ai dietitian + fit³ experts

Z / Runner 🏃🏻‍♂️

🏃🏻‍♂️ half-way point into 2020s; can ai sheet usher us into era of precision nutrition? biograph.com medicine 3.0, alpha.ai eduction 3.0, when nutrition 3.0? Dietary patterns should be hyper-personalized (see comments below), reducing adipose tissue post dexa scan is different than amending biomarkers post fx_health report. Can ai (sheet) extract RCTs + epidemiological reviews? Medical doctors, dietitians, fitness influencers, and (parents) are at wars b/c they have their true "food god". Only ai can make sense of all the complex/complicated personal variables: pathologies [pre-diabetic (s), hypertension (z)] … weight loss in Q2 vs. performance opt in Q4 vs …

• Quick GPTing, I've limited knowledge of (DASH, MIND, Mediterranean) vs. elimination based (paleo, carnivore, ketogenic, vegan).

• Can ai (sheet) profile biochemical density of daily diet (omega, trace elements, mono/poly fatty acids, polyphenols, personal favorite (broccoli sprouts for sulforaphane) … I have started to pull here supplements).

• For ai (journal), we can help people migrate away from ad libitum daily diet, we need a FSD moment for daily eats; spending ~30 mins / day "thinking" about food is 30 mins i want to get back.

• 🏃🏻‍♂️ CGM did help, but too intrusive for everyday use.

• 🏃🏻‍♂️ my marathon block, 50/40/10 (higher volume of endurance training, need more glycogen substrates, and will start to integrate maurten hydrogel into weekly long run)

• 🏃🏻‍♂️ 10/10 conviction: diet debates are noisy b/c people argue "food religion" instead of use cases. I assume mark, peter, chris, walter can reach consensus on "best for what (longevity, performance)?, for whom (age, gender, ethnicity)?, and for how long?

• 🏃🏻‍♂️ $100 billion dollar problem/soultion in front of us: agree ai (sheet) can settle this if we frame and double down on our core thesis: interpretation matters than prescriptions. Ai Interpretation → 10^10 evidence-based interventions.

🏃🏻‍♂️ obvious/obvious (obv^2): there is no single "best" diet. Weight 1oss, performance, longevity, healing (medical), growth (age appropriate).

• longevity, general consensus and evidence density still favors mediterranean (guide) + DASH: diff between the 2.
• Weight loss: CR (basic math or simple i/o). Just create a sustained deficit; you, aaron and myself all benefited in past 3+ months.
• Performance: high protein + complex carbs, combined w/ training time/recovery time. Strict keto is most effective for elite/professionals.
• 🏃🏻‍♂️ 1.6 - 2x for protein, carb (100 g / hr), even-paced fueling (info for elite to sub-elite endurance sports).
• lifestyle : agree w/ aaron keto can be dangerous if not done right (but then again so are a pair of scissors). Latest episode per peter attia has great info on keto. If targeted, can treat folks w/ neuro/metabloic issues. Not good for you & I (CVD, hypertension, LDL).
• Sloppy, unintentional diet = failures; high saturated fat due to keto/paleo, no fiber (shocked that some MDs, popular influencers live by it). Micronutrient gaps (personally guilty, suspect Li may have similar profile) b/c our diets are too restrictive.
• 🏃🏻‍♂️ i suspect zinc, iron, magnesium… much of our modern food lack these nutrients b/c the soils are so depleted.
• Age matters: for our kids, no evidence/logical chain to impose extreme restrictions. Do follow: low sugar, limited fast food, no hyperprocessed).
• Middle age (us): protein + high fiber + complex carbs are non-negotiable.
• 🏃🏻‍♂️ legume + nuts, but the specific should be calculated per baseline metabolic.

Stephen / Basketball 🏀

• 🏀 ai dietitian is valuable, but humans in the loop: social eating at fit³. i find myself overeating when tired, excited, mindless.
• 🏀 too awkward to wear it during ball games.

🏀 the more valuable debate ai (sheet) can settle is: daily intake % of carb vs protein vs fat – for n=1 fitness: basketball performance, muscle gain, fat loss, age 47..

• mark fadil (smi) strongly argues 60-20-20. he thinks it should not be controversial or ambigious, fully backed by research of chris gardener (stanford) and walter willet (harvard).

• peter attia 30-50-20 (?), rhonda patrick 30-35-35, david sinclair 15-45-40, valter longo (usc) 60-10-30.

• again, does it matter? what's the error bound? i may argue that, simply: min 15% for each will be fine for most.

• 🏀 mark says 60-20-20 for me, but also those of similar types and goals – if not most population.

• 🏀 yes, data > analytics > interpretation > debate > reasoning > ..

• 🏀 not strictly: microbiome strongly affects absorption, hormones fully dictates assimilations.

• 🏀 but how high and at what %?

• 🏀 for 170 pounds

• 🏀 what gaps in particular?

• 🏀 my zinc: 70mcg/dL.

• 🏀 how much and what variety of fiber?

• 🏀 did stool / microbiome analysis?

🏀 mark really liked fit³ and may even help. the top 3 values: food + train + tech.

• hourly but social – food & drink: offset the space + ops cost; most importantly, visitors stay 1+ hours and socialize over chef counters or farm tables – no phone allowed on premise upon entrance. mark and his wife are a strong believer in simple but nutritional food; they may help bootstrapping that.
• daily but 10% – train & stretch: ..
• weekly but 10x, tech & expert: ai, biomarkers, $10k - 100k devices (cold-hot contrast / shock wave / .. therapy). 3-hour onsite: speak to 100 people, hands-on assessment + treatment to 20. tracy is interested, too: her 2026 goal is organizing weekend or even weeklong retreats. we can convince mark to have share / split his new clinic space (close to cal ave cal train).
• sadly, to start lean and do the best by day _, we should pick 2 (or even 1) out of 3.
• mark likes the small, local gyms for social: axis in menlo park, fit and gains in palo alto.

🏀 "which articles and newsletters do you (wish to) read these days? daily feed of knowledge (not news) is the strongest consumer use of gen ai. besides prioritization (based on long-term goals), summarization (based on existing contexts), recommendations (based on active feedback) – what top features or initial prompts will be immediately useful?"

🏀 "share your top 10 receipes and your last 10 shopping cart receipts? (grocery) shopping is the most profitable use of (gen) ai. yes, ai (sheet) should track nutricient goals / calorie intakes / home inventory – as your concierge, if not coach or chef. dealing with instacart shoppers or item sustitutions should be a strong use (insteaf of limitations or frictions) of agents."

🏀 "how many times do you eat out monthly? ai with photo understanding (instagram) + menu parsing (yelp) + nutricient logging (google) is the most use by age 20-40. you knew i was most obessed with foursquare mayors, restaurant ranking, michelin recipes.."