Hello Reader,

As our team at Athletica puts the final pieces in place for the launch of our new platform, including the new UIX, conversational AI-Coach, and Athletica U, I wanted to pause and share a broader perspective.

There is a lot of noise right now around “AI coaching.” Much of it is well intentioned, but some of it risks distracting from the real work that matters: helping athletes and coaches make better decisions day to day.

The piece below is longer than a typical email, but it sits at the core of what we are building at Athletica. It outlines how sport science has evolved, and why the next step is not more data or smarter predictions, but better coaching choices grounded in physiology, training theory, and context. This is what we mean by Sports Science 3.0.

This will be my last long-form note here for a little while, as we shift focus to letting our team properly walk you through what’s coming with the platform launch in the new year.

I hope you enjoy the read.

Paul

Why the next era of coaching won’t be defined by data, dashboards, or AI hype, but by better decisions.

The rise of the “AI coach”

We are entering a phase in sport where nearly every platform will describe itself as an “AI coach.”

The promise is seductive: sync your data, trust the system, and let the algorithm decide what you should do next.

On the surface, this sounds like progress.

But beneath it sits a quiet assumption that deserves scrutiny: that better predictions automatically lead to better decisions.

They don’t.

Prediction is not the same as decision-making

For most of coaching history, the hard part was not knowing what could happen. It was deciding what should happen next. For a particular athlete, in this moment, with this history.

That problem has not gone away. It has simply been hidden under more computation.

Coaching has always been a judgment problem

AI is exceptionally good at pattern recognition. It is far less equipped to understand context, purpose, and consequence without constraint.

Left unconstrained, intelligent systems will always optimize for confidence and continuity. Training, however, demands restraint.

Sometimes the right decision is not the most obvious one.
Sometimes it is not the most aggressive.
Sometimes it is to do less, not more.

From automated prescription to decision support

This is why I believe the future of coaching is not automated prescription.

It is decision support grounded in physiology, training theory, and human judgment.

Expert-designed structure is not a relic of the past.
It is the foundation that keeps adaptation meaningful rather than reactive.

AI should not replace that structure.
It should operate within it.

When progress meant measuring more

For most of my career in sport science (above photo), progress was easy to define.

We measured more.
We measured better.
We measured faster.

Physiology moved out of the lab.
Sensors moved onto athletes.
Dashboards multiplied.

And yet, despite all of this, the most common question I still hear from coaches and athletes is unchanged:

"What should I do today?"

The Limits of Measuring More

Sports Science 2.0 gave us powerful tools.

• Heart rate
• Heart rate variability
• Power meters
• GPS
• Continuous glucose monitoring

But tools alone don’t make decisions.

In practice, more data often led to:

• More interpretation burden on coaches
• More second-guessing by athletes
• More confidence in numbers, less confidence in judgment

The issue was never data availability.
It was integration.

Decision-making under uncertainty

Coaching is not about knowing everything.
It is about choosing well under uncertainty. Choosing well within the context.

When to push.
When to hold back.
When to deviate from the plan.

Those decisions are shaped by physiology, but also by:

• Athlete profile
• Training history and fitness
• Psychological state
• Life stress
• Confidence
• Injury risk, etc…
  ​

Sports Science 2.0 described these factors.
It did not connect them.

Sports Science 1.0 vs 2.0 vs 3.0

Sports Science 1.0 focused on experimentation in controlled settings, building foundational physiological principles through laboratory research and group averages.

Sports Science 2.0 shifted science into the field, emphasizing measurement through wearables, sensors, and dashboards, dramatically increasing data availability but often leaving interpretation to coaches.

Sports Science 3.0 focuses on decision support, integrating physiology, training theory, athlete context, and observed load–response to help coaches and athletes choose better actions day to day.

This progression captures the arc of sport science over the past three decades.

What We Mean by Sports Science 3.0

​Sports Science 3.0 is partly a technology upgrade.

But equally, it is a philosophical shift.

From:

• Measurement → decision support
• Metrics → context
• Static plans → adaptive systems
  ​

At its core, Sports Science 3.0 asks one question:

How do we leverage foundational principles with sensor technology and AI to help coaches and athletes make better day-to-day training decisions?

Context Over Content (Now Made Operational)

For years, coaches have said, “It depends.”

That was never a weakness.
It was wisdom.

What’s changed is that context no longer has to live only in the coach’s head.

With longitudinal data, athlete feedback, and computational models, we can now:

• Learn how this athlete responds over time
• Use deviation as information, not failure
• Adjust training based on response, not intent

Context doesn’t disappear.
It becomes actionable.

What AI Means in Sports Science 3.0

Here, AI does not mean automated prescription or systems that decide training in isolation. It refers to decision support systems that are constrained by physiology, training theory, and human judgment. These systems integrate data, context, and expert structure to inform decisions, not replace the coach or athlete responsible for making them.

Where AI Fits (and Where It Doesn’t)

AI is not the coach.

AI is coaching infrastructure.

Its role is to:

• Integrate complexity humans can’t hold in working memory
• Recognize patterns across time, not moments
• Provide guardrails and guidance, not instructions
  ​

Great coaches don’t disappear in Sports Science 3.0.
They scale.

The Athlete Is Not a Dataset

Athletes are not passive data generators.

Their comments, perceptions, effort, and confidence are not noise.
They are signal.

Sports Science 3.0 treats athletes as active participants in the system (see below), not objects being monitored.

When athletes understand why a decision is made, trust follows.

Why This Matters Now

We are entering an era where everyone will claim to have an “AI coach.”

Many will be built on:

• Conversational layers placed over generic plans
• Prediction engines detached from training principles
• Confident outputs without physiological accountability
  ​

The real differentiator will not be artificial intelligence alone.

It will be judgment, informed by expert-designed training structure and systems, athlete context, and observed load–response over time.

A Living Example

These ideas aren’t theoretical.

In our recent year-end conversation, Sports Science 3.0: The Next Chapter, Martin Buchheit and I reflected on how this shift is already unfolding, in rehab environments, endurance sport, team sport monitoring, and AI-supported coaching systems.

We discussed:

• Why prediction is not the same as decision-making
• Why athlete context matters more than any single metric
• Why AI must be constrained by evidence, not novelty
• Why the future of coaching is adaptive, not automated

If you want to hear these ideas evolve in real time, that conversation is available here.

A Closing Thought

If Sports Science 1.0 was about experimentation, knowledge generation, and foundational principles, and Sports Science 2.0 was about measurement in the field, then Sports Science 3.0 is about choosing better.

Better for performance.

Better for health.

Better for longevity.

Those choices should be evidence-informed, efficiently delivered, and grounded in a rationale that is fully explainable.

That is the shift.

That is the work ahead.

We are grateful for the support of our Athletica community, whose members are active participants in this journey with us.

​​​​​​