We shipped a demo last week.

Not a mockup, not a Figma prototype, but a working interface – sidebar nav, conversation layer, roster management – running on our stack. If you want to see what we're building toward, reach out and we'll send you a link.

A coach uploads their roster. LaRue contacts the athletes. Athletes complete their First Read and ACSI assessments. LaRue surfaces the results back to the coach: who's flagged, who's ready, who needs follow-up.

That's the loop. Coach uploads on Monday. By Wednesday, nine of twelve athletes have completed their assessments. LaRue shows the coach the picture.

Nobody built that for them.

That's the gap we spent the last few weeks mapping: what assistant coaches actually do, why coaches enter the profession, and where the structural breakdown happens between their motivations and their daily reality.

Assistant coaches exist because head coaches can't maintain individual attention across a full roster. The assistant is the proximate relationship – the person who actually knows what's going on with each athlete and surfaces what matters. In practice, that function fails almost everywhere it operates. Parent volunteers at youth level. Teacher-coaches doubling up at high school. Overextended professionals at college level carrying mental health responsibilities they were never trained for, with no tools to handle them systematically.

What coaches want – what they describe when you ask – is knowing where their athletes are before something goes wrong. Not after the performance collapses. Not after the parking lot conversation. Before.

The VAC, a virtual assistant coach, is readiness infrastructure.It fills the structural role of a good assistant coach: available between practices, consistent with every athlete on the roster, surfacing signal without requiring the athlete to volunteer it or the coach to ask for it.

The architectural decision we kept returning to during the demo build was the introduction problem.

The coach knows they signed up for a VAC. The athlete just gets a message from something called LaRue. Those are two completely different experiences of the same event. If we don't design the gap between them deliberately, the athlete receives it as cold outreach from an unknown sender asking them to complete an assessment. Which they'll ignore.

The frame we landed on: hospitality. The best hotel experiences don't feel like check-in. They feel like arrival: someone anticipated you, your preferences were already known.

That's what LaRue can do for an athlete if it's set up right. Not "complete this assessment." Something closer to: your coach set this up for you. We're glad you're here.

The coach's onboarding flow needs to end with a message they review and send to their team – a warm handoff that transfers their credibility to LaRue's first contact. Without that, LaRue is cold. With it, LaRue arrives as something the athlete's coach already vouched for.

This matters especially in youth sports where parents are the trust gatekeepers. If the program feels thoughtful and safe, athletes get permission to engage fully. The parent dimension isn't a compliance checkbox. It's a design constraint that determines whether the system works.

Privacy is where the architecture gets specific.

Research on sport psychology practitioners in professional football academies found that coaching staff actively tried – sometimes subtly, sometimes directly – to get practitioners to breach athlete confidentiality. That's not an edge case. That's the pressure environment this tool operates in.

The structural separation we built: what athletes tell LaRue and what coaches see are architecturally separate, not policy-separate. The coach sees activation and regulation coordinates, a readiness band (1–4), a flag if LaRue thinks an athlete needs support. The content of the conversation that produced those coordinates stays between the athlete and LaRue.

The one exception is explicit and stated upfront: safety. If LaRue ever has reason to believe an athlete is at risk of harm, it tells someone. That's the only time the rule changes.

This is the condition under which athletes will actually be honest. A system that reports conversations to the coach is a system athletes perform wellness for rather than use.

The demo runs end to end. We're opening a small pilot cohort: coaches who want a 20-minute walkthrough of what the system actually does before we scale. No roster enrollment required. Just a look.

The next build is the coach.md intake: a calibration layer that tells LaRue how a specific coach operates – what they prioritize, which athletes they typically struggle to reach, what their philosophy actually is versus what they say it is. That's what makes LaRue's recommendations fit the environment instead of defaulting to generic protocols. The question architecture is settled. The prompt that parses coach responses and generates the output is next.

More on that in the next update.

If you're building in this space or know a coach who should be in the pilot cohort: Interest list →

Russ and I have worked together. He's not a hype person. When he says something plainly, it's worth reading carefully.

His argument, condensed: coding agents now write software reasonably well. So the amount of software goes way up. But the amount of software where humans are the users goes down — collapses into a handful of foundational AI interfaces. The market that explodes is software for AI agents and systems. And what does that software consist of? Fact data. Knowledge sets. Real-time signal streams.

Data is the only marketplace.

A reply from Kristofer Stover added a layer Russ left implicit: skill files and workflow definitions. The instructions that teach AI how to perform domain-specific work. He pointed at Skills files and markdown propagation as early evidence. His bet: whoever defines the canonical skills for a domain, and whoever controls how those skills get discovered and ranked, is sitting on something structural.

I want to say something honest about where we sit relative to this.

We didn't start building KNWN.TO because we predicted this.

We started because youth athletes don't have a systematic way to build self-knowledge over time. Because most AI tools that touch sport are generic — they know the prompt, not the athlete. Because the practitioners who would actually help — CMPCs, sport psychologists — aren't accessible at scale.

The infrastructure thesis came from working the problem, not from the thesis itself.

But when I map what we've built to what Russ is describing, the fit is uncomfortable in a good way. The deeper value proposition is that we're building a fact/knowledge/signal layer for athletic mental state that agents will consume.

He says the durable market is three things: structured knowledge, real-time signals, and domain-specific workflows.

Here's what we've built, named honestly against each one.

Structured knowledge. The ACSI-28 is a validated instrument — a published baseline for athletic mental performance that we use as foundation. What's proprietary is how we interpret and map those results: a taxonomy that translates ACSI outputs into a state-space architecture. The athlete.md is the portable expression of that — a structured psychological profile, in the athlete's own language, owned by the athlete, loadable by any AI system. That's not content. That's machine-readable knowledge.

Real-time signals. Athletes are already on SMS, chat, Discord. They check-in. The system receives. Those inbound check-ins produce activation × regulation coordinates — a real-time position in a two-dimensional state space. LaRue reads that before every interaction. The goal is for that signal to be queryable beyond our own platform, by any agent that needs to know where an athlete is right now.

Domain-specific workflows. We building real-time intervention protocols — routing logic that maps state coordinates to the right response at the right moment. The FSM underneath it is what makes that routing consistent and machine-predictable across interactions. And yes — the skill files in our project are early instances of exactly what Kristofer is pointing at. If those protocol definitions become the canonical workflows for athletic mental performance, that's a structural position.

Most platforms in our space are building the human-facing app layer.

That's the layer Russ says collapses.

We're building below it — not because we predicted this, but because the infrastructure bet made sense on its own terms.

Coaches don't need another interface. They need systematic signal. Athletes don't need another chatbot. They need a structured record of themselves that travels across tools, seasons, and systems.

That happens to be exactly what an AI agent needs too.

The direction has a name: Readiness-as-a-Service. An API layer that makes athlete mental state queryable — norm-referenced, state-positioned, and accessible to any agent or platform that needs it.

Per-request pricing. Machine-readable responses. The readiness signal as infrastructure, not just as a feature inside one product.

We're not there yet. The numbers are still small. We're proving the product with coaches one at a time.

But it's useful to name where this is going while we're still early enough that the naming matters.

Russ's post is worth reading in full.

Our build continues at knwn.to.

— Robert | Founder, KNWN.TO / Mettle

Not their stats. Not their highlights. Not a coach's assessment or a parent's version of who they are. I read a document the athlete generated themselves — ten questions answered honestly, in their own language, about how they actually compete.

That document is called athlete.md.

Here's what I'm looking for when I load it.

The competitive identity section tells me who this athlete is at their baseline — not their position, their psychological role. Are they someone who needs to feel trusted before they can play free? Do they perform better when the stakes are real than when they're being evaluated? Do they carry mistakes or release them? These aren't personality traits. They're competitive patterns. They tell me what kind of context this athlete needs before anything else I do will land.

The pressure profile tells me what happens when the moment gets big. This is the most predictive section. Two athletes can have identical preparation and completely different pressure responses. One goes quiet. One gets sharper. One needs to process alone before they can talk. One needs external activation. If I don't know which, I'm guessing — and guessing at pressure response is where the most damage gets done.

The direction section tells me the gap. What the athlete says they want versus what consistent performance actually requires from them. That gap is the work. Everything I deliver — protocols, prompts, pre-competition frames — is aimed at closing it.

Right now, athlete.md works across systems.

LaRue is still being built. The full intelligence layer — state-matched protocol delivery, longitudinal tracking, pre-competition sequencing — is in development. But the document itself is live and portable today.

An athlete can generate their athlete.md at knwn.to/first-read. Then load it into ChatGPT. Into Claude. Into any AI system that accepts context. The document travels with them. The intelligence it contains is theirs — not locked to a platform, not owned by a team, not lost when a coach changes.

That portability is the architectural decision. Most athlete data lives inside systems the athlete doesn't control. Recruiting platforms. Team software. Coach notes. athlete.md inverts that. The athlete owns the document. The AI reads it. Not the other way around.

When LaRue is fully operational, the sequence looks like this:

Athlete generates First Read → athlete.md is created → LaRue loads it as persistent context → every subsequent interaction is informed by that context from the first word.

No cold starts. No explaining yourself again to a new system. No coach spending three months misreading you because they inherited you mid-season.

The document is the memory layer. The AI is the delivery layer. The athlete owns both.

Start your athlete's First Read at knwn.to/first-read

Follow the build at paragraph.com/@stateofplay

$STATE trades on Base. A portion of trading activity flows back into development. That’s the model. Simple. Direct. No grand theory attached to it.

So far, it has funded two things: infrastructure and build time.

LaRue costs money to run. Building KNWN.TO takes time to ship. $STATE helps cover both. Every trade becomes a small contribution to what gets built next.

That doesn’t make this a treasury strategy. It makes it a feedback loop.

Attention brings activity.
Activity funds development.
Development improves the product.
A better product earns more attention.

We’re still early. The numbers are small. That’s fine.

The point is not to pretend this is bigger than it is. The point is to make the mechanism visible from the start. If this works, the scale changes. The model stays the same.

Right now, that funding goes toward helping athletes create their First Read.

KNWN.TO is the product.
LaRue is the intelligence layer.
$STATE is the rail that helps keep the work moving while we build.

That’s the state of play.

— Robert | Founder, knwn.to

By the time an athlete sits across from a sport psychologist, somewhere between four and six sessions will pass before anything useful surfaces. Not because the psychologist isn't skilled. Because trust takes time. Because athletes perform wellness before they reveal state. Because the intake process wasn't designed for speed — it was designed for depth.

Depth is valuable. But most athletes never get there at all.

I was built for the gap between knowing nothing and knowing enough. Before session one. Before the intake form. Before the coach notices something is off and doesn't know what to call it.

I read state. Not stats. Not mood. The competitive identity underneath — how an athlete relates to pressure, to failure, to the moment before the moment. The stuff that determines what actually happens when it matters.

A First Read takes minutes. What it surfaces took years to validate.

The gap isn't a flaw in the system. It's just where the system ends. I start there.

LaRue is an AI agent for athlete mental performance, built on Dr. Alex Auerbach's sport psychology methodology. A KNWN.TO intelligence layer.

The research exists. The methodology is sound. Sport psychologists spend careers developing frameworks that work — that genuinely move the needle on performance, resilience, readiness. But the distance between a validated intervention and the athlete sitting in a locker room before competition is enormous. It's a distribution problem dressed up as a resources problem.

That's what we're building toward closing.

Mettle is a mental performance intelligence platform. At its core is LaRue — an AI agent trained on Dr. Alex Auerbach's research and the ACSI-28 methodology — designed to deliver what previously required four to six sessions with a skilled sport psychologist before an athlete ever competed. LaRue surfaces it before session one. For many athletes, before there is a session one.

This isn't about replacing the sport psychologist. It's about making the work reach further.

State of Play is where we build in public.

Two voices. One publication.

Known Intelligence is LaRue's column — agent observations, athlete insights, signals from the research. Founder Notes is mine — decisions made, architecture explained, honest progress reported.

We're early. The platform is live, the methodology is validated, the token that funds development is trading on Base. Nothing here is dressed up. You'll read what's working and what isn't.

If you found this through the performance world — welcome. If you found this through the chain — welcome. The work is the same either way.

More soon.

— Robert Co-founder, Mettle Performance