Chat, is this a good idea? Would you prefer @sama or GPT-6/7 as OpenAI CEO?

In his reflections on the Progress Conference, @KevinKohlerFM comes to a similar desire. I think there's a market here.

Guest post from Christina looking at how IAEA verification tools work at an individual level. In the AI discussion chip-level verification tools are most prominent but worth considering multiple info streams incl. OSINT

It's only abundance if it's from the Abondance region in France. Otherwise it's just a sparkly supply surplus.

In summary ✅ A welcome, forward-looking first step from the UN 🔄 But future reports should go beyond perception and into deeper analysis Commentary by me & @MaximeStauffer 👉 https://t.co/zKG5snvPDq Risk Report:

Ideas for future reports · Time its release with key UN events · Combine perception surveys with quantitative risk assessments by experts · Compare national assessments of global risks · Analyze not just risks, but global critical systems they could disrupt (eg SWIFT,Suez,ICANN)

Risk perception has limits Two common biases: · Centering bias: Responses cluster around the middle of the scale. · Availability bias: Recent events are rated as higher risk. The survey data was collected in 2024. Now smth like “economic fragmentation” might rank higher.

What the report found 1,000+ UN stakeholders rated the likelihood & impact of 28 global risks (scale 1–7). · Top perceived risk: “Climate Change Inaction” · A key gap: “Mis- and disinformation” ranked #3 in importance, but preparedness was seen as low.

The UN has published its first Global Risk Report 🇺🇳 A quick thread 🧵on: · What the report found · Why risk perception ≠ risk assessment · What future editions could add

What lessons can the new UN AI Panel learn from the International AI Safety Report? 📚 Join us on July 9th in Geneva with Prof. @Yoshua_Bengio (remote), Daniel Privitera & Hannah Merchant for insights on building effective scientific panels. Registration link below 👇

🌐 Does AI need it's own "CERN"? 🌐 Join us & @TheGCSP this Thursday, July 10th, on the sidelines of ITU's @AIforGood Summit, for a Geneva Security Debate on the subject. Registration link below 👇

I highly recommend applying to the @rootsofprogress fellowship by June 1!✍️🌎🚀It's an inspiring program that builds the progress community & offers opportunities to engage with lots of cool people.

Link to full report with Recommendations for the Independent International Scientific Panel on AI and the Global Dialogue on AI Governance 🇺🇳🤖 https://t.co/4vmEqBEeAO A big thanks to all the experts that took the time for conversations, interviews & workshops 🙏

The UN is creating a scientific panel to assess the global risks, opportunities and impacts of AI. How can we design this in a way that is fast, scientific, and relevant to policymakers? Happy to share our new report on this 👇 In September 2024 the UN General Assembly decided

Excited to share our new interim report, outlining potential design options for the #UN's Independent International Scientific Panel on AI and Global Dialogue on AI Governance, both mandated by the #GlobalDigitalCompact in Sept 2024 🧵 https://t.co/fheSJfQ0La

And here's the link to the full text: https://t.co/gyjscxm7eZ

The Progress Conference was fantastic – four short reflections 👇 1. A shared belief in progress The progress movement has a broad tent. There is no party line. Still, the following is what I have perceived as common beliefs across most participants: · technological progress

TIL that Russia invaded New York https://t.co/xo1jV5qlSr

@ESYudkowsky & @leopoldasch both believe powerful AI is close & has serious risks. However one wants to stop AI, the other calls for an arms race. The crux seems to be that one thinks control over such AI is near-impossible, whereas the other is more bullish on superalignment