Public transcript

FULL FIRESIDE: Open AI CEO Sam Altman & VC Vinod Khosla at IIT Delhi

By: Diya TV
Length: 53 min
Language: English
Text from: the video's captions
Transcribed: 12 October 2026

Read the full transcript Open on YouTube

Notes in English.

Summary

This fireside chat at IIT Delhi features OpenAI CEO Sam Altman and venture capitalist Vinod Khosla, an IIT Delhi alumnus, discussing how AI will transform work, education, the economy and society. They urge students to develop agency, curiosity, risk-taking and fast learning rather than relying on technical skills alone. Sam Altman then takes student questions on AI safety, jobs, health AI, green energy, creativity and India's role in AI, ending with a rapid-fire round.

Key points

Sam Altman says India is OpenAI's second largest market with 100 million ChatGPT users and the fastest growing Codex market in the world.
Vinod Khosla tells students that no technical skill they learn will stay relevant and they should learn how to learn really fast.
Altman says the cost of solving a hard reasoning problem via OpenAI's API dropped about 1000x in 18 months.
Both speakers identify curiosity, agency, risk-taking and persistence as the most important skills for young people.
Altman outlines three AI safety pillars: technical alignment, new security architecture for agents, and resilience through democratized access.
Altman expects a hugely deflationary economy by 2035 and believes GDP will be a poor measure of progress.
Altman dismisses data center water concerns as a sticky meme but says the energy transition to fusion and renewables is real and near.
In rapid fire, Altman names resilience as the one skill every IIT student should build this year.

Questions it answers

00:00Welcome and introductions

Who are the guests and why are they at IIT Delhi?

Sam Altman, CEO of OpenAI, and Vinod Khosla, founder of Khosla Ventures and IIT Delhi alumnus, are welcomed for a fireside conversation. The director, Professor Rangan Banerjee, notes the huge student interest and describes IIT Delhi's AI initiatives, including a partnership with OpenAI.

Vinod Khosla co-founded Sun Microsystems and was the first institutional investor in OpenAI.
IIT Delhi is entering a partnership with OpenAI and runs AI-focused programs including a School of AI.
Khosla set up one of the first computer laboratories in the IIT system as a student.

07:12The scale of coming change

How big a change do Altman and Khosla expect AI to bring?

Both argue the world underestimates AI's impact. Altman says his one-year-old child will never know being smarter than a computer, and that scientific discovery and the economy will both be automated. Khosla suggests GDP forecasts are far too low and imagines India's transformation by its 100th anniversary of independence in 2047.

Altman says GDP will be a poor metric because AI is so deflationary.
He expects a hugely deflationary economy by 2035.
Khosla compares the change between now and 2040 to the delta between today and 1970, and calls that conservative.

14:30What students should learn

What should young people focus on learning if technical knowledge becomes obsolete?

Khosla says to learn how to learn really fast, since technical expertise will be outpaced by AI. Altman emphasizes staying at the forefront of AI tools, building projects constantly, and learning to make something people want. Both agree curiosity and agency are the most important skills.

Altman says learning to use AI tools is today's equivalent of learning to program in his school days.
Khosla says any technical skill you learn, AI will get better at.
Altman calls agency a learnable skill with returns that have never been higher.

19:00Agency and risk-taking

Why do the speakers stress agency and risk-taking?

Altman describes how unwavering conviction and knocking down obstacles made OpenAI happen despite widespread skepticism. Khosla shares his own student examples of creating programs at IIT Delhi through sheer will, and encourages students, especially in India, to embrace risk and persistence.

Altman says OpenAI seemed like a very bad idea for its first four or five years.
Khosla cites a Harvard Business School case quoting him: 'My willingness to fail allows me to succeed.'
Khosla mentions Emergent exceeding 100 million ARR in 8 months as proof of startup opportunity.

25:00India and OpenAI's market

How does Sam Altman view India's role in AI?

Altman says India is already OpenAI's second largest market with 100 million ChatGPT users and is the fastest growing Codex market. He expects India to become the largest market and praises the building energy of young Indian developers.

India has 100 million ChatGPT users, a third of them students.
Codex is growing faster in India than anywhere in the world.
Altman expects India to become OpenAI's largest market over time.

26:00Cost, access and safety

How does OpenAI reconcile premium pricing with broad access, and how does it keep AI safe?

Altman says the cost of a given level of intelligence has dropped about 1000x in 18 months and OpenAI aims to drive it toward zero. On safety, he outlines three areas: technical alignment, new security infrastructure for agents, and resilience through democratized access so good actors can counter bad ones.

The cost of solving a hard reasoning problem via OpenAI's API dropped about 1000x in 18 months.
Altman identifies alignment, security architecture, and resilience as the three safety pillars.
He argues democratized access to powerful AI is the safest path forward.

34:00Health, jobs and human advantage

How does Altman respond to questions about health AI, jobs and human relevance?

On health, he says ChatGPT works well without medical records and OpenAI would love to connect to Indian platforms. On jobs, he says some will vanish, many will persist because humans want humans, and history shows we underestimate how quickly new jobs appear. Humans remain strong at high-level strategy and caring about other people.

A huge volume of ChatGPT usage is health-related queries with no connectors.
Altman says anyone only excited or only fearful about AI is not being thoughtful.
He notes people who use AI proficiently have a huge advantage over those who don't.

43:00Energy, creativity and rapid fire

What does Altman say about green AI, design, water use, and quick predictions?

Altman says the world is a few years from fusion and cheap renewables, and dismisses water scarcity concerns from data centers as outdated. On creativity, he says creators find AI a great tool and human imprint still drives value. He ends with rapid-fire answers, including hoping the next trillion-dollar AI company comes from India.

Altman says AI data centers are not significant water consumers and the water concern is a sticky meme.
He says art made purely by AI has effectively zero value to people, while human-made work with AI tools is worth a lot.
In rapid fire, he picks education as the bigger five-year change and resilience as the one skill to build.

Notes

The Magnitude of Change

Sam Altman calls AI the biggest shift in what a young person can accomplish in his lifetime; his one-year-old will never know being smarter than a computer.
Vinod Khosla frames the minimum change to 2040 as the delta between today and 1970 — and calls that conservative.
Altman argues GDP will be a terrible metric because AI is so deflationary; he expects a hugely deflationary economy by 2035.
Scientific discovery and the economic engine that commercializes it will both be automated, at rates humans have no intuition for.

What Students Should Learn

Khosla: learn how to learn really fast; technical expertise will largely be irrelevant by 2030.
Altman: learning to use AI tools is the modern equivalent of learning to program; staying at the forefront of tools is a compounding accelerant.
Both stress curiosity and agency as the key skills; Altman calls agency learnable and says its returns have never been higher.
Altman's career lesson: be right on the idea, stick with unwavering conviction, and knock down obstacles with no loss of enthusiasm.
Khosla quotes himself from a Harvard Business School case: "My willingness to fail allows me to succeed."

India and OpenAI

India is OpenAI's second largest market: 100 million ChatGPT users, a third of them students; fastest growing Codex market in the world.
OpenAI announced offices in Bangalore and Mumbai beyond Delhi, plus a Tata partnership on compute and data centers.
IIT Delhi is entering a partnership with OpenAI.

Safety, Cost and Access

Cost of a hard reasoning problem via OpenAI's API dropped about 1000x in 18 months; goal is to drive intelligence cost toward zero.
Three safety pillars: technical alignment, new security architecture for agents (e.g., prompt injection), and resilience through democratized access.

Jobs, Health, Energy, Creativity

Some jobs will vanish; many will persist because humans want humans; new job categories always emerge faster than expected.
ChatGPT health works well without medical records; a huge volume of queries are health-related.
World is a few years from fusion and cheap solar/wind plus storage; water concerns about data centers are a sticky but false meme.
Art made purely by AI has near-zero value to people; human imprint drives worth.

Rapid Fire

Hopes the next trillion-dollar AI company comes from India; most overrated fear is killer robots; underrated opportunity is zero-person companies; education changes more than healthcare in five years; one skill for IIT students: resilience.

Text from the video's own captions. Summary and notes written by AI from the transcript, so check anything important against the recording.