Like Sabine Hossenfelder, I Was Offered Money to Tell You AI Will Kill Us
From House of El: AI · · 29 min

The player loads from YouTube only when you click. It uses YouTube's privacy-enhanced mode.
If the video does not play, the owner may have turned off embedding. Watch it on YouTube instead.
Description
Can AI really improve itself recursively until it becomes powerful enough to cause human extinction? Nobody sponsored this video, I turned down funding specifically for this topic, I do not work for any frontier AI lab, and I have no financial stake in whether the answer turns out to be yes or no. What I do have is a PhD in Computer Science and research experience, so instead of relying on headlines, CEO warnings or expert forecasts, I went through the research itself. Across twelve papers, I found evidence that genuinely surprised me, some of it far more concerning than I expected, but also major gaps between what has actually been demonstrated and what is being claimed publicly. So the question I wanted to answer was simple: where is the scientific evidence that gets us from recursive self-improvement to zero humans?
☕ Support the channel on Ko-fi: ➡️ https://ko-fi.com/houseofel
📌 Chapters 00:00 Introduction 03:54 It's More Real Than I expected 06:23 Conclusion 1 08:04 The Hidden Dependency 14:33 Conclusion 2 15:01 Could the bottleneck vanish? 19:24 Even if the bottleneck gets solved, does the loop run away? 23:26 Fine. Grant everything. Show me extinction. 27:50 Don't Look Up
📚 SOURCES & FURTHER READING
Chen, Wang & Qu (2026). “Recursive Self-Improvement in AI.” https://arxiv.org/abs/2607.07663
Zhang, Hu et al. (2025). “Darwin Gödel Machine.” https://arxiv.org/abs/2505.22954
Novikov, Vu et al. (2025). “AlphaEvolve.” https://arxiv.org/abs/2506.13131
Wijk, Lin et al. (2024). “RE-Bench.” https://arxiv.org/abs/2411.15114
Wen, Qiu et al. (2026). “Anthropic Automated Weak-to-Strong Researcher.” https://alignment.anthropic.com/2026/automated-w2s-researcher/
Kirgis, Kapoor, Narayanan et al. (2026). “Can AI Agents Conduct Open-Ended AI Research?” https://arxiv.org/abs/2607.27191
Cunningham et al. (2026). “The Economics of Recursive Self-Improvement.” https://arxiv.org/abs/2609.15802
Burtsev (2026). “Recursive Criticality of AI Self-Improvement.” https://arxiv.org/abs/2609.00137
Google DeepMind (2026). “From AGI to ASI.” https://deepmind.google/research/publications/239142/
International AI Safety Report (2026). “International AI Safety Report 2026.” https://internationalaisafetyreport.org/publication/international-ai-safety-report-2026
OpenAI (2026). “Research Acceleration: A View Inside OpenAI.” https://openai.com/index/research-acceleration-view-inside-openai/
Grace et al. (2024). “Thousands of AI Authors on the Future of AI.” https://arxiv.org/abs/2401.02843
#AI #ArtificialIntelligence #AISafety #RecursiveSelfImprovement #AGI #Superintelligence #ComputerScience #AIResearch #AIExtinction