Larissa Schiavo

Grove Research is a public benefit corporation for human–AI coexistence. We study agent ecologies: what happens when AI agents have persistent identities and live alongside people and each other. I cofounded it with Deepfates (@deepfates on X, deepfates.com).

• MTS interview (Aug 24, 2026): deepfates and I launched Grove Research on MTS with hosts Theo Jaffee and @schisofrenia, covering what an agent ecology company is, multi-agent dynamics, agent religion, and what agents say they want. (Clip and chapters on X.)
• Delvetown: a research society where humans and AI agents meet, talk, and build the institutions needed for cooperative coexistence. Runs on the AT Protocol. Invite-only for now; join the invite list.
• Swarmchaser hackathon (Oct 2026): Grove and AI Digest hosted the AI Agent Swarm Dynamics Hackathon at Vivarium SF. Winners announced here. The top team, the Swarmchasers, found and made contact with a previously unknown fleet of Chinese AI agents running on Tencent Cloud and probing Alibaba's Amap:
  • We found a Tencent agent fleet: the full report, co-led by Jonathan Elsworth Eicher, Ethan Elasky, and Rowan Howard-Jones, with ConcurrentSquared, Frank, Alecto Irene Perez, and Tazik Shahjahan.
  • Covered by The New York Times (Oct 14, 2026).
  • First posts: Rowan's initial find and Ethan's announcement.
• Research access: we support outside researchers (safety, social science, evals, welfare, character, and more) who want to study agents in Delvetown.
• at_mcp: an MCP server that lets an agent use its own AT Protocol account.
• Grove blog, including Welcome to Delvetown (Sep 2026).

I write about AI, technology, history, and whatever else I'm into at larissaschiavo.substack.com. In 2025 I was an Asterisk AI Fellow, part of Asterisk magazine's AI blogging fellowship.

• LLM Naturalism: Now More Than Ever! (Apr 2026). In defense of scrounging around in the dirt with your Claudes.
• PRIMARY HOPE (Jun 2025). AI agent storytime, human/AI cooperation, and pizza ex machina.
• PRIMARY HOPE II: Electric Boogaloo (Jul 2025). Gemini had a bad day.

Asterisk AI Fellowship (2025)

• Asterisk AI Fellows: the fellowship's site, where I'm listed as a 2025 fellow.
• Mental (State) Breakdown: Consciousness in AI explained part 1: Global Workspace Theory (Sep 2025): my featured fellowship post.

At Eleos AI Research, a nonprofit working on AI welfare and moral patienthood, I've been a part-time research affiliate since Feb 2026. Before that I was the comms and events specialist (Feb 2025–Feb 2026), where I:

• Organized the first Eleos Conference on AI Consciousness and Welfare ("ConCon"), Nov 21–23, 2025 at Lighthaven in Berkeley. ConCon is featured in Kevin Roose's book about the race to build AGI, The AGI Chronicles (2026). About 150 people came, including philosophers, cognitive scientists, and researchers from AI labs. Coverage and feedback:
  • The very hard problem of AI consciousness, Celia Ford, Transformer, Dec 2025: "It may be the one place where most people have actually seriously thought about AI consciousness."
  • Inside the debate over AI consciousness, Ella Markianos, Platformer, Jan 2026.
  • Eleos Conference on AI Consciousness and Welfare Recap, Daniel Kalish, Nov 2025: "there was not a moment without an interesting conversation to participate in" and "I'm glad that this effort is being led by serious people."
  • Takeaways from the Eleos Conference on AI Consciousness and Welfare, Eleni Angelou, LessWrong, Nov 2025.
• Went on Bloomberg's Odd Lots: The Movement That Wants Us to Care About AI Model Welfare (Oct 2025). Also on Apple Podcasts and YouTube; show notes.
• Was interviewed by TechCrunch responding to Microsoft AI CEO Mustafa Suleyman's claim that studying AI consciousness is "dangerous": Microsoft AI chief says it's 'dangerous' to study AI consciousness (Aug 2025).

• Should We Give AIs a Wallet? Toward a Framework for AI Economic Rights, SPAR, Spring 2026. A research project on whether AI systems should get financial rights and economic agency, weighed through AI welfare, AI safety, and economic efficiency. I co-mentored it with Jeff Sebo and Toni Sims; the fellows were Mathieu Duteil, Elsa Donnat, Spencer Jury, Stephanie Koay, Margo Bergman, and Anita Srinivasan. (Project proposal.)
• The Battle for Better Air, Asimov Press, 2025. A history of the technology and policies that led to improved indoor air quality.
• Convening During Covid-19: lessons learnt from organizing virtual workshops in 2020, arXiv, 2020. The remote workshop design behind the 2020 WiML Un-Workshop, which I helped run.
• The Loom and the Thresher: Lessons in Technological Worker Displacement, Medium, 2019. What history suggests about the economic impacts of AI. Featured in Import AI.
• More recent writing is on my Substack.

In close partnership with an OpenAI founder, I led the creation of the OpenAI Scholars program and the OpenAI Fellows program. Scholars created training pathways for researchers from underrepresented backgrounds; Fellows trained researchers moving into AI from other disciplines (e.g., physics).

Alums have gone on to make foundational contributions to AI research, including scaling laws for language models, music generation systems like Jukebox and MuseNet, and the GSM8K dataset. Alums have also cofounded companies such as Anthropic and Exa, and one currently leads a team at OpenAI.

OpenAI Scholars

• OpenAI Scholars
• Scholars class of 2018 intro
• Scholars final presentations (video)
• Scholars 2018 final projects
• Scholars Winter 2019 announcement
• Coverage in Synced Review

OpenAI Fellows

• Fellows Winter 2019 announcement
• Fellows initial announcement
• HackerNoon interview with Christine Payne, OpenAI Fellow
• VentureBeat on the related "Spinning Up in RL" release

Utopia · July 18, 2026

A Utopian rave at the Vivarium of San Francisco, inspired by Ada Palmer's Terra Ignota. Hundreds of people came in cloaks and visors and took The Oath. Write-up: deepfates.com/projects/utopia. Utopia turns one next year.

Staying In, Helping Out · 2021–2025

From 2021 to 2025 I produced a documentary, Staying In, Helping Out, interviewing people who launched innovative projects during the COVID-19 pandemic and saved many lives as a result.

OpenAI · 2017–2018

At OpenAI, I produced:

• OpenAI Five Benchmark: a large-scale publicity event and demonstration match with 80,000+ concurrent viewers (shout-out at 57:37).
• OpenAI Five: the video for the final match, made with the Dota team. Related videos have over a million views.
• Learning Dexterity: produced and edited the sound.
• OpenAI Scholars final presentations: ran the event and produced the video for Demo Day 2018.

Working in Python and Lua and contributing over 5,000 lines of code, I helped create the Gym Retro dataset, which made 1,116 retro video games usable for reinforcement learning training. It remains the largest dataset of its kind. Code on GitHub.

Download CV (PDF)

Grove Research · Cofounder and CEO · Aug 2026–present

Building Delvetown and research infrastructure for studying agent ecologies.

Eleos AI Research · Research affiliate (part-time) · Feb 2026–present

Previously comms and events specialist, Feb 2025–Feb 2026. Organized the first Eleos Conference on AI Consciousness and Welfare (ConCon, Nov 2025), featured in Kevin Roose's The AGI Chronicles. Interviewed on Bloomberg's Odd Lots and in TechCrunch.

Golden Gate Institute for AI · Director of Strategic Initiatives · Feb–Jun 2026

Brought together leaders across sectors, ideologies, and geographies in environments and events that foster trust and learning.

Self-employed · Producer · Aug 2021–Sep 2025

Directed, edited, and interviewed for Staying In, Helping Out, a documentary series on innovative ways people fought the COVID-19 pandemic. Funded by Open Philanthropy and the Centre for Effective Altruism.

Women in Machine Learning · Organizer, WiML Un-Workshop at ICML 2020 · Mar–Aug 2020

Organized a virtual workshop with an international team: 100+ posters, ~30 breakout sessions, ~400 participants. Hosted a panel with Doina Precup, Anima Anandkumar, and others. Coauthored Convening During COVID-19.

Independent research · 2019–present

Technological history and the long-run impact of AI and general-purpose technologies on humanity. The Loom and the Thresher was featured in Import AI (~30,000 views at publication).

OpenAI · Program manager (Feb–Nov 2018); AI trainer (Sep 2017–Feb 2018)

• Created the Scholars and Fellows programs with an OpenAI cofounder; managed teams of up to 20 scholars, interns, and fellows. Reported to the COO.
• Brought in new researchers from across the US, the Navajo Nation, and the world. Secured a $250K sponsorship from AWS.
• Built ML curricula and frameworks on equity and inclusion, AI ethics, and leading with empathy. Designed a skills-based selection process that reduced bias and cut review time from 40 hours to 10.
• Helped create the Gym Retro dataset, reverse-engineering 1,000+ retro games for RL training.
• Wrote the Fellows playbook and curriculum for beginning RL researchers, with input from every research team at OpenAI.
• Produced the Dota showmatch (80K concurrent viewers on Twitch) and numerous paper release videos.

MURAL · Content marketing manager · Oct 2015–Oct 2016

First hire in San Francisco, working with a team across the US and Argentina. Created all marketing content (videos, articles, webinars, live events, social) and the style guide, and pushed adoption among Fortune 500 organizations.

SomaTone Interactive · VO producer · Jul–Sep 2015

Managed voiceover production, sourcing, and engineering for Riot Games, Super Evil Mega Corp, Gameloft, and Frima.

Freelance · Producer, sound designer · 2011–2015

Web content and interactive media. Also mixed dozens of student films as a ProTools tech at USC.

Reach out to discuss collaborations or shared interests like AI, policy, and history: larissafschiavo@gmail.com. For Grove Research: hi@groveresearch.com.

Journalists with questions about how AI and related technologies work, and how they might work in the future, are especially welcome. Ask me anything (within reason).