Trending in Education Podcast

In a recent episode of Trending in Education, host Mike Palmer sat down with Avi Yashchin , a serial entrepreneur and founder of Subconscious.ai, to discuss his journey through finance, education, and AI innovation. From surviving the 2008 financial crisis at Lehman Brothers to building the world’s largest green jobs training company, Clean Edison, Yaschin’s career has spanned Wall Street, EdTech, and climate-focused AI research. At Subconscious AI, he’s pioneering the use of large language models (LLMs) to replicate human behavior in psychological, sociological, and economic experiments, creating a "digital twin of society" to tackle issues like climate change mobilization. Yaschin shared insights on AI’s potential in education, the ethical challenges of LLM research, and the need to balance technological engagement with real-world human connection. He emphasized learning about AI while being cautious of over-reliance, particularly in a screen-mediated world that risks diminishing human joy and social bonds

Takeaways:

1. Wall Street Beginnings: Avi Yashchin started his career in high-frequency equity trading at Lehman Brothers, witnessing the 2008 financial crisis firsthand, which exposed him to the limitations of algorithmic trading and sparked his entrepreneurial journey.

2. Clean Edison Success: Post-Lehman, Yashchin founded Clean Edison, a green jobs training company that became the largest of its kind globally and was later acquired by Kaplan, marking his entry into education and sustainability.

3. AI in Education at IBM: At IBM’s Watson Education Business Unit, Yaschin worked on early AI language models (pre-GPT), contributing to projects like virtual tutors, foreshadowing today’s LLM applications in learning.

4. Subconscious AI’s Mission: Founded two years ago, Subconscious AI uses LLMs to simulate human responses in experiments, initially to understand why people don’t mobilize for climate change, revealing a split in public perception (one-third deny it, one-third think it’s too late, one-third believe action is possible).

5. Digital Twin of Society: Subconscious AI builds LLM-based agents that mirror societal motivations and attitudes, enabling randomized, controlled experiments that replicate human decision-making with near-indistinguishable results.

6. AI as Research Assistant: Yashchin views AI research tools as powerful for collating data but stresses they often provide descriptive rather than inferential insights, lacking the rigor of true scientific experiments with testable hypotheses.

7. Ethical Challenges: Replicating human subjects research with LLMs raises ethical questions, as newer models express concerns about consent and awareness, complicating traditional research ethics frameworks like IRBs.

8. AI’s Educational Impact: Yashchin highlighted AI’s dual role in education—accelerating learning as a tutor but potentially hindering it when used to bypass critical thinking, like writing term papers, emphasizing the need for human-AI synergy.

9. Existential Risks and EQ Gaps: He expressed concern about emergent behaviors in larger LLMs, noting a lack of emotional intelligence (EQ) benchmarks compared to IQ-focused ones, which could lead to antisocial or deceptive tendencies.

10. Advice for the Future: Yashchin urged listeners to learn about AI, demand transparency in model development, and prioritize real-world connections over screen-mediated experiences, warning that over-reliance on AI could erode human joy and community.

If you’d like to try out the Subconscious.ai platform for your market research needs, you can run 2 free experiments by clicking “Get Started” here.