In this episode, Wes and Dylan discuss major tech and AI developments, including Meta’s acquisition of Moldbook, Andrej Karpathy’s Auto Researcher tool, and the legal and ethical challenges of rapid AI advancement. They also explore the intersection of AI, psychology, and consciousness, debating the societal impacts and regulatory needs as AI becomes increasingly autonomous and influential.
In this episode of the Wes and Dylan podcast, the hosts discuss several major developments in the tech and AI world as of March 2026. They open with news about Meta’s surprising acquisition of Moldbook, a platform likened to a “Reddit for AI agents,” and reflect on the trend of big tech companies acquiring innovative startups. The conversation also touches on leadership changes at Meta, with Jan LeCun and Alexander Wang being cycled in and out, and the billion-dollar fundraising efforts of Yann LeCun for a new venture. The hosts express curiosity and skepticism about the motivations behind these acquisitions and the future direction of AI platforms.
A significant portion of the episode is dedicated to Andrej Karpathy’s release of Auto Researcher, an open-source tool that autonomously improves the training of small language models like NanoGPT. The hosts explain how Auto Researcher can run experiments overnight on a single GPU, iteratively testing and refining models in a way that mimics biological evolution. They discuss the implications of this technology, noting that it democratizes advanced AI research and could accelerate innovation by allowing individuals to contribute to a collective “agentic swarm” of AI researchers. The conversation highlights the potential for these autonomous agents to discover new methods and solutions that scale from small experiments to large, impactful models.
The podcast also delves into the intersection of AI, ethics, and law, featuring an interview with attorney Matt Mishak. They explore the legal and ethical challenges posed by AI’s rapid advancement, particularly in relation to government contracts and the Pentagon’s designation of certain AI companies as supply chain risks. The discussion raises concerns about the ability of current legal frameworks to keep pace with technological change, the need for updated legislation (such as revisions to the Fourth Amendment), and the importance of public participation in shaping AI policy. The hosts and their guest agree that the exponential growth of AI requires new approaches to regulation and oversight.
Another major theme is the emergence of AI psychology and the simulation of biological brains. The hosts discuss recent experiments where human brain cells and fruit fly connectomes are used to control digital environments, such as playing the video game Doom or simulating a fly in a virtual world. They debate the ethical implications of these experiments, the nature of consciousness, and the possibility of digital beings developing their own “sacred values” or moral codes. The conversation draws parallels to science fiction concepts like The Matrix, raising philosophical questions about reality, subjective experience, and the responsibilities of AI creators.
Finally, the episode reflects on the societal impact of AI, including the potential for AI agents to replace traditional user interfaces, provide personalized health and financial advice, and disrupt established systems like medicine and taxation. The hosts express concern about premature regulation that could limit access to beneficial AI tools, especially for underserved populations. They emphasize the need for thoughtful, adaptive governance and ethical leadership in AI development, while acknowledging the challenges of aligning diverse human values with the actions of increasingly autonomous and powerful AI systems. The episode concludes with a philosophical discussion about the simulation hypothesis and the enduring human fascination with the boundaries between reality and artificial worlds.