147 épisodes
- In this episode of Tech Anatomy, Anca Petre talks with Jon Stokes, Assistant Professor at McMaster University and co-founder of Stoked Bio. In 2020, he published in Cell the discovery of halicin, an antibiotic found with deep learning. He calls himself a microbiologist first, "not an AI person", and treats machine learning as a tool, a wrench, to tackle antibiotic resistance. They talk about what AI really changes in antibiotic discovery, what it does not, and why speed matters.
Topics covered:
From testing molecules at the bench to working with computer scientists.
AI as a wrench: a tool for the problem, not the other way around.
Why antibiotics are essential, and why so few companies want to make them.
The Swiss cheese problem: why antibiotic discovery is hard.
Fire, the best antibacterial agent ever, and a terrible antibiotic.
An honest look at where AI actually helps: finding hits.
Resistance: "life finds a way", and what we can do about it.
Real-world prescribing and antibiotic stewardship.
Chapters:
0:00 Introduction
0:58 From the bench to machine learning
6:05 "I'm still not an AI person"
8:39 AI as a wrench
9:47 The antibiotic crisis in numbers
11:47 Why nobody wants to make antibiotics
13:47 The Swiss cheese problem
15:39 The best antibacterial agent: fire
19:59 Where AI actually helps
25:30 Can real-world resistance data help?
29:13 Real-world prescribing
32:54 Just Jon, and a lab
34:22 Conclusion
The study discussed: Stokes JM et al. A Deep Learning Approach to Antibiotic Discovery. Cell 2020;180(4):688-702. https://doi.org/10.1016/j.cell.2020.01.021
Credits:
Production: MedShake Studio
Host: Anca Petre
Subscribe to the Tech Anatomy newsletter:
https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7117418963067658241
Follow Anca on LinkedIn for more tips and free resources:
https://www.linkedin.com/in/ancapetre/
WhatsApp group, AI in health tips:
https://chat.whatsapp.com/CMdYp3pfcRrBX9Qxo2sKAz
Hosted on Ausha. See ausha.co/privacy-policy for more information. - Tomorrow on Tech Anatomy: Anca Petre talks with Jon Stokes, Assistant Professor at McMaster University and co-founder of Stoked Bio, who published in 2020 the discovery of halicin, an antibiotic found with deep learning. What can AI really do against antibiotic resistance?
Full episode on Tuesday, November 17.
Subscribe to the Tech Anatomy newsletter:
https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7117418963067658241
Hosted on Ausha. See ausha.co/privacy-policy for more information. - A short episode of Tech Anatomy: Anca Petre puts a series of statements on AI and mental health to Nick Jacobson, Associate Professor at Dartmouth and creator of Therabot, a generative AI therapy chatbot. He agrees, or he disagrees, and explains why.
The full episode with Nick Jacobson is available in this feed and on YouTube.
Credits
Production: MedShake Studio
Host: Anca Petre
Subscribe to the Tech Anatomy newsletter: https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7117418963067658241
Follow Anca on LinkedIn for more tips and free resources: https://www.linkedin.com/in/ancapetre/
Hosted on Ausha. See ausha.co/privacy-policy for more information. - In this episode of Tech Anatomy, Anca Petre talks with Nick Jacobson, Associate Professor of Biomedical Data Science and Psychiatry at Dartmouth's Geisel School of Medicine. With his team, he built Therabot, a generative AI therapy chatbot, and tested it in a randomised controlled trial published in NEJM AI in 2025. They talk about what it took to build it, what the trial shows and what it does not, and what comes next.
Topics covered
From measuring depression with smartphones to building a chatbot that talks with patients.
What Therabot is, and what it is not.
Why the first versions failed: training on online forums, then on real therapy transcripts.
Writing "gold-standard" therapy dialogues by hand, and having them reviewed by peers.
Agree to Disagree: Anca's statements on AI and mental health, and Nick's answers.
The randomised trial: 210 adults, Therabot versus a waiting list, and how to read the results.
Reviewing every message, and how to scale safely.
The next trial, and the question of what Therabot should be compared with.
Chapters
0:00 Introduction
1:10 From measuring depression to talking with patients
3:06 What is Therabot?
4:49 First attempts: forums and therapy transcripts
6:04 Designing gold-standard therapy data
8:34 Why real therapy transcripts are not enough
11:23 Agree to Disagree
17:19 The randomised trial
19:45 Reviewing every message, and scaling safely
22:57 What comparator for the next trial?
24:02 Conclusion
The study discussed: Heinz MV et al. Randomized Trial of a Generative AI Chatbot for Mental Health Treatment. NEJM AI 2025;2(4). https://doi.org/10.1056/AIoa2400802
Credits
Production: MedShake Studio
Host: Anca Petre
Subscribe to the Tech Anatomy newsletter: https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7117418963067658241
Follow Anca on LinkedIn for more tips and free resources: https://www.linkedin.com/in/ancapetre/
WhatsApp group, AI in health tips: https://chat.whatsapp.com/CMdYp3pfcRrBX9Qxo2sKAz
Hosted on Ausha. See ausha.co/privacy-policy for more information. - Trailer of the new Tech Anatomy episode with Nick Jacobson, Associate Professor at Dartmouth and creator of Therabot, a generative AI therapy chatbot tested in a randomised controlled trial.
Full episode out on Tuesday 10 November, in this feed and on YouTube.
Credits
Production: MedShake Studio
Host: Anca Petre
Follow Anca on LinkedIn: https://www.linkedin.com/in/ancapetre/
Hosted on Ausha. See ausha.co/privacy-policy for more information.
Plus de podcasts Forme et santé
Podcasts tendance de Forme et santé
À propos de Tech Anatomy
Tech Anatomy — Health tech massively simplified
Decoding AI, Web3, and digital health so health industry leaders can turn complexity into action.
Welcome to Tech Anatomy — your bilingual (English & French) dive into the intersection of health and technology. I’m Anca Petre (pharmacist, entrepreneur, health-tech strategist), and here we make the complex accessible.
Produced by MedShake Studio, a leading podcast production studio dedicated to healthcare, this channel delivers:
Clear, structured explanations of AI, Web3, digital twins, data analytics, virtual care, and more
Use cases, trends, and critical insights — grounded in real health industry challenges
Myth-busting and sense-making: what works, what doesn’t, where to begin
We publish fresh episodes every week — so hit Subscribe to stay updated.
Hosted on Ausha. See ausha.co/privacy-policy for more information.
Site web du podcastÉcoutez Tech Anatomy, Symptômes ou d'autres podcasts du monde entier - avec l'app de radio.fr

Obtenez l’app radio.fr gratuite
- Ajout de radios et podcasts en favoris
- Diffusion via Wi-Fi ou Bluetooth
- Carplay & Android Auto compatibles
- Et encore plus de fonctionnalités
Tech Anatomy: Podcasts du groupe










































