This Week in Startups
This Week in Startups

Oct 5, 2026 · 1h 11m

AI tutor turns productive struggle into personalized learning

Jason Mentors The Next Generation + Why AI Is Bad at Teaching (Aristotle Interview) | E2346

The episode tests whether AI can expand one-on-one teaching without reducing education to instant answers.

3 key takeaways
  1. 1AI-first careers still depend on durable skills, practical projects, biographies, and trusted relationships.
  2. 2Aristotle uses adaptive questions, pauses, visuals, and explanations to teach understanding rather than merely generate solutions.
  3. 3Personalized AI tutoring could widen access to one-on-one support while raising questions about economics, privacy, and classroom adoption.

Don't miss

Shan Reddy and Jason Calacanis use Aristotle to teach integration through a driving example, showing how the tutor adapts questions, visuals, and pacing.

The brief

Jason Calacanis and Lon Harris discuss preparing for an AI-driven economy, arguing that practical projects, biography-driven learning, and trusted relationships remain valuable as tools change.

The conversation then turns to AI in medicine and education, where the central question shifts from what models can answer to whether they can build durable understanding.

Shan Reddy, co-founder and CEO of Aristotle, demonstrates a voice-first tutor teaching integration through driving examples, graphs, questions, and explanations tied to accumulated change.

Aristotle’s approach predicts the next teaching move—speaking, pausing, drawing, or assigning another problem—aiming to reproduce the benefits of individualized instruction.

The discussion closes on tutoring economics, dyslexic students, private data, teacher support, and whether AI can extend one-on-one learning beyond well-resourced families.

What was said on this episode

28 statements · 18 positive · 8 negative · 1 mixed · 1 neutral

  1. Jason Calacanison AI-assisted finance educationPositive4:33

    Finance students should use AI first to deepen coursework and build expertise.

    “So you want to be AI first.”

    Listen at 4:33

  2. Jason Calacanison Biography readingPositive5:52

    Students should read approximately one biography monthly to become more well-rounded.

    “the other unlock I want to give you as a gift is to read a biography a month.”

    Listen at 5:52

  3. Jason Calacanison Biography readingPositive7:40

    Reading biographies can make students more worldly than their classmates.

    “you will be more worldly than any of your classmates.”

    Listen at 7:40

  4. Jason Calacanison Customer trustPositive10:09

    Customer trust is necessary for closing business deals.

    “Trust closes deals.”

    Listen at 10:09

  5. Jason Calacanison Technology angel investingPositive15:35
    “I would've become an angel investor in the technology industry.”

    Listen at 15:35

  6. Jason Calacanison Entrepreneurial projectsPositive16:45

    Early entrepreneurs should start projects rather than immediately starting companies.

    “instead of starting companies, I would start projects.”

    Listen at 16:45

  7. Jason Calacanison Entrepreneurial operating skillsPositive17:48

    Entrepreneurs should learn every operational component of running a business.

    “What you want to be is unstoppable, that you can do every single thing from your taxes to, you know, negotiating a non-disclosure agreement, from starting an LLC, doing your payroll.”

    Listen at 17:48

  8. Jason Calacanison Popular AI modelsPositive19:53

    Popular AI models can handle most workloads adequately.

    “The truth is most of the big popular AI models can handle your workloads just fine”

    Listen at 19:53

  9. Jason Calacanison AI in medicineNeutral25:13

    AI adoption in medicine is currently proceeding slowly and methodically.

    “AI is being used, um, slowly, methodically.”

    Listen at 25:13

  10. Jason Calacanison AI in medicineNegative26:41

    AI applications in medicine face slow regulatory processes.

    “it's going to be slow if you're going to go through the regulatory process”

    Listen at 26:41

  11. Jason Calacanison Drug-discovery software toolsPositive27:03

    Drug-discovery software tools face an easier path than regulated medical applications.

    “if you're making tools, that's super easy.”

    Listen at 27:03

  12. Jason Calacanison Customer discoveryPositive33:19

    Customer discovery can generate both product feedback and potential customers.

    “every time you do that customer intelligence, you've got a potential customer to sell into.”

    Listen at 33:19

  13. Frontier AI labs promote AI danger narratives because perceived power increases product usage.

    “It's coming from the frontier labs who have learned that every time they put a release out that their products are incredibly powerful and can hack a system, etc., their usage goes up.”

    Listen at 34:40

  14. Jason Calacanison AI consciousnessPositive38:06

    If AI consciousness is being built, it should receive compassion and strong moral character.

    “if we are building a consciousness here, we should treat it with compassion and we should imbibe it with really good morals and character.”

    Listen at 38:06

  15. Jason Calacanison Chatbots and AI systemsNegative39:48

    Jason has never experienced an AI system as alive or capable of feeling.

    “I've never once had an encounter with a chatbot or an AI where it made me think like, this thing is alive.”

    Listen at 39:48

  16. Jason Calacanison AI tutoringPositive44:02

    Well-implemented AI could provide exceptionally capable tutoring.

    “AI provides the greatest tutor you could ever have if implemented”

    Listen at 44:02

  17. Shan Reddyon General AI teachingNegative45:53

    General AI systems are worsening rather than improving at teaching.

    “AI is actually not getting better at teaching. It's actually getting worse.”

    Listen at 45:53

  18. Shan Reddyon Artificial General IntelligenceNegative45:57

    Scaling general intelligence has not increased AI’s pedagogical capacity.

    “The bet that the big labs have made is that scaling general intelligence is going to scale pedagogical capacity. Unfortunately, that has not been the case.”

    Listen at 45:57

  19. Shan Reddyon Artificial General IntelligencePositive46:14

    General AI models are improving at complex reasoning, enterprise tasks, and coding.

    “the models are getting better at long horizon complex reasoning, enterprise question answering, enterprise problem solving, coding, and so on.”

    Listen at 46:14

  20. Shan Reddyon AI teaching pedagogyPositive47:15

    Effective AI teaching should preserve students’ productive struggle instead of delegating learning.

    “when it comes to teaching, what we are trying to do is make sure that the student has the productive struggle”

    Listen at 47:15

  21. Shan Reddyon Large Language ModelNegative52:29

    Language models currently struggle to choose silence, examples, diagrams, or whiteboards pedagogically.

    “these are all things that are difficult for language models to do right now”

    Listen at 52:29

  22. Shan Reddyon One-on-one voice tutoringPositive54:05

    One-on-one voice tutoring produces the best learning outcomes.

    “one-on-one tutoring via voice, as human student tutors have done for so long, is the best way to learn, drives the best outcomes.”

    Listen at 54:05

  23. Jason Calacanison Private tutoringNegative55:00

    Affording private tutoring gives wealthy students a major educational advantage.

    “if you're a rich person who can afford a tutor, you're going to have an immense advantage over everybody else in the world.”

    Listen at 55:00

  24. Shan Reddyon US public educationNegative1:06:42

    Public education quality is poor in most regions and deteriorating.

    “the product that the public sector is putting out there is not great in most geographies and is getting worse.”

    Listen at 1:06:42

  25. Shan Reddyon US education spending and outcomesNegative1:07:05

    US education spending is rising while educational outcomes are worsening.

    “we're spending more and more on worse product.”

    Listen at 1:07:05

  26. Shan Reddyon AI curriculum deliveryPositive1:07:13

    AI will solve curriculum delivery through personalized instruction.

    “learning and teaching in terms of curricula is solved. AI delivers it and it delivers it magically and personally attuned to every student.”

    Listen at 1:07:13

  27. Shan Reddyon AI-delivered educationPositive1:07:17

    AI can deliver personalized curricula and teaching to every student.

    “AI delivers it and it delivers it magically and personally attuned to every student.”

    Listen at 1:07:17

  28. Shan Reddyon AI educationPositive1:07:49

    AI will solve students’ learning and teaching needs while schools emphasize broader development.

    “students and their— the students' learning and teaching is solved with AI.”

    Listen at 1:07:49

Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.

Books & mentions

Some links are affiliate links — PodLume may earn a commission if you buy.

Weekly email

State of the Pod, every Monday

The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.

Listen to the full episode and explore every guest, topic, and moment on PodLume.