June 24, 2026
Open-source or open season?
For Most of the World, Open-Source AI Is the Only Way Forward
AI for everyone or AI for the rich? Commenters say the fight just got very real
TLDR: At the UN, Yann LeCun argued that shared, open AI is the only realistic way for most countries to keep control of their digital future instead of depending on a few powerful companies. Commenters loved the anti-monopoly message but fought over one big catch: if ordinary people and poorer countries can’t afford the hardware, is it really freedom?
Yann LeCun walked into the United Nations and basically said the quiet part out loud: if artificial intelligence ends up controlled by a handful of giant companies in the US and China, the rest of the world could be stuck renting its digital future. His answer is open-source AI — systems whose core parts are shared publicly so countries, universities, and smaller companies can help build them instead of just buying access. Delegates from places like Morocco, Sierra Leone, Jamaica, and Spain backed the idea, with one warning that society’s tech future should not be decided by “a few techno bros.” Yes, that line absolutely lit up the room.
But the comments? That’s where the real fireworks were. One camp was fully in revolutionary mode, declaring AI should not become an “enclosure of the digital commons” because it was trained on humanity’s shared knowledge in the first place. Another camp slammed the brakes and asked the practical question: who can actually afford the machines needed to run this stuff? If a decent AI setup costs a fortune, is this freedom or just freedom for rich labs? And then came the spicy maximalists, scoffing at small home setups and demanding giant public models with one-click access for normal people, not “nerd tools” and “node graph garbage.”
The vibe was equal parts political manifesto, anti-monopoly panic, and open-source chest-thumping. Even the side comments had receipts, with one user dropping a UN video link like a fact-check grenade. The crowd seems to agree on one thing: if AI is going to shape daily life, people do not want a tiny club holding the keys.
Key Points
- •At United Nations Open Source Week in New York City, Yann LeCun argued that open-source AI is the only viable path to global AI sovereignty, cultural diversity, and long-term safety.
- •LeCun said most countries cannot afford to build frontier-scale large language models independently, but could collaborate on shared open platforms.
- •He warned that AI controlled by a small number of proprietary systems from major companies in the US and China could threaten cultural and linguistic diversity, democracy, and human rights.
- •LeCun described a federated model in which countries and institutions contribute to a global AI model by exchanging parameter vectors rather than transferring their underlying data.
- •The article presents Project Tapestry as an early implementation of this approach, with reported interest from multiple countries and companies including IBM, NVIDIA, AMD, and Intel, and a hoped-for production target of early 2027.