Coach Roster Scale and the 2 mg Retatrutide Lesson

Coach Roster Scale and the 2 mg Retatrutide Lesson

Why AI should widen a coach’s reach without replacing appetite, load, and judgment calls

The clearest field finding in the KB is mundane and useful: a 2 mg retatrutide dose produced “no appetite whatsoever” across low-carb days, with fatigue noted alongside it. The mechanism is appetite suppression, and the practical thesis is sharper than the hype around “AI coaching” usually allows: software should scale observation and pattern-spotting across a roster, but it cannot be allowed to make the judgment call that decides whether a body is ready to push food, hold dose, or change course.

That is the real leverage question for coaches. Not whether AI can replace expertise. Whether it can make one coach more available to more athletes without flattening the decisions that actually matter.

Justin Harris’ messaging gives the most honest version of that workflow. He tried retatrutide himself “so I have better first hand experience for clients,” then immediately moved to the coaching implication: if body composition is moving in the right direction, “run with it and lean out a bit while it’s easy.” That is not a pharma victory lap. It is a coach using a new input to improve decision quality. The key move is not the compound; it is the judgment around when to use it, why, and for whom.

That distinction matters because the same tool can create opposite constraints depending on phase. In a gaining block, appetite suppression is not a feature. It is a bottleneck. In a cutting phase, appetite suppression can make adherence easier, at least for the people who were previously fighting hunger. The KB language is explicit about this tension: Harris said he was “not sure” he liked the idea of forcing appetite lower, and he was holding his opinion on “helpful in gaining” until he had more data. That skepticism is the standard coaches should adopt for every AI feature that promises to “optimize” something before the actual phase demand is defined.

AI coaching inherits the same problem. It is easy to build a system that produces more check-ins, more reminders, more summaries, and more enthusiasm. That is throughput, not leverage. Coach leverage happens when the tool helps you see the right signal faster, across more athletes, without diluting the judgment that tells you what the signal means. The wrong use case is a chatbot that speaks with confidence on every variable. The right use case is a system that logs, surfaces, compares, and escalates while keeping the coach in charge of the call.

The KB has a second useful clue: Justin’s off-season philosophy is built around teaching the body to “digest and assimilate a massive amount of clean food.” He frames that as a pathway to higher food tolerance, better metabolic handling, and better prep outcomes later. Whether you agree with every part of that phrasing or not, the coaching structure is clear. The off-season is not just about adding tissue. It is a skill phase. Food intake, meal timing, and digestive capacity are being trained as a capacity that later changes the options available in prep.

That is exactly where roster-scale coaching can go wrong if it becomes too automated. If AI turns the job into generic check-in collection, it will miss the phase-specific logic: some athletes need more food tolerance, some need appetite control, some need tighter insulin coordination, and some need more restraint around novelty. A model can remind you that a client reported fatigue. It cannot decide whether fatigue is a nuisance, a cost worth paying, or the reason to pause a lever. That decision depends on context the model does not own.

The same applies to insulin sensitivity notes in the KB. Joe Webb reported that on a high day, the same insulin dose that had been fine the week before now dropped blood sugar more noticeably, forcing him to bring meals closer together and reduce the dose on the next high day. Justin’s response was not to chase the number blindly. It was to observe the change and adjust the plan around what actually happened. That is the kind of nuance AI often misses when it overweights the latest datapoint and underweights the pattern across weeks.

So what should a coach actually automate?

Automate collection. Automate organization. Automate flagging. Automate the boring parts that make a roster hard to manage: missed check-ins, repeated meal timing drift, unusual appetite changes, sudden fatigue, bodyweight trends, and response notes that need to be compared against last week rather than read in isolation. Use AI to expand your bandwidth so you can spend your real energy on interpretation.

What should stay human?

Thresholds, phase decisions, and tradeoffs. Whether an athlete should keep running a tool that lowers appetite. Whether a food increase is pushing adaptation or just creating noise. Whether a change in dosing pattern is a one-off or the start of a new response curve. Whether the current strategy matches the actual phase goal. Those are not data-entry tasks. They are coaching judgments.

That last point is the real roster-scale advantage. A coach with AI can manage more athletes because the system compresses clerical load and surfaces exceptions early. But scale only helps if the coach refuses to let the software become the author of the plan. AI should narrow the distance between signal and review, not widen the distance between review and decision.

If you want a falsifiable standard, use this: AI is useful to a coaching roster when it makes the coach faster at noticing phase-relevant exceptions without increasing the number of bad default decisions. If the tool increases volume of communication but not quality of judgment, it is just more noise at scale.

That is the useful lesson from the KB sources. Retatrutide gave one athlete a dramatic appetite drop at 2 mg. Justin used that input to inform a phase decision, not to surrender judgment to the tool or the compound. Good AI coaching should work the same way: more reach, more structure, more pattern recognition, and the same hard human responsibility at the center.

Sources Used:

• raw/_consumed/2026-05-26/troponiniq_kb.md
• raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/transcripts/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.md
• raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/clients/joe_webb___members-rksigkykimaxwmo_t4_e8nwvbtc2j0etleutkyysads.json
• raw/Justin_TT1.txt

Sources Used

• /Users/justinharris/TroponinIQ/kb/supertrop/raw/_consumed/2026-05-26/troponiniq_kb.md
• /Users/justinharris/TroponinIQ/kb/supertrop/raw/Justin_TT1.txt
• /Users/justinharris/TroponinIQ/kb/supertrop/raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/transcripts/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.md
• /Users/justinharris/TroponinIQ/kb/supertrop/raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/clients/joe_webb___members-rksigkykimaxwmo_t4_e8nwvbtc2j0etleutkyysads.json