#113 Why AI Could Add Decades to Your Lifespan | Dr. Derya Unutmaz

Jul 19, 2026 Episode Page ↗
Overview

Dr. Derya Unutmaz, an immunologist and OpenAI collaborator, discusses how accelerating AI could revolutionize medicine by transforming drug discovery, shortening clinical trials, and enabling personalized treatments. He argues AI is a medical enabler that physicians may soon be ethically obligated to use, potentially adding decades to human lifespan.

At a Glance
13 Insights
2h 45m Duration
24 Topics
10 Concepts

Deep Dive Analysis

Introduction to Dr. Derya Unutmaz and AI's Role

Longevity Escape Velocity and Exponential AI Growth

AI's Acceleration of Biology and Drug Discovery

The Concept of Digital Twins in Clinical Trials

Defining AGI, ASI, and Physical Intelligence

Optimistic View on AI vs. Pessimistic Concerns

AI's Biological Intuition and Experiment Prediction

Ethical Obligation for Physicians to Use AI

Choosing and Integrating AI Models in Medicine

Why Cancer is Difficult to Cure

AI's Role in Curing Cancer and Personalized Treatment

Preventing Cancer Years Before Onset with AI

The Biosingularity Concept and Human 2.0

Bottlenecks and Information Loss in Aging

Engineering Humans for Longevity and Disease Resistance

Yamanaka Factors and Cellular Reprogramming

Challenges of Partial Reprogramming and Aging Hallmarks

Biosecurity Dilemmas and Guardrails in AI

Democratizing Healthcare with AI

Measuring True Aging and Functional Outputs

Reversing Brain Aging and Identity Preservation

The Ultimate Prompt for Extending Lifespan

Building a Personal Mini Digital Twin Today

Importance of Persistent Memory and Context in AI

Longevity Escape Velocity

A hypothetical point where advances in medicine extend remaining life expectancy faster than time passes, meaning for every year you live, more than a year is added to your potential lifespan.

In Silico Experiments

Computational simulations or modeling of experiments performed entirely on a computer, rather than in a lab or on living organisms, used to predict outcomes and test scenarios.

Digital Twin

A comprehensive virtual replica of a biological organism, like a human being, incorporating all available data (genetics, metabolism, immune system, microbiome, etc.) to simulate drug effects, side effects, and treatment efficacy.

Artificial General Intelligence (AGI)

An artificial intelligence that can learn one set of rules or knowledge and generalize that to solve problems in completely different areas, similar to human intelligence.

Artificial Super Intelligence (ASI)

An AI that has the intelligence of the combined totality of humanity, capable of self-learning and training itself millions of times faster than humans, leading to unprecedented problem-solving capabilities.

Physical Intelligence

The ability of an AI or organism to understand and interact with the physical world, possessing a world model that allows it to predict outcomes of physical actions without needing to test them millions of times.

Biosingularity

A concept predicting a point where biological and computational advances, driven by AI, accelerate exponentially to a degree that allows for the treatment of all diseases, complete reversal of aging, and the engineering of 'Human 2.0'.

Yamanaka Factors

Four specific transcription factors (Oct3/4, Sox2, Klf4, c-Myc) that can reprogram adult somatic cells into induced pluripotent stem cells (iPSCs), effectively reverting their age to an embryonic-like state.

Partial Cellular Reprogramming

A technique using Yamanaka factors for a shorter duration or in a controlled manner to rejuvenate cells and tissues without fully reverting them to a pluripotent stem cell state, thus maintaining cell identity while making them more youthful.

Sick Care vs. Health Care

The current medical system is primarily 'sick care' (treating diseases after they manifest) rather than 'health care' (preventing diseases before they happen), a paradigm AI is expected to shift towards preventative medicine.

?
Why might the next 10-15 years be critical for longevity?

The next 10-15 years are critical due to the exponential acceleration of AI and medical technology, potentially leading to 'longevity escape velocity' where each year lived adds more than a year to one's remaining life expectancy.

?
How is AI currently accelerating biological research and drug discovery?

AI models can analyze massive biological datasets, extract insights, generate hypotheses, and even simulate and recommend the most effective experiments, dramatically reducing the time and effort required for R&D and drug screening.

?
What is a 'digital twin' and how will it impact clinical trials?

A digital twin is a comprehensive virtual simulation of a human being, integrating all biological data; it will allow for in silico clinical trials, drastically shortening drug development times from years to months or weeks by predicting efficacy and side effects.

?
Have we already achieved Artificial General Intelligence (AGI)?

Dr. Unutmaz believes we have achieved 'level one AGI' because large language models (LLMs) can generalize knowledge across different domains, such as applying video game logic to biological experiments.

?
Why is Dr. Unutmaz optimistic about AI despite common fears?

He views AI as an incredible enabler that gives humans 'superpowers' and believes the true existential threat is humanity itself, not AI, which he sees as a collaborator that can save billions of lives and enhance human capabilities.

?
Can AI develop biological intuition like a human scientist?

Advanced AI models, particularly reasoning models like GPT-5.5 Pro, are approaching human-level biological intuition, capable of predicting complex experimental outcomes with high accuracy, similar to an experienced scientist.

?
Why might it become medical malpractice for doctors not to use AI?

As AI models become highly reliable at diagnosing diseases and designing treatment protocols as effectively as specialists, failing to use these tools could lead to preventable misdiagnoses and mistreatments, making it ethically and eventually legally irresponsible.

?
How should physicians choose which AI models to use?

Physicians should continuously update their AI knowledge, always use the latest, top-tier 'pro' or 'thinking' models for complex problems, and consider specialized platforms that integrate the best available AI for medical applications.

?
Why is cancer so difficult to cure, and how can AI help?

Cancer is challenging because it's hundreds of diseases, and cancer cells are part of the body, making them hard to target without harming normal cells. AI can accelerate personalized treatments by modeling mutations, screening compounds, and designing highly specific therapies like mRNA vaccines.

?
Can AI predict cancer and other diseases years before they form?

Yes, studies using large biobanks show that AI can predict diseases like cancer and neurodegenerative conditions years in advance by analyzing vast amounts of biomarker data, enabling proactive preventative interventions.

?
What is the 'biosingularity' and 'Human 2.0'?

Biosingularity is a predicted point where exponential advances in biology and computation, driven by AI, allow for the complete understanding and engineering of human biology, leading to 'Human 2.0' with enhanced, re-engineered biological systems.

?
Is reversing aging possible, and what are the bottlenecks?

Yes, aging is considered reversible, as demonstrated by cellular reprogramming. The bottleneck is understanding and correcting the 'information loss' that causes the breakdown of the body's natural repair programs, and then engineering precise interventions for complex, multi-system changes.

?
What are Yamanaka factors and their potential for age reversal?

Yamanaka factors are proteins that can revert adult cells to an induced pluripotent stem cell state, effectively erasing their age. Partial reprogramming using these factors aims to rejuvenate cells and tissues without losing their identity, offering a path to age reversal.

?
How can someone build a 'mini digital twin' today?

A mini digital twin can be built by consistently feeding a top-tier AI model (like GPT 5.5 Pro) with personal longitudinal data, including lab values, continuous glucose monitoring, activity, sleep, supplements, and diet, to establish baselines and receive personalized health advice.

?
What is the importance of persistent memory and context for AI in personal health?

Persistent memory and expanded context handling allow AI models to remember an individual's historical health data over long periods, enabling them to identify trends, correlate interventions with outcomes, and provide more accurate, personalized health predictions and advice.

1. Try to Live 10-15 More Years

Dr. Unutmaz predicts that the next 10-15 years are a critical window where medical advances, especially due to AI, could lead to ’longevity escape velocity,’ meaning each year lived adds more than a year to remaining life expectancy. This suggests a strong incentive to maintain health during this period.

2. Physicians Should Use AI

It is becoming unethical, and will eventually be considered malpractice, for physicians not to use AI for diagnosis and treatment planning, as advanced models can diagnose better than average doctors and even specialists, reducing misdiagnoses and improving patient outcomes.

3. Prioritize Latest AI Models

For complex medical problems, always use the latest, top-tier AI models (e.g., GPT 5.5 Pro) and specifically the ‘pro’ or ’thinking’ models, as they offer deeper reasoning and insights compared to instant or older versions.

4. Use AI for Data Analysis

Scientists and researchers should leverage advanced AI models (like GPT-5 Pro) to analyze large biological datasets, identify insights, and generate hypotheses, significantly accelerating research and development by condensing months or years of analytical work into minutes or hours.

5. Use AI for Experiment Design

AI models can simulate and recommend the most effective experiments from hundreds of possibilities, saving significant time and resources in biological research by identifying the best options instead of trying many.

6. Integrate AI into Hospital Systems

Hospitals should implement enterprise-level AI for continuous patient monitoring, data analysis, and providing real-time information to nurses and doctors, similar to how big tech companies use AI.

7. Continuously Update AI Knowledge

Doctors must constantly update their knowledge of the latest AI models and their capabilities, just as they update medical knowledge and drug information, to ensure they are using the most advanced and effective tools in their practice.

8. Focus on Preventative Medicine

Leverage AI’s ability to predict diseases years before they manifest (e.g., cancer, neurodegenerative disease) based on comprehensive biomarker data. This allows for early lifestyle, dietary, or other interventions to prevent disease onset.

9. Build a Mini Digital Twin Today

Start constructing a personal ‘mini digital twin’ by consistently collecting and feeding AI models with daily health data such as lab values, glucose levels, steps, sleep, supplements, and food intake. This helps the AI personalize advice.

10. Maintain AI Context for Personal Data

When building a mini digital twin, keep all personal health data within the same AI context or window, or use a personal database that the AI can access. This allows the model to remember historical data and identify changes before and after interventions.

11. Establish Personal Baselines

Use continuous data collection (e.g., glucose monitors, regular lab tests) to establish personal ‘set point normals’ for various biomarkers, rather than relying solely on population-based ranges. AI can then provide more personalized advice based on individual fluctuations.

12. Experiment and Validate AI Advice

For non-harmful daily uses, conduct personal experiments (e.g., stopping a supplement for a period) to validate AI’s predictions by observing before-and-after changes in your health data. This helps build trust and refine AI’s personalized recommendations.

13. Cultivate Optimism for Longevity

Being optimistic is identified as one of the best things one can do for aging, with studies showing its positive impact on resilience and longevity.

The next 10 years, you can think of it as more advanced than the last century.

Dr. Derya Unutmaz

We will come to a point in the next, I would say, probably eight to 10 years, where every year you live is going to add more than a year to your life.

Dr. Derya Unutmaz

There's only one existential threat to humanity, and that's humanity.

Dr. Derya Unutmaz

AI is is is an incredible enabler, it gives you superpowers.

Dr. Derya Unutmaz

90% or 95% is based on what's known, how you process that knowledge, but there's that extra five 10% totally dependent on your intuition.

Dr. Derya Unutmaz

Right now, it's unethical for physicians not to use AI anymore.

Dr. Derya Unutmaz

Cancer is going to be 100% curable, probably less than a decade.

Dr. Derya Unutmaz

We don't have health care. We have sick care, right? So we never take care of healthy people.

Dr. Derya Unutmaz

Biology, we think is a miracle, but it's a it's a bad kind of a legacy engineering, right?

Dr. Derya Unutmaz

Living an extra one year could make you reach that threshold where there's going to be the ability to, to treat many diseases and reverse your aging and give you another decade, give you another 20 years.

Dr. Derya Unutmaz

Building a Mini Digital Twin

Dr. Derya Unutmaz
  1. Provide AI with as much personal health data as possible on a daily basis.
  2. Include lab values (cholesterol, glucose, etc.), daily steps, sleep data, supplements taken, and types of food eaten.
  3. Maintain this data within the same AI context window or as a personal database that the AI can access, allowing it to remember historical information.
  4. Use the AI to analyze changes based on new data and provide personalized suggestions.
  5. Establish personal 'set point normals' for biomarkers by collecting frequent measurements (e.g., continuous glucose monitoring, regular lab tests).
  6. Conduct personal experiments (e.g., temporarily stopping a supplement) and observe before-and-after data to validate AI predictions, for non-harmful daily uses.
10-15 years
Predicted window for reaching longevity escape velocity Dr. Derya Unutmaz's prediction
50 years
Potential additional lifespan predicted if longevity escape velocity is reached Dr. Derya Unutmaz's prediction
5-10 years
Lifespan extension observed with GLP-1 drugs for obese or chronic conditions Mentioned by Dr. Derya Unutmaz
15-20 years
Maximum predicted timeframe for complete aging reversal Dr. Derya Unutmaz's prediction
10 times safer
Safety factor for self-driving cars (current) Compared to human drivers, aiming for 100 times safer
2 hours
Longest reasoning time Dr. Unutmaz pushed GPT-5.5 Pro for complex data analysis For analyzing huge datasets and generating insights
40 pages
Length of report generated by GPT-5.5 Pro from a large immunological dataset Containing analysis and insights
12 million
Estimated number of misdiagnoses in the US alone annually Mentioned by Dr. Derya Unutmaz
700,000
Number of people who suffer or die from misdiagnoses in the US annually Mentioned by Dr. Derya Unutmaz
1 in 5,000 to 1 in 10,000
Rate of myocarditis in young people vaccinated for COVID-19 Mostly non-fatal
1 in 1,000 or 1 in 10,000
Chance of a young person dying from COVID-19 Mentioned as a comparison to vaccine side effects
1 out of 5 or 1 out of 10
Proportion of people who truly benefit from statins for high cholesterol Illustrates lack of personalization in current medicine
500,000 people
Size of the UK Biobank dataset used to predict diseases Used in a study where AI predicted 1,000 diseases
1,000 diseases
Number of diseases predicted before they happened using UK Biobank data and AI Based on retroactive analysis of collected data
200 plasma proteins
Number of plasma proteins analyzed in the UK Biobank study for disease prediction Compared to typical 10-20 proteins in standard checkups
25 years ago
Years ago Dr. Unutmaz started his 'biosingularity' blog Predicting AI's impact on biology
2035
Predicted year for AI to treat all diseases according to Dr. Unutmaz's early calculations From his 'biosingularity' blog
2045
Predicted year for AI to completely reverse aging according to Dr. Unutmaz's early calculations From his 'biosingularity' blog
2050
Predicted year for reaching 'Human 2.0' according to Dr. Unutmaz's early calculations From his 'biosingularity' blog
1996-1997
Year Dolly the cloned sheep was created A pivotal moment for understanding biological regeneration
2016
Year Yamanaka factors were discovered (or became widely recognized for reprogramming) A pivotal moment for understanding cellular age reversal
123 years old
The oldest recorded human lifespan A French woman
115-116 years old
Considered the current limit of human lifespan Discussed as a general observation
300 people
Number of people in the world aged 110 and older Highlights the rarity of supercentenarians
50-fold
Increase in effectiveness/efficiency of Yamanaka factors with AI-designed mutations Reported in a recent study using an AI model
1 million tokens
Current context window limit for some AI models For models like Gemini or Claude, allowing for large data inputs