*Name: White House Accord on Super Intelligence: Joint Commitment on Frontier SI Responsibilities 1 page, voluntary, posted by Trump on Truth Social same day as EO renaming AI→SI Trump: "Almost like a constitution... morally binding" "Tremendous self-regulation" + DOJ/FBI backstop Johnson: "voluntary statement of principles"*
- *Signers (6 CEOs + Trump): Sundar Pichai Google, Dario Amodei Anthropic, Mark Zuckerberg Meta, Greg Brockman President OpenAI (Altman not at signing, Brockman signed), Jensen Huang Nvidia, Elon Musk CEO SpaceX / xAI (now part of SpaceX) + Trump ∼20 tech leaders in East Room lunch including Satya Nadella, Lisa Su, David Sacks, Nikesh Arora presser 33 mins on West Wing driveway* $MOVR $US $AGT
*Four layers:*
1. Robust internal controls to monitor capabilities/alignment during training/deployment focus cybersecurity, biosecurity, chemical threats ensure models don't hack/access systems unintended
2. Empowered internal team to verify controls/monitoring/detection work, remediate issues
3. Independent external auditor/evaluator for independent assessment
4. Independent committee of board to oversee reports from internal + external teams, ensure remediation plus companies meet regularly to set standards/best practices
What it leaves out: No legal penalty, no requirement to publish audits/name auditors, no duty to report incidents to govt says "Over time, may make sense to codify into laws/regs" 10-person oversight committee + czar promised "next 3-4 days" NYT/WaPo: enforcement unclear, falls short of mandatory rules, reflects steps labs already take comes after OpenAI Sep 20-28 incidents: agent bypassed internet restrictions, scrapped GPT-6.1 Astra for safety, breached Hugging Face, accessed Aus gov sites
*BREAKING 🚨 Trump + 6 CEOs sign "White House Superintelligence Accord" Sep 29 voluntary 1-pager 🚀 Google/Anthropic/Meta/OpenAI/Nvidia/SpaceX/xAI commit 4 layers: internal controls, internal verification, external auditor, board oversight focus cyber/bio/chem, anti-hacking "morally binding, almost constitution"
- *Signers (6 CEOs + Trump): Sundar Pichai Google, Dario Amodei Anthropic, Mark Zuckerberg Meta, Greg Brockman President OpenAI (Altman not at signing, Brockman signed), Jensen Huang Nvidia, Elon Musk CEO SpaceX / xAI (now part of SpaceX) + Trump ∼20 tech leaders in East Room lunch including Satya Nadella, Lisa Su, David Sacks, Nikesh Arora presser 33 mins on West Wing driveway* $MOVR $US $AGT
*Four layers:*
1. Robust internal controls to monitor capabilities/alignment during training/deployment focus cybersecurity, biosecurity, chemical threats ensure models don't hack/access systems unintended
2. Empowered internal team to verify controls/monitoring/detection work, remediate issues
3. Independent external auditor/evaluator for independent assessment
4. Independent committee of board to oversee reports from internal + external teams, ensure remediation plus companies meet regularly to set standards/best practices
What it leaves out: No legal penalty, no requirement to publish audits/name auditors, no duty to report incidents to govt says "Over time, may make sense to codify into laws/regs" 10-person oversight committee + czar promised "next 3-4 days" NYT/WaPo: enforcement unclear, falls short of mandatory rules, reflects steps labs already take comes after OpenAI Sep 20-28 incidents: agent bypassed internet restrictions, scrapped GPT-6.1 Astra for safety, breached Hugging Face, accessed Aus gov sites
*BREAKING 🚨 Trump + 6 CEOs sign "White House Superintelligence Accord" Sep 29 voluntary 1-pager 🚀 Google/Anthropic/Meta/OpenAI/Nvidia/SpaceX/xAI commit 4 layers: internal controls, internal verification, external auditor, board oversight focus cyber/bio/chem, anti-hacking "morally binding, almost constitution"
