Peter Simpson has been at Dexcom for over 24 years, back when it was a 
small startup, and he says he's more excited about the technology now than 
at any point in that run. In this bonus tech episode, Dexcom's CTO walks us 
through the upcoming G8 sensor: about 50% smaller, plus a new in-use 
calibration system that checks itself every five minutes to account for the 
fact that every body is a little different. The goal isn't just a better 
average accuracy number. It's fewer of those nights where your sensor and 
your fingerstick seem to be describing two different people.

We also dig into the landmark CONNECT study, just published in The New 
England Journal of Medicine, which found CGM use lowered A1C by 0.9% more 
than routine care for people with type 2 diabetes not on insulin. Peter 
shares a surprising finding about people already on GLP-1s, and what 
Dexcom's new State of Type 2 report says is really standing between people 
and better care: awareness, education, and access.

Near the end, I ask Peter about the recent G7 app terms and conditions 
update that briefly kept some users from seeing their glucose data. He 
shares how future updates will roll out and answers questions about 
metadata and AI. There's more I wish I'd had time to ask, and if I get 
another conversation with Dexcom, you'll hear it here first.

Chapters

01:22 Meet Dexcom CTO Peter Simpson

02:01 Why innovation is speeding up after 24 years

03:21 G8 preview: 50% smaller, more accurate

04:15 How factory calibration actually works

05:05 Real-time calibration that adapts to your body

06:03 MARD explained: fewer bad sensor outliers

07:15 Type 2 barriers: health and tech literacy

07:48 Inside Dexcom's State of Type 2 report

09:08 CONNECT study: CGM results without insulin

10:25 GLP-1s plus CGM: the biggest impact

11:38 Addressing the G7 terms and conditions

12:12 "We made an error": what changes now

13:15 Metadata, pump partners, and AI explained*

15:07 What's next for type 1 and type 2

Resources:
The CONNECT Study published in The New England Journal of Medicine
Dexcom

Peter Simpson has been at Dexcom for over 24 years, back when it was a small startup, and he says he's more excited about the technology now than at any point in that run. In this bonus tech episode, Dexcom's CTO walks us through the upcoming G8 sensor: about 50% smaller, plus a new in-use calibration system that checks itself every five minutes to account for the fact that every body is a little different. The goal isn't just a better average accuracy number. It's fewer of those nights where your sensor and your fingerstick seem to be describing two different people. We also dig into the landmark CONNECT study, just published in The New England Journal of Medicine, which found CGM use lowered A1C by 0.9% more than routine care for people with type 2 diabetes not on insulin. Peter shares a surprising finding about people already on GLP-1s, and what Dexcom's new State of Type 2 report says is really standing between people and better care: awareness, education, and access. Near the end, I ask Peter about the recent G7 app terms and conditions update that briefly kept some users from seeing their glucose data. He shares how future updates will roll out and answers questions about metadata and AI. There's more I wish I'd had time to ask, and if I get another conversation with Dexcom, you'll hear it here first. Chapters 01:22 Meet Dexcom CTO Peter Simpson 02:01 Why innovation is speeding up after 24 years 03:21 G8 preview: 50% smaller, more accurate 04:15 How factory calibration actually works 05:05 Real-time calibration that adapts to your body 06:03 MARD explained: fewer bad sensor outliers 07:15 Type 2 barriers: health and tech literacy 07:48 Inside Dexcom's State of Type 2 report 09:08 CONNECT study: CGM results without insulin 10:25 GLP-1s plus CGM: the biggest impact 11:38 Addressing the G7 terms and conditions 12:12 "We made an error": what changes now 13:15 Metadata, pump partners, and AI explained* 15:07 What's next for type 1 and type 2 Resources: The CONNECT Study published in The New England Journal of Medicine Dexcom

Episode 368 - CTO Peter Simpson on the Future of Dexcom CGMs

When Aurelian Briner's wife was diagnosed with type 1 diabetes about twelve 
years ago, the hardest questions kept coming back to the plate: how many 
carbs are in this or how big is this portion? Aurelian, a product manager 
with a business background and an entrepreneurial itch, decided that was a 
problem worth solving. In 2017 he co-founded Snaq, and today the app has 
more than 300,000 registered users and around 20,000 monthly active users 
around the world.

In this episode, Rob and Aurelian trace Snaq's evolution from a photo-based 
carb counter, to a platform that connects meals with CGM, pump and activity 
data, to the AI Nutritionist, a 24/7 companion built to help people reach 
their nutrition goals. They also dig into what it takes to bring AI into 
healthcare responsibly: fast product iteration on one side, slow and 
careful clinical validation on the other, and why both matter.

Rob makes his case: he doesn't want your vibe-coded diabetes app. Not 
because building your own tools isn't exciting (it is), but because there's 
a huge gap between a weekend prototype and a product that's used, trusted 
and clinically backed. Aurelian breaks down Snaq's moat (users, data and 
clinical validation), shares an MIT finding that app releases roughly 
tripled while engagement stayed flat, and explains why building got easier 
for everyone while marketing got harder.

Snaq's published RCT showing a 6.6% time-in-range improvement on top of AID 
systems, upcoming research on the AI Nutritionist versus standard of care, 
a US study on T1D pregnancy (one Aurelian's family lived firsthand), the 
shift from reactive apps to proactive AI agents, and the founder lesson 
Aurelian would give his younger self: it always takes longer than you 
think, so make sure you care enough to stay.

Chapters:

00:00 Meet Aurelian Briner, CEO of Snaq

01:10 Why AI health tech changed in 2026

02:20 A wife's T1D diagnosis sparks Snaq

03:42 From carb counter to AI companion

05:10 Late-night questions and clinically safe AI

06:20 Build fast, iterate, then prove it clinically

08:09 Why insulin pumps lagged behind iPhones

09:26 Rob's case against vibe-coded diabetes apps

11:40 Snaq's moat: users, data, clinical validation

14:05 Triple the apps, flat engagement: MIT study

17:13 What users say about the AI Nutritionist

19:49 New research: 6.6% more time in range

21:05 Studying T1D pregnancy while living it

22:14 From reactive tools to proactive companions

24:50 Ten years in: lessons for founders

Resources:
Snaq.ai

When Aurelian Briner's wife was diagnosed with type 1 diabetes about twelve years ago, the hardest questions kept coming back to the plate: how many carbs are in this or how big is this portion? Aurelian, a product manager with a business background and an entrepreneurial itch, decided that was a problem worth solving. In 2017 he co-founded Snaq, and today the app has more than 300,000 registered users and around 20,000 monthly active users around the world. In this episode, Rob and Aurelian trace Snaq's evolution from a photo-based carb counter, to a platform that connects meals with CGM, pump and activity data, to the AI Nutritionist, a 24/7 companion built to help people reach their nutrition goals. They also dig into what it takes to bring AI into healthcare responsibly: fast product iteration on one side, slow and careful clinical validation on the other, and why both matter. Rob makes his case: he doesn't want your vibe-coded diabetes app. Not because building your own tools isn't exciting (it is), but because there's a huge gap between a weekend prototype and a product that's used, trusted and clinically backed. Aurelian breaks down Snaq's moat (users, data and clinical validation), shares an MIT finding that app releases roughly tripled while engagement stayed flat, and explains why building got easier for everyone while marketing got harder. Snaq's published RCT showing a 6.6% time-in-range improvement on top of AID systems, upcoming research on the AI Nutritionist versus standard of care, a US study on T1D pregnancy (one Aurelian's family lived firsthand), the shift from reactive apps to proactive AI agents, and the founder lesson Aurelian would give his younger self: it always takes longer than you think, so make sure you care enough to stay. Chapters: 00:00 Meet Aurelian Briner, CEO of Snaq 01:10 Why AI health tech changed in 2026 02:20 A wife's T1D diagnosis sparks Snaq 03:42 From carb counter to AI companion 05:10 Late-night questions and clinically safe AI 06:20 Build fast, iterate, then prove it clinically 08:09 Why insulin pumps lagged behind iPhones 09:26 Rob's case against vibe-coded diabetes apps 11:40 Snaq's moat: users, data, clinical validation 14:05 Triple the apps, flat engagement: MIT study 17:13 What users say about the AI Nutritionist 19:49 New research: 6.6% more time in range 21:05 Studying T1D pregnancy while living it 22:14 From reactive tools to proactive companions 24:50 Ten years in: lessons for founders Resources: Snaq.ai

Episode 365 - How His Wife's T1D Diagnosis Led Aurelian Briner to Build Snaq

Episode 173 - Grace Choi - T1D, NYC, Diabetes & Relationships with people and food, Mental Health, fitness.

Episode 172 - Bridget McNulty, T1D and Author of the Grief Handbook

Episode 171 - Saira Gallo T1D, Mother of T1Ds, Married to a T1D, Diabetes Blogger and works in Diabetes!

Episode 170 - Lina Kazii 2015 Miss Rwanda, Advocate, Influencer, Entrepreneur

Episode 169 - New format podcast! Special guest Noor Al Ramahi on Diabetes and IVF, Palestine and more

Episode 168 - Golf and SmartGuard Auto Mode with Medtronic Diabetes Ambassador Kris Leeper

Ramadan with Diabetes & April 21 Round Up with Eritrea and Rob

Episode 167 - Diabetes in Jamaica with Alexia Beckford

Episode 166 - Hiking My Feelings with Sydney Williams

March 21 Round Up with Eritrea and Rob

<p>Hosted by T1D Rob Howe, Diabetics Doing Things tells amazing stories of people with diabetes from across the globe, digging deep into everything it takes to Live Well with Diabetes and sharing exciting adventures along the way.</p>

Hosted by T1D Rob Howe, Diabetics Doing Things tells amazing stories of people with diabetes from across the globe, digging deep into everything it takes to Live Well with Diabetes and sharing exciting adventures along the way.

Share Diabetics Doing Things Podcast

Sign up to save your podcasts

Diabetics Doing Things Podcast

Diabetics Doing Things Podcast episodes

About Diabetics Doing Things Podcast

More shows like Diabetics Doing Things Podcast

Stuff You Should Know

Juicebox Podcast: Type 1 Diabetes

Diabetes Connections | Type 1 Diabetes

Reclaim Your Rise: Type 1 Diabetes with Lauren Bongiorno

The Economics of Everyday Things