Coalition for a Baruch Plan for AI

A coalition of NGOs to foster a uniquely timely, bold and effective treaty-making process to build the global federal organization that we need to master AI.

[Geoffrey Hinton @geoffreyhinton]

The idea of an intelligence explosion caused by recursive self improvement has been around for a long time but until very recently it did not seem imminent. Now many leading researchers think it may happen quite soon. You can read our paper about it here:
https://t.co/sgUpugjpRY

[Rufo Guerreschi @rufoguerreschi]

🚨 Last week 29 nations + the EU signed a call for an international body to control frontier AI. Xi told Trump AI "must be kept under human control." Washington said no to both.

What if Beijing signs the call, brings in 37 countries of its open WAICO AI alliance, and decisively moves onward for a bold AI safety treaty?

Meanwhile, leading US labs AI architectures have reached a "controllability cap", and a consequent "capability cap", which massively reduces US negotiation power via advanced training chips and access to the most advanced models.

How would such a scenario affect the US economy? Should Trump co-lead instead soon, before the G20 in Doral?

https://t.co/3um4dIIx47

#AIgovernance #baruchplanforai #Superintelligence #ASI

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

Proud to have been interviewed by the unique John Sherman. Listen to his great forsight in early 2023 when most of us in the movement got our wake up call and dived deep down in existentila crisis.

[John Sherman @ForHumanityPod]

I wrote this nearly three years ago and recorded this video more than six months ago.

In recent months I have wondered if it was too harsh.

Given the events of the last several weeks, from Hugging Face through Jacob Coxon's incredibly brave breakthrough moment, it's not too harsh at all. It's what needs to be said.

Thank you, @hilbertspaess. If we wind up saving the world, your role and September 9th, 2026, will never be forgotten.

[David Krueger 🦥 ⏸️ ⏹️ ⏪ @DavidSKrueger]

I'm genuinely curious about this!

[sarah @littIeramblings]

I appreciate the candour of this tweet but I think the “>10%” phrasing still undersells things

as of 2022, Evan assigns an ~80% chance to existential risks from AI. maybe most of this 80% is accounted for by existentially terrible things other than AI killing everyone, but the distinction between ‘existential’ and ‘extinction’ isn’t one that most people outside of this discourse have thought about or understand.

and even though Evan says the risk is ‘greater than’ 10%’ I think this lower bound is still the thing that ends up getting the memetic power. a cursory or not-very-careful read gives the impression of a 90% chance everything will be fine (still terrifying odds, but sadly soft selling the true level of expert alarm). I think expressing probabilities as a range - 10-20%, 10-50%, whatever - is a better strategy

we seem to be in a very rare moment of public attention on this issue right now (partly thanks to this tweet!) so I think now would be time for experts and researchers to throw out their most unfiltered takes.

[AI Notkilleveryoneism Memes ⏸️ @AISafetyMemes]

ANTHROPIC ALIGNMENT LEAD: 80% (!) chance we are all about to die.

When he said ">10%" he was softpedaling it

Also, when Coxon said we might die "this decade" he meant in the next ~3 YEARS

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

Come meet us in Washington DC this Sept 14-19.

The first Trump-era US-China talks devoted solely to AI are being lined up for the coming weeks, ahead of President Xi's September 24 visit to Washington. Their agenda — cyberattacks, distillation, labs sharing threat data — is a fraction of what Xi, the leading labs and 63% of US voters have said the moment requires.

Closing that gap is the niche we occupy: designing the treaty-making process the moment demands, and converging the dozen "humans first" figures around the White House who can put it on the President's desk.

We'll be in Washington September 14–19 for the Pro-Human Assembly, meeting prospective co-hosts, co-steering partners and funders of our Q4 Deal of the Century Roundtables.

Come find us: [email protected]

📄 September newsletter: https://t.co/JgRWdfOXmb

🏛️ Our Upcoming Roundtables: https://t.co/pqaAnCs6HO

#AIgovernance #baruchplanforai #Superintelligence #ASI

[Rufo Guerreschi @rufoguerreschi]

This week I was interviewed by the great John Sherman

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

@danfaggella it will lead the its internationalization or loss of control. Natiolization with just be a small parenthesis

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

@danfaggella please call us humans, would you?!

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

🚨 New: the 53-page Executive Summary of our Strategic Memo v2.8 is out.

The thesis is simple: it's either both, or neither.

Preventing AI loss of control and extreme power concentration takes two things together:

1️⃣ a bold, timely US–China-led AI safety treaty
2️⃣ treaty-backed citizen co-ownership of AI firms

Read it 👉 https://t.co/whh8RVKXFE

#AIgovernance #baruchplanforai #Superintelligence #ASI

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

🎙️ May 25: Pope Leo XIV presents his AI encyclica alongside @AnthropicAI's cofounder @ch402 on the opening panel.

Anthropic's own leaders estimate 25% chance AI destroys humanity. The Vatican has said almost nothing on loss-of-control.

Olah's speech could close that gap and catalyze a critical mass of key potential influencers of Trump's AI policy to champion a US-China-led AI treaty.

🧵👇
https://t.co/Ag7SDEoIFW
#AIgovernance #baruchplanforai #Superintelligence #ASI

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

🔴 The Pentagon-Anthropic clash proves it: the "human power grab" is happening without a treaty, not because of one.

Militarization. Nationalization. RSP rollbacks.

The solution? A conditional assurance contract among top AI labs CEOs leading to a US-China-led AI treaty.
Built on distrust, not trust.

🧵👇

https://t.co/A0X3QD4YO7

@AnthonyNAguirre

#AIgovernance #baruchplanforai #Superintelligence #ASI

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

🌍 Assume the hardest part happens: the US and China agree on an emergency AI deal. What comes next?

Every conventional treaty-making model fails:

❌ UN multilateralism — took 23 years to get the (weak) Paris climate accord. We have maybe 5 years before ASI.

❌ Club models (G7, "NATO for AI") — exclude China and lack global legitimacy.

❌ Expanding the bilateral deal — every added nation gains a de facto veto. Gets slower and weaker.

❌ Voluntary frameworks — evaporate under competitive pressure. Always.

There is exactly ONE model that meets all requirements: speed, legitimacy, realism, enforceability, and protection against value lock-in.

A Realist Global Constitutional Convention for AI.

What makes it "realist"? The two superpowers + a majority of middle powers pre-negotiate the vote weighting and rules before anyone sits down.

→ Voting weighted by GDP, tech leadership, and population

→ US + China = ~30% of votes — lead, never dictate

→ Hard deadlines: 12-18 months max

→ No vetoes. Supermajority rules.

This isn't utopia. It's the model that produced the US Constitution in 4 months. The Swiss Confederation. The EU. The African Union.

Every time it worked, the most powerful participants helped design the rules first.

New blog post makes the full case

👇 https://t.co/4gKupeLMYk

#AIgovernance #baruchplanforai #Superintelligence #ASI

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

🚨 "This is a question between me and Demis—which I am very confident we can work out."

At Davos 2026, Anthropic CEO Dario Amodei made a striking admission: he'd prefer to slow down the AI race. But he can't—because "we have geopolitical adversaries building the same technology at a similar pace."

Then came the revelation: if China weren't in the picture, coordinating a pause with Demis Hassabis would be straightforward.

@AnthonyNAguirre of the Future of Life Institute quickly followed with a revolutionary proposal for a mechanism that could let them do exactly that—an assurance contract where labs commit to slowdown terms that activate only when enough others sign. No one moves first, until (nearly) everyone moves together.

In our new analysis, we explore:

→ Why this configuration is revolutionary

→ Which AI leaders would likely join (Altman, Suleyman, Musk, Sutskever)

→ Why the contract alone isn't enough—but what it would trigger is

→ The enforcement mechanisms and diplomatic infrastructure a real treaty would require

→ What such agreement should demand from Trump and Xi

The mechanism exists. The philosophical alignment exists. What remains is the decision to act.

Read the full analysis ↓ https://t.co/lVoSbOZFMx

#AIgovernance #baruchplanforai #Superintelligence #ASI

[Rufo Guerreschi @rufoguerreschi]

If Trump succeeds in its The Deal of the Century for AI,

he can happily retire in NYC in 2029

tasting his unprecedented legacy, cherished the work over

https://t.co/j40NiauhZa

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

🚀 NEW BLOG POST: Why Trump sudden leadership via COVID’s Operation Warp Speed could repeat with a bold AI Treaty

In January 2020, Trump called COVID "totally under control." By May, he launched Operation Warp Speed — the fastest vaccine development in history.

Experts said it was impossible. He did it in under a year.
The same pattern can repeat with AI.

Trump isn't an ideologue — he's a pragmatic dealmaker who pivots fast once he grasps the full scale of a threat. And the numbers are staggering:
🔹 77% of Americans are concerned AI could threaten humanity
🔹 53% believe AI will "destroy humanity" someday
🔹 57% rate AI's societal risks as "high" — and rising monthly

What will be Trump's "hospitals overflowing" moment on AI?

We believe it's one or a mix of three forces:
📊 Spiking citizen concern (already underway)
⚠️ A major AI accident or crisis
🤝 A powerful joint pitch from key potential influencers of Trump's AI policy — such as Musk, Vance, Bannon, Pope Leo XIV, Altman, Amodei

Warp Speed worked because Trump gave it speed, autonomy, and unlimited resources. The same instincts applied to AI could produce the most consequential treaty since the Nuclear Non-Proliferation Treaty.

The demand for leadership exists. What's missing is the supply.

Read the blog post → https://t.co/no2ibS13ib

hashtag#AIgovernance hashtag#baruchplanforai hashtag#Superintelligence hashtag#ASI

[Jonathan Liedl @JLLiedl]

Douthat asked Vance about Leo XIV and AI. Fascinating response:

"The American government is not equipped to provide moral leadership, at least full-scale moral leadership, in the wake of all the changes that are going to come along with A.I. I think the Church is."

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

5/ Enter Pope Leo XIV.

He chose his name to signal AI as central to his papacy.

His top advisor Paolo Benanti has gathered leading AI scientists around a call for a bold AI treaty.

The Vatican already convened world religions and big tech on shared AI ethics.

[Coalition for a Baruch Plan for AI @BaruchPlanForAI]

4/ Enter JD Vance.

When asked about AI extinction risk, he cited the AI 2027 report and said the government is "not equipped to provide moral leadership" — "the Church is."

Peter Thiel called this a dangerous "Caesar-Papist fusion." That reaction alone tells you the stakes.