.	2
Dylan Field	Co-founder & CEO Figma	Controls Config, the single densest annual gathering of exactly the designers the client needs to recruit, and sets Figma's build-versus-buy posture on AI quality.	Config keynote; extensive public podcast and press presence.	2
Grant Lee	Co-founder & CEO Gamma	Just bought a design research lab rather than buying design data — the clearest recent datapoint on how an app-layer company solves the taste problem, and a direct competitor for the same talent.	Quoted in TechCrunch; active founder press and podcast presence.	2

Researchers

The people who built the aesthetic reward models and know what the annotation really cost.

Name	Role	Why they matter	How to reach	Pri
Aspen Hopkins	Researcher MIT CSAIL	Lead author of Contra Labs' flagship creative-preference paper and an MIT CSAIL researcher — the academic credibility of the leading competitor rests substantially on her name. Whether she is staff, contractor or collaborator is the most useful unknown in the entire map.	MIT CSAIL personal page and arXiv corresponding-author address on 2606.30561; academic collaboration is a legitimate opening.	1
Jordan Taylor	PhD researcher Carnegie Mellon University	Lead author of the FAccT '26 audit showing LAION-Aesthetics filters by gender and reinforces Western and Japanese realist bias. That paper is simultaneously the best argument for the client's product and the framework by which the client's own panel will be attacked.	arXiv corresponding-author channel on 2601.09896; FAccT 2026 is a public venue.	1
Baiqi Li	Researcher, GenAI-Bench Carnegie Mellon University	First author of GenAI-Bench, whose 1,600 prompts came from professional designers — and which Google DeepMind then used as the prompt set for Imagen 3. That is the single clearest proof that a lab will adopt designer-authored artefacts.	arXiv corresponding-author channel on 2406.13743; GenAI-Bench project site.	2
Christoph Schuhmann	Co-founder and organisational lead LAION	He personally built the LAION-Aesthetics predictor on ~5,000 ratings and chose its architecture by visual inspection. That model still filters most open image training data — the client's entire product is the argument that his shortcut was not good enough.	LAION Discord and open community channels are genuinely open; he responds publicly.	2
Jenia Jitsev	Co-founder, LAION; scientist, Jülich Supercomputing Centre LAION	The most academically established LAION co-founder and the credible co-publication partner for a replacement aesthetics predictor built on expert rather than volunteer ratings — which would displace the artefact the CMU audit just discredited.	Academic corresponding-author address on the LAION-5B paper; NeurIPS and ICML circuit; LAION's open community channels.	2
Jiazheng Xu	PhD researcher, Tsinghua University (KEG) Tsinghua University	First author of both ImageReward and VisionReward — the two most influential open aesthetic reward models. Every serious buyer's internal baseline is descended from his work, so he defines the bar a commercial taste model must beat.	arXiv corresponding-author channels; the THUDM/zai-org GitHub repositories are actively maintained public channels.	2
Zhangchen Xu	PhD researcher, University of Washington University of Washington	Co-author of the Visual Aesthetic Benchmark, which found a substantial gap between frontier models and expert human judges on aesthetic comparison. That gap is the client's market, quantified by someone with no commercial stake.	arXiv corresponding-author channel on 2605.12684.	2
Björn Ommer	Professor, Computer Vision & Learning, LMU Munich LMU Munich (Computer Vision & Learning Group / CompVis)	Supervised all three Black Forest Labs founders and the Stable Diffusion work. He is the academic hub connecting the open-image-model world to the European labs, and an unusually accessible route to BFL.	University faculty page, public lectures and academic conferences.	3
Haonan Zhu	Research, Lica World (now Gamma) Gamma	First author of the TASTE paper — the only published designer-annotated multi-dimensional preference dataset in existence. He built the thing the client wants to sell.	arXiv author channels on 2605.20731.	3
Hongsheng Li	Professor, CUHK MMLab CUHK MMLab	Senior author on HPS v2 and head of one of the most productive multimodal groups in Asia; a route into both academic credibility and the Chinese lab ecosystem.	Faculty page; CVPR/ICCV circuit.	3
Maarten Sap	Assistant Professor Carnegie Mellon University	Co-author of the LAION audit and a leading researcher on annotator subjectivity and disagreement — the exact methodological problem at the heart of selling taste.	Faculty page and academic conferences; publicly reachable by academic email.	3
Omer Levy	Associate Professor, Tel Aviv University Tel Aviv University	Senior author on PickScore; a well-known NLP and preference-learning researcher whose endorsement would carry into the labs.	University faculty page; academic conferences.	3
Percy Liang	Professor, Stanford; Director, CRFM Stanford CRFM	Senior author on HEIM and the most institutionally influential voice on how foundation models get evaluated. If aesthetic evaluation becomes a standard axis, it goes through his framework.	Stanford faculty page; CRFM public events and workshops.	3
Radha Poovendran	Professor, University of Washington University of Washington	Senior author on the Visual Aesthetic Benchmark; his group's involvement gives the aesthetic-evaluation gap an institutional home outside the commercial players.	UW faculty page; academic email.	3
Sam Jung	Researcher (AfterQuery / UPenn) AfterQuery	First author of UI-Bench and the person who designed the blinded pairwise TrueSkill methodology that is now the reference standard for judging generated interfaces.	arXiv corresponding-author channel on 2508.20410.	3
Tony Lee	Researcher, Stanford CRFM (HEIM lead author) Stanford CRFM	Lead author of HEIM, which put aesthetics on the same footing as alignment and safety as an evaluation axis for image models. HEIM is the framework a buyer already knows.	Stanford CRFM channels and the HELM/HEIM GitHub repository.	3
William Agnew	Researcher Carnegie Mellon University	Co-author of the LAION audit and a persistent public critic of data practices in generative AI. He will be among the first to scrutinise a company that claims to sell taste.	Academic channels and FAccT; publicly active on data-ethics debates.	3
Xiaoshi Wu	Researcher, Human Preference Score v2 CUHK MMLab	First author of HPS v2 — 798,090 preference choices over 433,760 image pairs — the largest open human preference benchmark for text-to-image and the standard the client's dataset will be compared to on size.	arXiv corresponding-author channel and the HPS v2 GitHub repository.	3
Yuval Kirstain	Researcher, Pick-a-Pic / PickScore Tel Aviv University	First author of PickScore, the preference model most often used as a baseline. Any claim that designer data beats crowd data has to be demonstrated against his work.	arXiv corresponding-author channel on 2305.01569.	3
Yuxiao Dong	Associate Professor, Tsinghua University Tsinghua University	Senior author on ImageReward and VisionReward and a leader of the group that keeps shipping the reference reward models. Academic partnership with his group would give a new entrant instant technical credibility.	Faculty page and arXiv corresponding-author address.	3

Practitioners

Senior creatives whose participation would make this credible — and the critics who would not.

Name	Role	Why they matter	How to reach	Pri
Ed Newton-Rex	Founder & CEO, Fairly Trained Fairly Trained	He resigned from Stability over training data ethics and now certifies companies on it. He is the most credible single person who could either certify the client's supply terms or dismantle them publicly — and certification is a genuine competitive moat if won.	Writes publicly at ed.newtonrex.com; speaks at academic and industry seminars; very responsive on X.	1
Elizabeth Goodspeed	Designer and writer; US editor-at-large, It's Nice That It's Nice That	She has written the definitive practitioner essays on taste and on why AI cannot supply it, in the publication designers actually read. Her judgement of this company will substantially set how the profession receives it.	Publishes a regular column at It's Nice That and a personal newsletter; speaks at DesignThinkers and similar conferences.	1
Karla Ortiz	Concept artist and illustrator; named plaintiff, Andersen v. Stability AI Independent practice	The most prominent artist litigant against generative AI training and the most effective organiser of creative opposition. She will be among the first to publicly frame a taste-data company as laundering exploitation, and knowing her exact objections is worth more than avoiding them.	Speaks publicly and often on AI and artists' rights; her public advocacy channels and interviews are open.	1
Scott Belsky	Partner, A24; founder of Behance; former Chief Strategy Officer, Adobe A24	The only person who has founded the largest creative portfolio network, run strategy at Adobe through the Firefly contributor-compensation decisions, and now sits at A24 — which has a reported $75M AI venture with Google DeepMind. He is the highest-leverage single introduction in this entire map.	Writes a public newsletter and speaks widely; Cornell Tech Council and A24 Labs are public affiliations.	1
Jingna Zhang	Photographer; founder, Cara Cara	Built a million-artist anti-AI platform without venture funding, precisely so no investor could push it toward AI. She controls the largest concentrated population of artists who will refuse this work, and her posture will shape whether the client is seen as a fair employer or an adversary.	Public and highly visible on X and Cara; frequent press interviews.	2
Teo Connor	Vice President of Design, Airbnb; Jury President, Digital Experience Design, D&AD Awards 2026 Airbnb	A rare cross-cluster node: a big-tech design leader who is also chairing a D&AD jury. She confers credibility with both the buyer side and the practitioner supply in one relationship.	D&AD jury programme and awards events; design conference circuit.	2
Boris Eldagsen	Artist; participant, OpenAI Sora Selects Independent practice	The proof case that a lab can pay a sceptical artist without buying his silence — he took the money and still criticised the copyright position publicly. That is exactly the relationship model the client needs to be able to survive.	Publicly outspoken artist with an active exhibition and speaking presence; quoted in Artnet.	3
Cat How	CEO, Founder & Executive Creative Director, How&How; Jury President, New Brand Identity, D&AD 2026 How&How	Runs an independent studio and chairs a D&AD jury — the profile most likely to say yes early, because independents have both the credential and the commercial flexibility that holding-company creatives lack.	D&AD jury programme; studio's public channels and speaking circuit.	3
Debbie Millman	Host, Design Matters; Chair and co-founder, Masters in Branding, School of Visual Arts; President Emeritus, AIGA School of Visual Arts	Design Matters is the longest-running and most credible design podcast, and she is President Emeritus of AIGA. An appearance there is the cheapest possible route to legitimacy with senior designers at scale.	Design Matters accepts guests through public channels; SVA and AIGA events.	3
Matias Corea	Co-founder of Behance; investor and Awwwards jury member Awwwards	Co-founded the largest creative portfolio network and now judges Awwwards — one of very few people who has built creative supply infrastructure and still sits on the credentialing side of it.	Awwwards jury programme; public design and startup events.	3
Noma Bar	Illustrator and graphic designer, Dutch Uncle; Jury President, Illustration, D&AD 2026 Dutch Uncle	One of the most recognisable illustrators working, chairing the D&AD illustration jury. Illustration is the most hostile segment to AI, so his position is a direct read on whether that segment can be recruited at all.	Represented by Dutch Uncle, whose agency channel is the professional route; D&AD events.	3
Teemu Suviala	Global Chief Creative Officer, Landor; Jury President, Brand Identity Refresh, D&AD 2026 Landor	Runs creative at one of the largest global brand agencies and chairs a D&AD jury. Agency creative leadership is the fastest route to a credentialed panel, because one relationship brings a roster.	D&AD jury programme; agency and awards circuit.	3
Veronika Burian	Creative Type Director & Co-founder, TypeTogether; Jury President, Type Design & Lettering, D&AD 2026 TypeTogether	Type designers are a tiny, fully enumerable, prestigious population, and typography is where generative models most visibly fail — which makes this the cheapest high-value vertical to staff. She is its credential.	D&AD jury programme; ATypI and type conference circuit.	3

Community

The people who own the channels. Distribution to the supply side runs through them.

Name	Role	Why they matter	How to reach	Pri
Felix Lee	Co-founder & CEO ADPList	ADPList is a global network of senior designers with publicly bookable mentor profiles — the lowest-friction, highest-trust recruiting surface in existence for exactly this supply, and it is not a portfolio site so quality signal is stronger.	He maintains a public ADPList mentor profile that anyone can book — a literal open door.	1
Sacha Greif	Founder Sidebar	Sidebar has been the daily habit of working designers for over a decade and he also runs the State of CSS/JS surveys — meaning he owns both a distribution channel and the methodology for surveying developer and designer populations at scale.	Sidebar's sponsorship page is a public, priced channel; he is responsive on X and in indie-maker communities.	2
Andy McCune	Founder & CEO Cosmos	Cosmos is where taste-forward creatives actually curate references, and it has explicit anti-AI content controls. It is simultaneously the best-targeted recruiting pool and a demonstration of how this audience feels about AI.	Public founder with an active press and investor profile; company channels.	3
Héctor Ayuso	Founder and director OFFF Festival	Has programmed OFFF for over twenty years and effectively curates who counts in European digital and motion design. Conference programmers are the cheapest way to reach many credentialed creatives with one relationship.	OFFF speaker and partnership channels; he gives frequent public interviews and podcasts.	3
Jenny Brewer	Editor, It's Nice That It's Nice That	Commissions the coverage that shapes how the graphic design profession understands new companies. Whether the client is written up as an opportunity for designers or as an extraction scheme is substantially her call.	It's Nice That's public editorial contact and pitching channels.	3
Katy Cowan	Founder and editor Creative Boom	Creative Boom is the UK's most-read publication for freelance creatives and studios — precisely the population the client needs to recruit — and she hosts its podcast, so she is both a gatekeeper and an interview channel.	Creative Boom pitches and the Creative Boom Podcast are open public channels.	3
Zack Onisko	CEO Dribbble	Dribbble still has 16M+ registered designers and sells sponsorship, newsletters and design contests through an account manager. It is the largest paid recruiting channel available, even though its quality signal has decayed.	Dribbble's advertising page routes to an account manager; he is publicly active on community growth topics.	3

Investors

The partners who funded the competitors and wrote the category thesis.

Name	Role	Why they matter	How to reach	Pri
Sarah Catanzaro	General Partner Amplify Partners	She co-led Taste Labs' seed and wrote the public thesis for the category, and she is also an investor in Runway. She is the single best-informed outside observer of this exact market and holds a genuine cross-cluster introduction path from investor to lab.	Publishes under her own byline at Amplify and speaks frequently at data and ML events; publicly reachable.	1
Sarah Guo	Founder and Managing Partner Conviction	Backed Design Arena and is one of the most public investors on evaluation and data as a category, with a podcast that reaches exactly the buyer audience.	Co-hosts the No Priors podcast and posts constantly; a genuinely open public channel.	2
Sarah Leary	Partner, Unusual Ventures; co-founder of Nextdoor; director, Contra Unusual Ventures	She led Unusual's seed relationship with Contra and sits as a director — the single clearest read on how the incumbent competitor's board thinks about the Labs pivot, and a warm path to Contra if the client wants one.	Unusual Ventures publishes her founder stories; active public speaker on marketplaces and community.	2
Chris Dixon	General Partner Andreessen Horowitz	Led the $33M seed into Yupp, which collected millions of preferences a month and still died for lack of buyers in April 2026. He paid for the most expensive lesson available about demand for preference data.	Writes publicly and at length; a16z podcast and public channels.	3
Ted Wang	Partner, Cowboy Ventures; director, Contra Cowboy Ventures	A Contra director and ex-Fenwick partner; unusual depth on both the governance and the legal structuring of creative-data businesses, which is the client's largest unmodelled risk.	Public investor writing; Cowboy Ventures channels.	3

How this was built, and its limits

Method

Public professional information only

Names come from company sites, ATS boards, paper author lists, award jury pages, mastheads and podcast credits. No personal contact details and nothing from a private profile. Where a role could not be confirmed, the record says so rather than asserting it.

Known holes

Google DeepMind, and the design community

DeepMind publishes the richest human-evaluation methodology in the industry and no individual could be attached to it. Taste Labs lists eleven first names and one was resolvable. Sixteen of seventy people have no verified edge at all, almost all of them practitioners and critics — which is the gap the business has to cross.