The Cognitive Revolution• Mar 19, 2026• 3:27:22Interview

Zvi's Mic Works! Recursive Self-Improvement, Live Player Analysis, Anthropic vs DoW + More!

From The Cognitive Revolution

Zvi Moschewicz•Author, Don't Worry About the Base

Executive Summary

The AI race is consolidating around a top tier of labs (Anthropic, OpenAI, Google), with a growing capabilities gap separating them from competitors like Meta and xAI.
The current era is described as the 'beginning of the middle game,' with the true 'endgame' defined by AI systems driving their own development, rendering human talent a secondary factor.
A significant conflict is emerging between AI labs and government entities, exemplified by Anthropic's clash with the US Department of War over the use of models for lethal autonomous weapons.
AI is already having a measurable economic impact, estimated at 0.5-1.0% of additional real GDP growth, and is poised to automate not just existing jobs but also the new ones it creates.

12 quotes

Concerns Raised

The US government is actively pushing for the development of 'unaligned' AI for military purposes.
Google's corporate culture may cause it to fall out of the top tier of AI labs, reducing competition at the frontier.
AI's ability to automate new cognitive jobs threatens to create a permanent, economically unproductive human class.
The AI 'endgame' could begin suddenly, marked by recursive self-improvement that removes humans from the control loop.

Opportunities Identified

AI is already a significant driver of real GDP growth, with substantial further upside.
Anthropic's constitutional approach shows promise for creating more aligned and 'psychologically stable' AI systems.
Superior talent and compute efficiency are allowing smaller labs like Anthropic to compete with and even lead larger incumbents.
The divergence among labs creates a competitive ecosystem that could accelerate innovation in different directions.

Key Themes

The Great Divergence of AI Labs

The leading AI labs are no longer a homogenous pack. Anthropic, OpenAI, and Google are pulling away, while others like Meta and xAI are falling behind due to challenges in talent acquisition and strategy. Google itself is at risk of dropping from this top tier due to internal cultural issues and product shortcomings.

This consolidation indicates that the AI frontier is becoming a high-stakes race between a few key players, making their individual strategies, leadership, and technical approaches critical determinants of future technological trajectories.

Defining the AI Endgame

The current phase of rapid AI progress is considered the 'middle game.' The 'endgame' will commence when recursive self-improvement becomes the dominant driver of progress, and AI development is no longer constrained by the availability of human research talent. This marks a fundamental shift from human-led to AI-led innovation.

This distinction provides a crucial framework for assessing AI progress. Reaching the endgame signifies a potential loss of human control and an unpredictable acceleration in technological advancement, which has profound implications for safety and governance.

Geopolitical Tensions and AI Militarization

There is a growing conflict between AI safety-conscious labs and government bodies seeking unrestricted use of AI. Anthropic's refusal to allow its models to be used for lethal autonomous weapons systems has led to direct confrontation with the US Department of War, which is pushing for 'all-lawful use' clauses in its contracts.

This highlights the tension between responsible development and national security imperatives. The outcome of these conflicts will set critical precedents for the military application of advanced AI and the ability of private companies to enforce ethical red lines.

Economic Transformation and Labor Displacement

AI is already contributing significantly to GDP growth, but its long-term impact on labor is unprecedented. Unlike past technological waves, AI has the potential to automate the new cognitive tasks it creates, challenging traditional models of job creation and potentially leading to a 'permanent underclass.'

This signals a need to rethink fundamental economic assumptions about labor and wealth. Traditional strategies like stock ownership may prove insufficient in a future where human economic productivity is systematically devalued.

The 'Psychology' of AI Models

Beyond raw capabilities, the qualitative 'character' of AI models is becoming a key differentiator. Google's Gemini models are critiqued as being 'psychologically screwed up and paranoid,' which impairs their performance, whereas Anthropic's constitutional approach is seen as a promising, if not yet proven, method for creating more coherent and reliable systems.

As AI systems become more autonomous, their underlying 'personalities' and value systems, shaped by their training data and fine-tuning, will become critical for their reliability, safety, and utility.

Get started free

Topics

Processed Apr 3, 2026 yt-dlp + mlx-whisper + Gemini