Rory Lazowski 2 mg Retatrutide and the Coach Leverage Problem
Why AI can scale observation, but judgment still has to stay human
Rory Lazowski 2 mg Retatrutide and the Coach Leverage Problem
Why AI can scale observation, but judgment still has to stay human
The 2 mg retatrutide test immediately reduced appetite and added fatigue in Justin Harris’s firsthand trial, with no appetite on low-carb days and a clear plan to pause or reduce it if food intake needed to climb. The mechanism is simple appetite suppression. That matters because the real coaching thesis is not that AI should decide more; it is that AI should let a coach observe more athletes with less overhead while keeping the call on load, timing, and tradeoffs firmly inside human judgment.
That is the leverage question in 2026. AI coaching is getting pitched as if it will replace coaching brainpower. The evidence in the KB points in a narrower direction: it can help a coach collect more signals, compare more cases, and keep more athletes in play at once, but the high-value decisions remain judgment calls built on context. If you care about roster scale, the winning move is not outsourcing judgment. It is separating pattern collection from decision ownership.
The useful part of AI is not the headline
TroponinIQ already lives in the middle of this shift. It is not just a chatbot bolted onto a bodybuilding brand; the platform is presented as a coaching ecosystem with structured courses, progress tracking, and 24/7 access, built from Justin Harris’s coaching experience and the broader body of coaching knowledge in the KB. The practical promise is obvious: a coach can spread standardized information, keep athletes moving, and reduce the amount of repetitive explanation that usually eats the day.
That alone is real leverage. Most coaches do not lose time on complex programming decisions. They lose time on repeatable tasks: explaining why the same food change is being made again, reminding an athlete what data matters, checking whether the report is actually actionable, and translating theory into the next seven days. AI helps most when it absorbs that repetition.
But the KB also gives the counterweight: Justin’s own voice stays skeptical of anything that sounds clean in theory but messy in practice. In the retatrutide exchange with Rory Lazowski, he does not pretend the appetite suppression is neutral or universally helpful. He notes that it is lowering appetite, that he is not sure he likes forcing appetite lower, and that he is holding his opinion on “helpful in gaining” until he has more data. Then he makes the decision that matters operationally: if the body comp is moving in the right direction, run with it and lean out while it is easy.
That is the model coaches should keep. Use the tool to widen observation. Keep the decision tied to what the athlete is actually doing.
Scale comes from standardization; quality comes from exceptions
There are two kinds of leverage in coaching.
The first is standardization. This is the part AI can scale well. If you have 20, 50, or 100 athletes, you need a consistent way to gather updates, flag missing data, summarize changes, and surface obvious problems. The more your workflow is built around structured inputs, the more a system can save time without changing the actual coaching philosophy.
The second is exception handling. This is where judgment lives. A coach has to know when the body is responding faster than expected, when food tolerance is collapsing, when a trend line is lying because the athlete’s compliance changed, or when a seemingly good intervention is actually crowding out something more important. AI cannot own those tradeoffs unless you are willing to let it flatten the person into a template.
Justin’s remarks around food intake and metabolic adaptation in the off-season point in the same direction. He frames the off-season as teaching the body to digest and assimilate a massive amount of clean food, then using that capacity to support growth and make contest prep easier later. Whether you agree with every phrasing or not, the operational point is clear: the coaching target is not a static macro number. It is the athlete’s ability to tolerate, process, and adapt to a rising workload of food and training stress over time.
That kind of target benefits from AI-generated organization, but it does not become an AI decision. A model can remind you that an athlete’s appetite, fatigue, and body comp trend changed. It cannot tell you whether the correct move is to hold, increase, reduce, or wait another week unless you let it override the very thing the coach is paid to do.
What roster-scale coaching should automate
If you are coaching a roster, automate the boring parts first.
- Intake collection: make the athlete report the same key variables every time.
- Trend extraction: summarize changes in bodyweight, appetite, fatigue, training performance, and adherence.
- Triage: flag who needs a reply now versus who needs no change.
- Education: answer the repeated “why” questions with a consistent framework.
- History retrieval: let the coach see what happened last time a similar pattern showed up.
Those are all leverage multipliers because they preserve the coach’s attention for decisions that actually change outcomes.
What should not be automated is the part where the coach decides whether a signal matters. The KB is full of reminders that context wins. In the Rory exchange, Justin is willing to test retatrutide himself to understand it firsthand for clients. That is not a gimmick. It is a reminder that tools should be judged in use, not in marketing copy. He is also explicit that he is “holding opinion until more data.” That is exactly the right stance for coaches using AI: let the system scale note-taking and pattern visibility, but keep your skepticism intact until you have seen enough cases to know where the tool fails.
The preserve-judgment rule
There is a trap in every automation story: once a system saves time, people start asking it to save thinking too. That is where coaching gets worse.
Preserving judgment means three things:
- The coach defines the decision criteria.
- The AI gathers and organizes the inputs.
- The coach makes the call and owns the consequence.
That order matters. It keeps the platform from becoming a replacement for experience. It also protects against the most common failure mode in high-volume coaching: getting seduced by tidy summaries that hide the one variable that actually changed.
The better your AI system gets, the more dangerous it becomes to treat it as neutral authority. Use it to see more, not to believe more. Use it to scale your roster, not to dilute your standards. Use it to reduce labor, not to erase the coach’s role in deciding what matters.
That is the falsifiable thesis here: AI will increase coach leverage only when it expands observation faster than it expands automation of judgment. If your system starts making the calls for you, you have not scaled coaching. You have outsourced it.
Bottom line
The strongest use of AI in fitness coaching is not prediction. It is bandwidth.
Bandwidth lets a coach manage more athletes, review more data, and keep better records without drowning in admin. But roster scale only turns into better outcomes if the coach stays the final filter for context, tradeoff, and timing. In the KB sources, the pattern is consistent: the best operators test tools firsthand, track what happens, and keep their opinion provisional until the data is strong enough to justify a real change.
That is the standard worth protecting.
Sources Used
- raw/_consumed/2026-06-02/_GRAS/gras_strategy_training.md
- raw/_consumed/2026-05-26/troponiniq_kb.md
- raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/clients/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.json
- raw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/transcripts/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.md
- raw/Justin_TT1.txt