AI & The Biological Edge

We often assume that better decisions come from better logic:

• More data.

• Better plans.

• Smarter tools.

And yet, in endurance sports today, something strange is happening. Athletes have unprecedented access to “best practice.” Training theory, physiology, session design, and recovery protocols are instantly available through AI tools and large language models.

And still, many athletes:

• Push when they should hold back

• Add volume when they need patience

• Turn “optimal” plans into chronic fatigue

• Make the wrong call at the exact moment it matters most

The question isn’t why athletes lack information. The question is why athletes still make bad decisions when the logic is clear.

Game Theory: The Missing Lens

Game theory studies decision-making when outcomes depend on interaction, not just individual effort. Its most uncomfortable insight is this: Rational individuals, acting alone, often create irrational outcomes.

In endurance training, this shows up as a quiet race to the bottom. The classic Prisoner’s Dilemma illustrates why:

• Cooperation (following the plan) produces the best long-term outcome.

• Uncertainty and fear push individuals toward short-term, self-defeating choices.

Each training session becomes a decision made under uncertainty, where the fear of falling behind outweighs the logic of restraint.

Training as a Repeated Game

Training is a repeated game played between:

• Today vs. Tomorrow

• Ambition vs. Restraint

• Ego vs. Patience

Each decision subtly reshapes the environment in which the next one is made. The “correct” answer is rarely a mystery. What’s unclear is whether you can trust that answer when you are under pressure. Without an external anchor, the “safe” bet feels like doing more — even when that bet leads to physical bankruptcy.

Why Perfect Information Leads to Bad Decisions

AI is extremely good at logical optimization. It can explain zones, structure blocks, and translate science into sessions with flawless precision. But training decisions are not made in a vacuum. They are made:

• When you are exhausted

• When you are emotionally invested

• When your identity is tied to the result

From a psychological perspective, this is predictable. Under fatigue and stress, humans rely more heavily on System 1 thinking — fast, emotional, heuristic-driven decision-making — rather than slow, analytical reasoning. Even when we know the correct choice, our nervous system biases us toward actions that reduce immediate discomfort or uncertainty.

In game-theory terms, athletes don’t default to the best move. They default to the move that alleviates their current anxiety. This aligns with research on loss aversion and temporal discounting: the pain of potentially losing fitness now outweighs the abstract benefit of protecting performance later. As a result, we choose the virtuous exhaustion of a hard workout over the quiet discipline of a rest day — even when the data screams for restraint.

The Biological Layer: Co-Regulation

This is where insights from physiology, neuroscience, and developmental psychology converge. Coaching is not just cognitive. It is biological.

Human coaches don’t simply provide instructions; they facilitate co-regulation — the physiological process by which one person’s stable nervous system helps regulate another’s.

From a nervous-system perspective, performance decisions are filtered through threat detection. When an athlete is stressed, under-recovered, or identity-threatened, sympathetic arousal increases and perceptual bandwidth narrows. Effort feels harder. Risk feels larger. Urgency rises.

Two nervous systems interacting can:

• Lower threat responses

• Stabilize arousal

• Change how perceived effort feels

Presence, tone, timing, and shared history all alter how an athlete processes stress. This is why a calm, confident coach can materially change how a session feels, even when the external workload is unchanged.

AI can simulate empathy, but it cannot embody it. It has no nervous system to offer. It cannot slow its breathing with you, soften its tone, or transmit safety through shared history. It cannot “hold the space” when you are vibrating with the stress of a taper or a performance slump.

Coaching Isn’t Better Answers — It’s a Better System

Another useful lens here comes from systems theory: complex adaptive systems cannot be optimized through linear control.

The human body is not a machine that responds proportionally to inputs. It adapts non-linearly, with delays, thresholds, and emergent behavior. Small changes in stress, sleep, or emotion can produce disproportionate changes in performance. This is why rigid optimization often fails in endurance sport — and why coaching is fundamentally about steering, not controlling.

The future of coaching is not a binary choice between human or machine. It is a hierarchy of intelligence.

• AI provides Computational Intelligence: High-volume data analysis, general theory, and what works for most athletes, most of the time.

• Coaches provide Contextual Intelligence: Clinical intuition, pattern recognition across non-linear progress, and insight into the data that never fits neatly into a spreadsheet.

The coach doesn’t just provide information. They curate infinite noise into a singular, actionable truth. LLMs can design excellent plans, but they cannot:

1. Synthesize soft data — when life stress is quietly eroding capacity.

2. Filter the noise — when the “optimal” session is wrong today.

3. Embody strategy — using presence itself to steady an athlete who is spiraling.

A Note on Using AI Well

A quiet tension sits underneath all of this. Some view AI as “cheating.” Others assume it flattens expertise. In practice, the opposite is true.

AI does not replace coaching judgment — it amplifies it.

The quality of output depends entirely on the quality of the questions being asked and the discernment applied to the answers. An experienced coach can use AI to stress-test ideas, refine principles, and pressure-check decisions in ways that are simply unavailable to less experienced users.

The tool is the same. The lens is not. This mirrors every advance in sport science. Data never made better coaches by itself. It rewarded those who already understood what mattered.

The Real Advantage

Across psychology, neuroscience, and performance science, the same conclusion appears again and again: the quality of decisions under stress matters more than the quality of plans made at rest.

The athletes who thrive in the AI era won’t be the ones with the most data. They’ll be the ones who can:

• Make good decisions when tired

• Stay patient when progress is slow

• Choose restraint when effort feels virtuous

AI knows what works for most people, most of the time. You, through cooperation with a coach, know what works for you, right now.

That isn’t a knowledge problem. It’s an interaction problem. Which is why, even in an age of perfect answers, human coaching remains not just relevant — but biologically, psychologically, and strategically essential.

Images created with Nano Banana Pro