full-time fit³ + sheet prototype

Z / Runner 🏃🏻‍♂️

🏃🏻‍♂️ demo ai sheet that continuously x-references: huberman lab for mind or consumer lab for supplements; show the ai sheet is capable of doing analysis not just data aggregation/automated scrapper.

• 🏃🏻‍♂️ 2025 is ending, time to be man in the arena.

• 🏃🏻‍♂️ conviction means tradeoff… dialing back on run and putting research into development.

• 🏃🏻‍♂️ gating & strong signals: tgi in weekly cadence, prototype app in hands, ai layer in development, and capital committed. I want to bring 5+ likeminded angels into the cafe or seek out venture (although no consensus at the moment).

• 🏃🏻‍♂️ serious about making tgi digital; otherwise, we will not control our own distribution destiny.

• 🏃🏻‍♂️ we need repeat what Bryan Johnson, Peter Attia, Huberman Lab, All-In taught us, control the narrative… podcast not for self-gratification but for growth.

• 🏃🏻‍♂️ options for sublet spaces among xoogler startups?

• 🏃🏻‍♂️ can we sublet an adjacent space near a fitness gym?

• 🏃🏻‍♂️ start w/ borrowed infra in Q1.

• 🏃🏻‍♂️ start "biographing" people in the longevity space.

• 🏃🏻‍♂️ how much do you want to scale this? If you want 100 customers in the next 30 days, highly suggest upwork & fiverr.

🏃🏻‍♂️ completely diff framing. Goal is not crawling/ranking/better db/prettier UI… agree these are table stakes and low value commodity in 2025. The real unlock is buzzing around "agentic interpretation" at the moment of decision/action. Meaning, biographing top n longevity researchers/practioners is not about creating profiles for people to read later (chatgpt/gemini overserving already). It is creating a living system that can instantly comment, feedback. Let's flow thru this:

• Continuously ingest what these people say, publish, recommend

• Track mounting evidence, contractions, updates and reversals over time (no one does this, market vacuum)

• Map their positions, protocols to each user's input (i took these pills, follow this exercise routine, doing keto, using this device…)

• The ai sheet (assume this is NOT the final product you're launching) runs a synthetic peer review among all the people

• Who explicitly supports/contradict, who warns, assumptions made, link to evidence.

• Completely different than "yelp" reviews, it is contextualized expert consensus/disagreements, on demand

• Tl;dr: agentic reviews > yelp reviews, citations > opinions, ai interpretation > data aggregation

• 🏃🏻‍♂️ social eating, fit^3 couple nicely w/ the daily habit; the core theme here is how to leverage the 100 ai agents… what are some of the alternative use cases you have in mind?

• 🏃🏻‍♂️ do > know.

• 🏃🏻‍♂️ solution, previously we highlighted the problem. If we're talking about using the ai sheet, then we can't assume people are hyper-motivated to spend hours / week using chatgpt.

• 🏃🏻‍♂️ no twitter, but at least 2-3 hours reading books, pa shownotes, podcast… they all say different things over the course of years… hence the entire industry is mired in fads and double-talk. DYOR is not a solution…

• 🏃🏻‍♂️ if longevity != evidence, then != trust. Vibes, charisma can only go so far, research consensus is the only shared anchor, especially when attention spans are collapsing & we are always skepticical. Case in point, folks like peter attia, david sinclair are not walys right, but they do anchor their arguments in journals, trials and primary sources.

• 🏃🏻‍♂️ assuming, not getting the point across, citations are NOT the problem/solution. No one clicks on citations (i rarely check every single gemini links). People need contextual commentary, visible revision… for example, when i shared keto diet maybe helpful, ..

• 🏃🏻‍♂️ goals != process… personal fitness and product development are just part of the process to unlock 10x. In reality, have to check-off certain milestones… motivation of 10x is great, but enjoying the pain is what i come to accept this year.

• 🏃🏻‍♂️ how many people know the 99% and can't or won't do the basics?

• 🏃🏻‍♂️ my question is how to turn knowledge (100 prompts) → social commentary? Not just hyperlinks to papers or books.

• 🏃🏻‍♂️ that's my point, citation is not about accuracy, it is about popularity. People trust blueprint protocol b/c BJ got 1 million yt subscribers. His hook is that he has a team of doctors testing his day/day. If the 100 agents can constantly be on the lookout, and pushing/nudging you daily, then that's worth something.

• 🏃🏻‍♂️ agree, these are diff points. Being correct/right != following… the market is full of noise… sleep, fitness, diet are still vibing on social.

• 🏃🏻‍♂️ hence nudge > nag.

• 🏃🏻‍♂️ for clarity, i don't need nudge to run, but i do check my LT1/LT2, heart rate and pace / mile to nudge for better performance (the watch literally shows perceived effort to push for lower intensity vs. higher intensity). tgi, you did a great job of telling folks who is coming (vcs, ceos…), so people find their own motivation/reaons to show up. But i get the point. No need to split the hair. No one likes to be nagged.

• 🏃🏻‍♂️ good question, curious about parallels between the 2 sports in terms of training (science based) and what's considered elite. For example, when you train, do you ever read advanced books, fundamentals of set plays?

• 🏃🏻‍♂️ we're cross-talking, yes, reading books won't make you better. I didn't just read to run faster and longer :-). Have you looked into science? For example, Daryl morey on 3pt, layups vs. mid-range, favor spacing over speed…

• 🏃🏻‍♂️ to counter, running clubs are huge, i just prefer to run and think solo.

• 🏃🏻‍♂️ On the topic of LT1-3, it is the modern, science based training… less subjective method to override the brain. We have the tools to measure the pain and understand the energy consumption per lactate threshold.

• 🏃🏻‍♂️ yes, i can see your fitness 10x in the past 6 monthsish time, but my question is the skill improvements and play making. words have meaning, nerds vs. jocks vs. bball savant (LBJ, rondo…).

• 🏃🏻‍♂️ basketball IQ is a real thing… i see why joker (advanced stats are amazing) and luka are so good, slow as a mule, but can out pace/play most.

• 🏃🏻‍♂️ xmas read. 🎄

Stephen / Basketball 🏀

🏀 exactly.

🏀 your full-time commitment, fit³ (三人行): "sleep at home, work at screen – but fitness together".

• 🏀 i don't worry about your pains or failures – those are guaranteed – but your choice of fight to win.
• 🏀 but time realistic? 24 hours - 7 sleep - 3 ira.yo - 2 run.train - 2 eat.drive - .. = 0 for ai products, health research, 3:30 marathon.
• 🏀 milestones by mar 5th thu (day 100)? own space with a 3-year lease, daily programs with _ paying customers.
• 🏀 if once a week and borrowed space, no way to do more than speaker series (like tgi-ml). so, no equipements or barely venture incubation before q3?
• 🏀 digital podcast / content / app?
• 🏀 yes, as go-to-market and distribution scaling – but not a product itself + assuming the product (social fitness in physical space) is scalable.
• 🏀 if daily and owned space, 5 angels * $100k not enough for lease deposit + rent + front desk (10-hour pay).
• 🏀 not just a few desk + whiteboard, but also a 10-people room with dumbbells?
• 🏀 or buy out a gym for fixed hours: 9-11am or 2-4pm or 8-10pm.

🏀 "my prototype of ai spreadsheet can now run 100 prompts in batch or parallel, paste to store images quickly without google files api, crawl a full site behind paywall, full control via keyboard shortcuts.. – soon, also: control remote browsers, retrieve via vector stores, zapier-like integrations. what (ai) tasks do you (wish to) run daily / weekly that my sheet can automate?"

• 🏀 no longer necessary: prompt "top 5 on basketball performance + ai fitness among ..". i was all obsessed with crawling / ranking / enriching everything at spotsetter and later tgi guests.

• 🏀 who among friends will pay me $100 to automate their ai tasks?

• 🏀 we have gone through this argument many times already; see "meta protocol" below. gemini 3 with proper prompt + context already gives 90% answer. any of the top questions (calorie deficiency) you care only make 10% in your fitness; any expert answers make only 10% difference – so 0.1% impact for your life. let's focus on the 10x: social eating and fit³?

• 🏀 exactly the right question: i can't find use cases that i care about / that makes >10% impact in your life. of course plenty tasks on fiverr / upwork for marketing campaigns, financial modeling – but they are low-level analytics or boring cron jobs, rather than using intelligence of any ai models.

• 🏀 code > uses. ai sheets are the right infra. i can keep building features that over time it will be useful to you and many others by default.

• 🏀 how many hours of (social) research are you still doing on twitter?

• 🏀 true, but you assume their / any answers matter >1%. sadly / gladly, that means you can stop reading but start building.

• 🏀 that's just citation dropping; much as we did name dropping when we were young.

• 🏀 so just ai product with better ux will fix everything; no science research or more info from web are needed anymore.

• 🏀 you want to be right, or you want to make money, or you want to be fit? 1% difference for the first (science papers), 10% (product differentiation) for the second, vs 10x impact (simple rules) for the third – in your life.

• 🏀 doing the basic and well-known is 99% than the latest 1%-better answers: turn off your brian to sleep 8 hours, chew more to enjoy the plain greens, train 2 hours on what you love.. plus social with 5 friends (on not necessarily about fitness).

• 🏀 99% around us. but 10x more or accurate knowledge is the opposite of what may change them; social + physical can.

• 🏀 those are two separate points: 1. if zi is expert and correct on a particular fitness thesis, the crowd will follow. 2. constant nagging is good coaching.

• 🏀 the former is valid but monetizing yourself as a superficial product. per "external motivation" below, i am against nagging.

• 🏀 no nudge needed for your running or come to drink at tgi.

• 🏀 is obsessing lt1/lt2 the right / common way to train for 3:30 marathons? i think and dream a lot (LOTS!) about basketball moves, but not the scores.

• 🏀 i did, and had a full substack post with 100+ references on basketball training. 0.1% impact compared to playing 3 times a week + 90-minute weekly training with jeramey.

• 🏀 yes, love "thinking basketball" and can consume 10x more content like that.

• 🏀 indeed, the social (10 people) of basketball vs the solidity of running.

• 🏀 counter to your counter, i love shooting solo 30 mins every 6pm; never did or liked it before.

• 🏀 nerds vs jocks. i am pleasantly surprised i can play 2+ hours now without pain. no science or motivations or measurements needed.

• 🏀 absolutely, my top 2026 goal: neuro training of spacing, pacing, acting.. in games.

• 🏀 ha, they all drink too much, too. indeed, that (neuro ball) may be the science and research i need to do + unlock. will train my ai (sheet) on this playlist.