Hi, I'm Kevin
I host Hack Club's AMA series.
Reach out at me@kevintlu.com if you're interested in being a guest!
Recent AMAs
Danijar Hafner
November 12, 2025 – In this AMA, Danijar Hafner, a staff research scientist at Google DeepMind, discusses his work on Dreamer 4, a significant advancement in world models and reinforcement learning. He explains how Dreamer 4 allows AI agents to learn and simulate environments, particularly in complex scenarios like Minecraft. The discussion covers the evolution of the Dreamer models, the challenges of long-term memory in AI, and the implications for robotics. Hafner also compares Dreamer 4 with OpenAI's approaches, shares insights about the culture at Google DeepMind, and offers advice for aspiring AI researchers.
Professor Karen Livescu
September 24, 2025 – In this AMA, Professor Karen Livescu discusses her research in speech and language processing, focusing on multimodal representation learning across spoken, written, and signed languages. She explores efforts to make AI language models more universal and capable. The discussion dives into challenges in multimodal learning, the importance of inclusivity in AI design, and the future of speech technology. There are key emphasis on collaboration, data efficiency, and ethical considerations in developing systems that bridge human communication and machine understanding.
Professor Michael Littman
September 16, 2025 – In this AMA, Brown Professor Michael L. Littman, pioneer in machine learning and reinforced learning, discusses his encounters with AI usage its effects on the education of student, productivity of workers, and development of core fundamental skills. Through wonderful personal stories, references to various studies, and opinion of other educators, he paints a comprehensive picture of where we stand in terms of AI risks.
Dr. Ilan Strauss
September 10, 2025 – In this AMA, Dr. Ilan Strauss discusses the economic implications of artificial intelligence, focusing on market structures, protocols for competition, and the evaluation of AI systems. He explores the potential impacts of AI on employment, taxation policies, and the contrasting economic models of China and the United States. The discussion also touches on the concept of degrowth and the future of AI in society, emphasizing the need for proactive measures to ensure equitable outcomes in an increasingly automated world.
Tim O'Reilly
August 29, 2025 – In this AMA, Tim O'Reilly, founder of O'Reilly Media, shares his journey from a tech writer to a leading figure in publishing and technology. He discusses the evolution of his company, the impact of AI on society, and the importance of ethical considerations in technology development. O'Reilly emphasizes the need for entrepreneurs to focus on meaningful work rather than just financial gain, and he introduces the AI Disclosures Project aimed at promoting transparency in AI technologies.
Armando Solar-Lezama
August 27, 2025 – In this AMA, Professor Armando Solar-Lezama discusses the concept of program synthesis, its evolution, and its intersection with AI and deep learning. He emphasizes the importance of collaboration between machines and developers, the role of formal guarantees in program synthesis, and the potential of neuro-symbolic programming in scientific discovery. The discussion also touches on the challenges of AI safety and the unique contributions of academia to the field of AI research.
Buck Shlegeris
June 9, 2025 – In this AMA, Buck discusses his work as CEO of Redwood Research on AI control and safety. He emphasizes the importance of understanding AI's potential to misalign with human interests and the need for effective control strategies. The discussion also touches on the role of interpretability in AI, the societal implications of AI advancements, and the responsibilities of AI companies in mitigating risks.
Nikita Dhawan
May 19, 2025 – In this episode, we spoke with Nikita Dhawan, a PhD student at the University of Toronto, about her groundbreaking discoveries in federated learning through FedFish and connecting the medical industry with AI through NATURAL. We explored her expertise in federated learning and its improvements versus the old, and we journeyed with her through her passion of improving medical efficiency using LLMs to interpret texts on medical forums.
Zeqi Xiao
April 22, 2025 – In this episode, I spoke with Zeqi Xiao, a PhD researcher at MMLab@NTU, about his groundbreaking work on WORLDMEM, a framework for long-term world simulation with memory. We explored how generative models are evolving to not just imagine, but remember and build coherent virtual worlds over time.
Eric Zhu
April 21, 2025 – In this episode, I sat down with 17-year-old Eric Zhu – founder of Aviato, a startup building the Bloomberg terminal for private markets, which raised $2.3M from 8VC, Soma Capital, and SoftBank. He's also a partner at Bachmanity Capital, a $20M venture fund, and the creator of the sport sperm racing (yes, really). We talk about his upcoming event on April 25th, with 4,000 spectators at the LA Center Studios.
Arthur Conmy
April 18, 2025 – In this episode, I hosted an Ask-Me-Anything with over 100 RSVPs. Our guest, Arthur Conmy, is a leading AI safety researcher at Google DeepMind and the University of Cambridge, known for pioneering automated circuit discovery and training over 400 sparse autoencoders for Gemma Scope. Whether you're new to AI or deeply into mechanistic interpretability, come learn how he's helping lift the lid on black-box LLMs.
Kevin @ CHAI Workshop 2025
