supplement scam + exploratory sheet

Z / Runner 🏃🏻‍♂️

🏃🏻‍♂️ connect fx/dexa w/ episodes from peter attia.

• 🏃🏻‍♂️ no, 10/10 conviction on byoD (data). "PERSONALIZED_HEALTH_INSIGHTS" is already connected on my prototype.
• 🏃🏻‍♂️ truish.

🏃🏻‍♂️ de-re/compose daily supplements → whole foods.

• 🏃🏻‍♂️ science is settled, foolish thesis; d3+k2, creatine, coffee, omega-3 (low risk, high reward, decades of RCT data)... no need to proof circle is round. Plus, all elite & professional runners (especially in Europe are taking d + fish oil).

• 🏃🏻‍♂️ maintain large-scale, programmatic, unsupervised extraction of ai sheet is the obvious & immediate use case. Mass consumer use cases near zero, ai sheet is a power-user/b2b. Facts: 2.5 billion chatgpt messages/day, less than 10% send more >1 message per week.

• 🏃🏻‍♂️ YES! Get the data, turn cacophony (shilling fads) into euphony (evidence-based guide). Best way to dispel lies is maximum truth seeking (elon).

• 🏃🏻‍♂️ agree, open/explore/play, but i need all the data from consumerlab…

• 🏃🏻‍♂️ turn the ai agent towards government procurement (currently a black box). sam.gov is canonical source for $700b annual flows (range $10k - $10b dod)... has api. grant.gov (nsf, nih) has $300b in grants, xml+api. usaspending.gov (shows past winners, recurring vendors, patterns). If elon, palmer… are "griffting" on spend, assume there are spend that can be captured in the wellness space for generating 10x revenue.

Stephen / Basketball 🏀

• 🏀 ok / ready to share your fx login now to test?

• 🏀 ha, your pdf is scrambled and your manus scrap got many numbers wrong.

• 🏀 PERSONALIZED_HEALTH_INSIGHTS is a markdown folder?

• 🏀 definitely useful, but show me your n=1 control trials that supplements have >1% on fitness. not just boosting vitamin d in blood test by taking it. your body or your sports may not need as much vitamin d – or have homeostasis mechanisms to push and maintain vitamin d at a lower level?

• 🏀 yes, proven on fish oil but not fish oil supplement. show me the study otherwise? "settled" but what's effect size on the highest-level impact metric (fitness)?

• 🏀 also, low use of ai (sheet or other tools) since N=1 – capital N for the numbers of (daily) uses. likely N<1 for life.

• 🏀 yes, on parallelizing and extraction – but that's just good system scaling, not intelligence the best model can offer: multimodal, sentiment, reasoning..

• 🏀 or, go to other extreme: systemically expose (destroy!) the whole supplement scams. ai sheet of 100k opposing debates on each supplement: researchers, experiments, contradictions.

• 🏀 yes, but do it on the most valuable and impactful contradiction first; as you said, low risk + (placebo) high reward on supplements – so most people don't care.

🏀 which 3 more features to add – so #ai-sheet is useful for _ every day?

• for now, i want active, interactive, exploratory use cases – not just passively monitoring biometrics or representing analysis.
• 🏀 ok, let's start with fish oil – what're your prompts on them?
• 🏀 ha, i doubt we can 10x $doge or take incubuments' lunch. assuming i can get these data, what're your prompts for redirecting their trillions our way?