Meal Timing and 1 IU: When to Change Nutrition, Not Chase It
AI coaching works best when it slows nutrition edits to the pace of the adaptation you’re trying to measure.
Meal Timing and 1 IU: When to Change Nutrition, Not Chase It
AI coaching works best when it slows nutrition edits to the pace of the adaptation you’re trying to measure.
The clearest coaching pattern in the Rory Lazowski and Joe Webb logs is not that nutrition changes need to be aggressive; it’s that they need to be timed to the adaptation window. Joe saw the same insulin dose start pushing his blood sugar down enough that meal 2 had to move up by 30 minutes, and he reduced the dose by 1 IU and then again on the next high day rather than treating each meal like a new emergency. That is the mechanism in one phrase: response lag. The thesis is simple and testable: in AI-assisted coaching, the best nutrition adjustment is the smallest change that matches the timescale of the signal, not the loudest reaction to today’s check-in.
This matters because most bad nutrition decisions are not about being wrong on principle; they’re about being too fast. Coaches see a high day, a flat look, a little more hunger, a lower appetite, or a meal that suddenly lands too hard, and the temptation is to rewrite the whole setup immediately. But the actual evidence in these logs points the other way. Justin’s replies keep separating signal from noise: if the body is still in the middle of adapting, don’t stack new changes on top of old ones before the first one has had time to show its effect.
Joe’s high-day example is the cleanest case. He didn’t overeat. He didn’t blow up the plan. He simply noticed improved insulin sensitivity, a meaningfully earlier blood sugar dip, and then a need to pull meal timing forward. The important part is what he did next: he planned to reduce insulin further on the next high day, not immediately rewrite the whole day after one data point. That is disciplined adjustment timing. In practice, it prevents two common errors at once: over-correcting a temporary shift and confusing a dose problem with a meal-timing problem.
That same logic shows up in Justin’s broader nutrition periodization work. On medium days, fruit is fine in the mix; on high days, carbs are high enough and insulin elevated enough that sugar “barely matters” and fruit can help keep food volume down. The point is not that fruit is magical or that one carb source is always superior. The point is that the right choice depends on the day’s structure and what problem you’re actually solving. If the issue is food volume, fruit may be useful. If the issue is whether a medium day is “perfect,” chasing tiny carb-source differences is probably wasted motion. The rule is to let day type do the work first, then optimize details only when they actually move something.
This is the exact kind of place where AI can help—or make coaching worse. A good system should notice a trend like “meal 1 is now too much on the same dose” and suggest a narrow change: adjust the dose, or move the meal, or both in a controlled way. A bad system will see the same note and keep suggesting more changes because it is optimized for responsiveness rather than stability. Coaches already know this from prep and rebound work: the body does not respond to every day as if it were a standalone experiment. If you change nutrition faster than the physiology can reveal the effect, you create noise and then make decisions on noise.
Justin’s own language in another case is blunt enough to be useful: higher year-round doses push most people onto the downward slope of a response curve, and that point is different for everyone and unpredictable. Translate that into nutrition management and the lesson is straightforward. There is no prize for being first to intervene. There is a cost to moving too quickly, especially when the change is small and the sign is ambiguous. If a client is eating fine, training fine, and body comp is moving in the expected direction, the default should be to hold the line long enough to see what the current setup actually does.
That does not mean being passive. It means matching the size and timing of the intervention to the size and timing of the problem. Joe’s note is not a request for a complete rework. It is a localized issue: same dose, lower glucose, meal needs to come sooner. The right answer is a localized fix, then a wait period, then a reassessment. That sequence matters more than whatever exact tool you use. The same principle applies to appetite shifts. In the Rory exchange, retatrutide dropped appetite hard, and Justin immediately framed the question around what that meant for the current phase: if food is being added, consider pausing or reducing dose until prep; if leaning out is the goal, running it while body comp is moving can make sense. Again: the adjustment is tied to the phase, not to a single day’s feeling.
For coaches using AI, the practical standard should be this: do not let the machine talk you into treating every check-in as a mandate. Ask three questions before changing nutrition. First, is the signal repeatable or just new? Second, is the change localized to meal timing, appetite, volume, or total intake? Third, how long does the current setup need before the result can be seen? If the answer to the third question is “not long enough yet,” the correct move is usually to wait, not to optimize.
That rule is especially important when the signal comes from a small adjustment that is already doing its job. Joe reduced the dose by 1 IU, then had to reduce again on the next high day. That is how controlled adjustment should look: one variable, one step, one observation cycle. If instead you slash intake, add a new carb source, change meal frequency, and alter timing all at once, you lose the ability to know what helped. In coaching, the goal is not to do more. It is to learn faster without making the system unstable.
So the daily angle is simple enough to use tomorrow: change nutrition on the schedule of the adaptation, not on the schedule of your anxiety. If the meal is landing early, move the meal. If the day type changed, adjust the day type. If the signal is not yet repeated, do not stack another correction just because AI made it easy to propose one. Stable coaching beats reactive coaching, and the best nutrition timing is usually the one that waits long enough to be real.
Sources Used
modules/03-knowledge/kahunas-coaching-deep-nutrition.mdraw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/clients/joe_webb___members-rksigkykimaxwmo_t4_e8nwvbtc2j0etleutkyysads.jsonraw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/transcripts/joe_webb___members-rksigkykimaxwmo_t4_e8nwvbtc2j0etleutkyysads.mdraw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/clients/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.jsonraw/_consumed/2026-05-31/kahunas-export/2026-05-31-w13-18m/transcripts/rory_lazowski___members-c5balaovjbdoeefqmfuqdhh2tbpmfdu16lnf0tnrtmw.mdmodules/08-voice/kahunas-coaching-deep-voice.md