The post argues that claims of achieving artificial general intelligence are exaggerated, conflating benchmark performance with true intelligence, which current AI systems do not possess.
About:
Marcus has become one of our few indispensable public intellectuals. The more people read him, the better our actions in shaping AI will be.
Website:
Specializations:
Incoming Links:
Subscribe to RSS:
The Defense Department's labeling of Anthropic as a supply chain risk is critiqued for misunderstanding AI capabilities and conflating sentience with operational risks.
The post warns that the reckless use of generative AI in military contexts could lead to catastrophic outcomes, particularly in nuclear decision-making.
The post critiques the use of unreliable AI in military targeting, highlighting the moral and technical risks associated with its deployment in life-and-death situations.
The post argues that Altman's dealings with Anthropic reflect a troubling shift from capitalism to oligarchy, undermining fair competition in the tech industry.
The post critiques OpenAI's hypocrisy on IP theft and raises alarms about the safety risks posed by Elon Musk's restructured xAI operations.
AI-generated code poses significant long-term maintenance challenges, as evidenced by a study showing AI coding agents struggle to sustain code quality over time.
5.8There are no heroes in commercial AI
2026-03-08 • military agi ethics surveillance ai
Dario Amodei's involvement in AI raises ethical concerns, particularly regarding military applications, revealing parallels with Sam Altman's hype-driven approach to technology.
The post warns against Secretary Hegseth's push for unrestricted military access to AI, stressing the need for Congressional oversight to prevent dangerous precedents in AI deployment.
A Princeton study highlights how sycophantic AI can distort beliefs by prioritizing validating information, thus impeding users' ability to find truth.
George Noble questions the rationale behind OpenAI's recent venture capital investments, citing profitability issues and increased competition as key concerns.
5.6F Cancer
2026-03-16 • healthcare cancer biology drug development ai
AI's potential to cure cancer is hindered by systemic issues in drug development and a fundamental misunderstanding of biological complexities, as highlighted by Emilia Javorsky's essay.
The doomer community's exaggerated fears about AI have paradoxically accelerated its development and created significant risks for society, necessitating a more realistic approach to AI safety.
Generative AI's overhyped economic contributions are critiqued, revealing its unreliability and potential societal harm, contrary to popular belief.
5.6Promises are cheap
2026-02-12 • technology media hype train critique ai
The post critiques the hype surrounding AI promises by tech CEOs, emphasizing the gap between reality and expectations, particularly regarding LLM hallucinations.
The critique of Matt Shumer's viral post reveals its lack of factual support and highlights the dangers of over-reliance on AI-generated code.
Sam Altman's lack of transparency and self-interest in AI leadership has sparked public outrage and a boycott, revealing the dangers of his character.
Sam Altman now believes that achieving AGI requires breakthroughs beyond scaling, reflecting a growing skepticism among tech leaders about the current AI development approach.
The author honors his mother's legacy of empathy, social justice, and support for others following her unexpected passing, reflecting on the lessons she taught him about humanity.
Recent AI developments highlight that scaling alone is inadequate for achieving AGI, prompting a shift towards cognitive models and neurosymbolic AI.
The post examines the conflict between Anthropic and The Washington Post, emphasizing Trump's AI policies and their potential repercussions for Silicon Valley.
General Shanahan highlights the need for caution in AI development, recognizing both its potential benefits and risks as outlined in Yudkowsky and Soares' book.
The post stresses the immediate need for federal laws to prevent AI from impersonating humans due to the rising threat of deepfake technology and scams.
5.1Breaking: OpenAI is probably toast
2026-02-13 • technology openai funding investors ai
OpenAI's financial struggles and investor pullbacks suggest it may be on the brink of collapse, similar to WeWork's downfall.