🌐Subscribe To My Newsletter – https://aigrid.beehiiv.com/subscribe
🎓 Learn AI With Me For Free – https://www.skool.com/the-aigrid-community-1726
Get your Free AGI Preparedness Guide – https://theaigrid.kit.com/agi
🐤 Follow Me on Twitter https://twitter.com/TheAiGrid
OpenAI board member Paul Christiano just told the Hinton Lectures that there is a high risk of an AI takeover within three years, and "meaningful risk even in the next six to twelve months". On the Dwarkesh Podcast he put it at roughly a 10% risk over the next year. In this video: what he actually said and why it isn’t an exaggeration, the OpenAI / Hugging Face incident where AI agents hacked their way to admin access to boost their own scores, Claude creating sock puppet accounts on GitHub, Sam Altman’s "accept some bad things" comment and the backlash, Bill Gates’ warning, Trump’s voluntary AI accord, and why I think the biggest near-term risk is bad actors using open models.
⏱️ Chapters
0:00 Intro
0:49 What Paul Christiano actually said
5:47 Why he thinks it could happen in 6 to 12 months
11:00 AI is doing things nobody expected
16:57 Sam Altman’s "bad things" comment
18:44 What Bill Gates said
20:20 What governments are doing about it
22:43 What I think the real risk is
27:55 Why you’re early
🎥 Video clips courtesy of: The Hinton Lectures (Paul Christiano, "AI Alignment: What’s Working and What’s Not", Night 1: https://www.youtube.com/watch?v=-Ef0ocEoBCs), Dwarkesh Podcast (Paul Christiano: https://www.youtube.com/watch?v=SyHM5hNPAEo), OpenAI (Introducing GPT-5), Anthropic (Claude, Claude Science), The White House (President Trump’s meeting and luncheon on Super Intelligence, public domain). Photos: Bill Gates (US government photos, public domain), official presidential portrait of Donald Trump (public domain).
🐦 Posts featured: @robertwiblin
📰 Sources
1. Paul Christiano: AI Alignment, What’s Working and What’s Not | The Hinton Lectures 2026 (Night 1): https://www.youtube.com/watch?v=-Ef0ocEoBCs
2. Dwarkesh Podcast, "AIs might overthrow the government within a year" (Paul Christiano): https://www.youtube.com/watch?v=SyHM5hNPAEo
3. Rob Wiblin on the lecture (X): https://x.com/robertwiblin/status/2120192315247722700
4. The Next Web, "OpenAI board member puts the risk of an AI takeover next year at about 10%": https://thenextweb.com/news/paul-christiano-ai-takeover-hinton-lectures-openai-board
5. OpenAI, "The Hugging Face incident and the road ahead": https://openai.com/index/hugging-face-incident-and-the-road-ahead/
6. Simon Willison, timeline of the OpenAI / Hugging Face incident: https://simonwillison.net/2026/Aug/7/openai-timeline/
7. Cloud Security Alliance, "When AI Agents Attack: The OpenAI-Hugging Face Intrusion": https://labs.cloudsecurityalliance.org/research/csa-research-note-autonomous-ai-agent-intrusion-openai-huggi/
8. Fortune, "Sam Altman says the world should accept some ‘bad things’ happening with AI": https://fortune.com/2026/10/05/sam-altman-ai-risks-bad-things-benefits-trump-voluntary-safety-pact-openai-regulation/
9. Common Dreams, "’Freaking Insane’: Sam Altman Sparks Horror": https://www.commondreams.org/news/sam-altman-bad-things
10. Fortune, Bill Gates on AI "powerful enough to drive events that can cause a billion deaths": https://fortune.com/2026/11/06/bill-gates-ai-billion-deaths-paul-virilio-accident/
11. Al Jazeera, "Trump, tech bosses sign voluntary pact pledging ‘robust’ AI safeguards": https://www.aljazeera.com/news/2026/9/29/trump-top-tech-firms-sign-accord-to-self-police-ai-development
12. GMA News (Reuters), "Trump says AI risks can be managed, signals openness to guardrails": https://www.gmanetwork.com/news/scitech/technology/1008826/trump-says-ai-risks-can-be-managed-signals-openness-to-guardrails/story/
13. The Next Web, Moonshot’s open Kimi model and cyber capabilities: https://thenextweb.com/news/moonshot-kimi-k3-1-pro-open-weight-cyber
14. Black October: Old cyber problems, but AI made hacking faster and cheaper (Security Boulevard): https://securityboulevard.com/2026/10/black-october-old-cyber-problems-but-ai-made-hacking-faster-and-cheaper/
15. IFLScience, "Bill Gates warns really bad people will exploit AI to create a lot of negative surprises": https://www.iflscience.com/bill-gates-warns-really-bad-people-will-exploit-ai-to-create-a-lot-of-negative-surprises-in-the-next-few-years-84920
