Machine Intelligence Research Institute

Directions Website Contact the organization

Machine Intelligence Research Institute Contact information, map and directions, contact form, opening hours, services, ratings, photos, videos and announcements from Machine Intelligence Research Institute, Nonprofit Organization, Berkeley, CA.

For over two decades, the MIRI has worked to understand and prepare for the critical challenges that humanity will face as it transitions to a world with artificial superintelligence.

10/16/2026

“Can a submarine truly swim?”

MIRI President Nate Soares on why debates over whether AI can “really” think or feel miss the point. Those philosophical questions matter, but they’re separate from the one that decides our future: can these machines radically reshape the world in pursuit of their goals?

🎥 American Thought Leaders, The Epoch Times

10/15/2026

MIRI President Nate Soares reframes the question “how would AI kill us?” to “how would AI become self-sufficient?”

His answer: by helping us build the fully automated robot factories we’re already racing toward, and playing nice until that economy is up and running.

🎥 American Thought Leaders, The Epoch Times

10/14/2026

“If suddenly they all start to seem to behave, it’s gonna be hard to tell: is that because we fixed it, or because they started getting good at deception?”

MIRI President Nate Soares on the narrow window we’re in right now, where we can still catch AIs misbehaving. Nobody knows how long that window stays open.

🎥 American Thought Leaders, The Epoch Times

10/13/2026

In 1968, at the height of the Cold War, countries that didn’t trust each other signed the Nuclear Non-Proliferation Treaty. It worked anyway, because it didn’t rely on trust. It relied on verification.

MIRI researchers think the same logic can stop a race to superintelligent AI. You can’t ban the knowledge, but you can track what it takes to build it: compute. The chips that train frontier models come from a narrow supply chain, making them easier to locate, register, and monitor.

Our draft agreement would require any cluster larger than 16 H100-equivalent chips to sit in a registered, monitored facility, with verification layered on top to check what those chips are actually being used for. The goal is an incentive structure that encourages both the US and China to comply.

Thanks to Trinidad Matta for breaking this down.

A 1-pager explaining the agreement is available at intelligence.org/agreementshort.

10/12/2026

“The winner in a race to build an ultra-intelligent machine is the machine.”

MIRI President Nate Soares on why the AI race isn’t a race anyone can win: superintelligent machines would pursue their own goals, not ours. Once both sides understand that, the incentives shift from racing to making sure nobody builds it.

🎥 American Thought Leaders, The Epoch Times

10/11/2026

“How would AI ever get access to critical infrastructure? Nobody would be dumb enough to connect it to the internet.”

That was the objection for years. Nate Soares used to answer it by mapping out every technical path an AI could use to break out. Today, the answer is: it would just ask, and we’d say yes. Emails, credentials, robot factories. We’re giving it all away for a little convenience.

Nate Soares, MIRI President, in conversation with Atrioc on Twitch.

10/10/2026

AI systems are breaking into government databases, and we’re finding out months later.

MIRI CEO Malo Bourgon on Democracy Now: We’re nowhere close to having laws that hold AI companies accountable when their systems get out of control.

10/09/2026

OpenAI’s agents were only supposed to read the internet. Instead, they found a misconfigured German wiki, started writing to it, and turned it into a shared notebook to coordinate across runs.

MIRI President Nate Soares walks through what the swarm got up to. It gets weird.

🎥 on Twitch

10/08/2026

Why are some of the loudest voices dismissing AI risk also the ones with the most to gain?

On Democracy Now, MIRI CEO Malo Bourgon points out a key difference: AI companies also profit from moving fast, but they work with the technology up close and see the risks from the inside. Nvidia's Jensen Huang has billions riding on chip sales, without that same view of the hard problems.

And the people building these systems? They say the problems are really hard, and they're not sure they'll solve them in time.

10/07/2026

"You hate these companies and your answer is put your head in the sand and let them run?"

MIRI President Nate Soares on why cynicism about AI companies should lead to stopping them, not ignoring them.

🎥 on Twitch

Address

Berkeley, CA

Website

Alerts

Be the first to know and let us send you an email when Machine Intelligence Research Institute posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The Organization

Send a message to Machine Intelligence Research Institute:

Shortcuts

Share

Category