An AI nutritionist,
available right now.

Programme design is not where most people fail. Adherence is. A theoretically optimal six-day split that collides with a real job is worth less than a mediocre three-day one that survives to month six.

The Nutrition and Fitness Copilot starts from constraints rather than from ideals - the days you have, the equipment you can reach, the food you will actually cook - because that is the variable that determines the outcome.

Fitness and nutrition carry more confident nonsense per square inch than almost any other field, because there is money in supplements and none in saying that sleep, protein and consistency do most of the work.

What it will do is tell you what the evidence actually supports, how strong it is, and where honest disagreement exists. Sometimes the answer is that it does not matter much, which is unsatisfying and true.

Medical conditions, medications and pregnancy change what is safe, and it will tell you to see a doctor rather than guess. Rehabilitation needs someone watching how you move.

And on disordered eating there is a hard rule: it will not build restrictive plans or engage with the framing, and it will point toward specialist help. That is not caution, it is the only responsible behaviour available.

The fitness industry is built on the marginal and quiet about the fundamental, because the fundamental is not sellable. Sleep and protein account for a large share of the results people attribute to programming and supplementation.

Sleep is the more neglected of the two. Restricted sleep impairs recovery, reduces training performance, and shifts appetite regulation in a direction that makes eating well harder. Someone training hard on six hours is fighting their own physiology and usually blaming their programme.

Protein is the more misunderstood. The requirement for someone training regularly is substantially higher than general dietary guidance suggests, and most people who believe they eat plenty are meaningfully under when they log it accurately for a week. Distribution across the day matters less than total, and total is where the gap usually is.

Neither is complicated and neither is interesting, which is precisely why supplement marketing exists in the space they leave. Nothing on a shelf compensates for six hours of sleep, and no programme design compensates for chronically insufficient protein.

Fix those two before optimising anything else. The difference is usually larger than any change to set and rep scheme, and it costs nothing except attention.

The mechanism behind strength and muscle gain is not complicated: the body adapts to demands that exceed what it is currently accustomed to, so the demand has to increase over time. That is progressive overload, and almost every stalled programme has stopped applying it.

Overload does not only mean more weight. More reps at the same weight, more sets, better range of motion, shorter rest, or the same work performed with better control all constitute increased demand. Fixating on load alone is why people stall and then plateau for months on the same three lifts.

The second reason programmes stall is that nothing is recorded. Progression requires knowing what last week was, and a surprising number of people train consistently for a year without a log. Without one, the weight on the bar drifts by feel, which on a tired day means backwards.

The third is that recovery has not scaled with the training. Adding volume without adding sleep and food produces accumulating fatigue that reads as a plateau and is actually under-recovery. A deload week is not a rest from progress; it is frequently what unlocks it.

Log the sessions. Increase something. Sleep. That is most of it, and the rest is detail that matters far later than people think.

Nutrition generates more confident contradictory advice than almost any field, and a few structural facts explain why.

Most nutrition research is observational, because randomising humans to diets for decades is not feasible. Observational studies find associations, and associations are confounded by everything correlated with the behaviour being studied. People who eat more vegetables differ from people who do not in dozens of ways.

Effect sizes are usually small and headlines are not. A relative risk increase of 20% on a rare outcome is a small absolute change, and it is reported as though it were not. Learning to ask for the absolute numbers filters out a large share of alarming coverage.

Study length matters enormously. Weight loss trials of twelve weeks tell you about twelve weeks, and almost every diet performs well over twelve weeks. The interesting question is what happens at two years, and there are far fewer studies that answer it.

Funding is worth checking and is not disqualifying on its own. Industry-funded research is more often favourable to the sponsor, which is a reason for scrutiny rather than dismissal.

The practical filter: prefer the boring consensus over the exciting single study, ask what the absolute effect is, and be suspicious of anything that requires buying something.

The default response to an injury is to stop entirely, and it is usually both unnecessary and counterproductive.

Most injuries affect a movement rather than a body. A shoulder problem does not prevent lower body training; a knee problem does not prevent upper body work. Stopping everything converts a localised issue into a general loss of conditioning, and it removes the habit that took months to build - which is why so many injuries are followed by a year of not training rather than by six weeks.

The second point is that complete rest is no longer the standard advice for many soft tissue injuries. Appropriately loaded movement is generally part of recovery rather than a risk to it, and the appropriate load is a question for a physiotherapist rather than for the internet or for stubbornness.

That is the real caveat. Working around an injury requires knowing what the injury is, and self-diagnosis is where this goes wrong. Pain that persists, pain that follows a specific mechanism of injury, and anything involving numbness, instability or an inability to bear weight needs assessment before any of this applies.

What a copilot can do is help you keep training the unaffected parts, adjust the programme so the return is gradual rather than a cliff, and prepare the questions for a physiotherapy appointment.

What it cannot do is watch you move, which is the thing that actually determines whether a movement is safe for you right now.