🪩 Party Guide is 🔥 View Luma

Do the Boring Stuff to Make Open Source AI Win

  • Inference Systems

"Open source AI just isn’t as good as OpenAI, Anthropic, and Gemini.”  

Needing to reach performance parity with closed source frontier models is the prevailing north star for open source AI - and this is an important goal!  Here, I’ll make the case for a different north star: ease of use.  Open source AI can win if it becomes dead easy to use, the simple default option for the 99% of AI use cases that don’t require the newest frontier mega-model.  The missing piece is being so easy to install, maintain, and use that customers aren’t tempted to default elsewhere.  That means our community moving beyond benchmarks to beautifying UI/UX, reducing jargon, building catchy apps, and meeting the modal end-user where they are - *not* just on GitHub or Hugging Face, but rather in the address bar of a browser or the app store.

Throughout, I’ll talk about some of the decisions we are making at Mozilla to support our “choice-first stack,” a portfolio of commercially-licensed open source AI projects meant to take steps toward making open source AI more approachable, maintainable, and usable.

John Dickerson

CEO

John Dickerson

Mozilla.ai

John Dickerson is CEO of Mozilla.ai. He brings a wealth of experience in scaling startups, developing practical and robust machine learning methods, deploying AI-based products into the enterprise, as well as providing broad AI/ML thought leadership in industry, academia, non-profits, and governments.

Previously, John was co-founder and Chief Scientist at Arthur as well as a tenured professor at the University of Maryland in the Washington, DC area.

At Arthur, he helped scale the company to 50+ employees, a presence in NYC, DC, and the US west coast, and $55 million raised from seed through Series B financing. Arthur develops industry-leading technology in data drift detection and mitigation, bias detection and mitigation, GenAI firewall features such as jailbreak and PII leakage detection, and explainability. Arthur’s ML-based products are deployed at some of the largest regulated enterprises in the US and worldwide.

At Maryland, he founded and led a large lab researching the intersection of ML and economics, with a core focus of designing incentives that promote “good” participation in complex systems. That lab produced 16 PhD graduates and secured $10M+ in funding from NIST, NSA, DARPA, ARPA-E, NIH, NSF – including an NSF CAREER award – in addition to industry funding.

He has worked extensively on theoretical and empirical approaches to organ exchange where his work has set US-wide policy; worldwide blood donation markets with Meta; game-theoretic approaches to counter-terrorism and negotiation, where his models have been deployed; and market design problems in industry (e.g., online advertising) through various startups.

John holds a BS in mathematics and a BS in computer science from the University of Maryland, as well as a PhD in computer science from Carnegie Mellon University. He splits his time between Seattle, Washington, USA and Western Europe.