Durable Sessions vs Exercise Intensity: Which Skill?

What does the Christensen Durable Sessions AI UX Framework do?

This skill diagnoses why AI chat and agent experiences break under real-world conditions—and it fixes them at the infrastructure layer, not the model layer. The central insight is the Single-Connection Trap: default direct HTTP streaming (SSE via the Vercel AI SDK, LangChain streaming, raw WebSockets) couples stream health to one client's connection. Drop the connection and the stream is gone.

The framework scores your product against Three Foundational Capabilities: Resilient Delivery (streams survive disconnects), Continuity Across Surfaces (session follows the user across tabs and devices), and Live Control (clients can steer, interrupt, or cancel mid-generation). It then prescribes a Durable Session—a persistent, independently addressable, fully resumable pub/sub channel that sits between the agent and client layers. Agents write events; clients subscribe. Neither holds a private pipe to the other.

It also resolves specific failure modes: the SSE Resume-Cancel Conflict (a closed connection can't distinguish 'disconnect' from 'stop'), and the Orchestrator Dual-Purpose Problem (an orchestrator forced to relay sub-agent progress). The output is a validated architecture you test by dropping connections, opening second tabs, and firing cancel signals.

What does the FoundMyFitness Exercise Intensity Optimization Skill do?

This skill designs a personal exercise protocol that maximizes longevity, VO2 max, and disease-risk reduction. It is grounded in biobank wearable data that establishes Intensity Exchange Ratios—fixed time equivalences between intensity tiers. For all-cause mortality, 1 minute vigorous = 4 minutes moderate = 53 minutes light. For cardiovascular disease it is 1:8:73; for type 2 diabetes, 1:9:94.

The headline finding: vigorous intensity wins across the board. There is no evidence moderate intensity beats vigorous on any endpoint studied. The skill also flags the 40% Non-Responder Problem—roughly 40% of people don't meaningfully improve VO2 max from moderate-only training, which makes at least one vigorous session per week effectively mandatory.

Zone 2 (breathy-conversation pace) is not dismissed; it builds the aerobic base and serves as active recovery. But it is complementary, not equivalent. The workflow audits your current intensity distribution, maps your goals to the right ratio, checks for the non-responder signal, and produces a concrete weekly template—anchored by vigorous sessions, supported by Zone 2, and personalized around adherence and burnout risk.

How do they compare?

Honestly, they don't compete—they occupy entirely different universes. One is a real-time systems architecture framework for software engineers; the other is an evidence-based health protocol for anyone with a body. There is zero functional overlap in inputs, outputs, or audience.

What they share is structural rigor. Both are diagnostic-then-prescriptive: audit the current state, identify failure modes, apply a principled fix, then validate. Both refuse false equivalence—Christensen insists Durable Sessions (not just WebSockets) are required for multi-surface visibility, and FoundMyFitness insists vigorous exercise cannot be fully substituted by Zone 2 volume. Both also privilege real-world constraints: Christensen targets the gap between a fragile demo and a production experience; FoundMyFitness insists a sustainable protocol beats an optimal one that gets abandoned.

On complexity, the AI UX framework is significantly heavier. It demands technical fluency in transports, pub/sub, and agent topology, and applying it means rearchitecting production systems. The exercise skill is far more accessible—classify your sessions, apply a ratio, build a schedule. Time-to-value follows suit: minutes to a workout plan versus days or weeks to a validated architecture redesign.

Which should you choose?

Choose based on your task—there is no judgment call here.

Choose the Christensen Durable Sessions AI UX Framework if you are building or auditing an AI product and any of these are true: mobile users lose responses when switching networks, a second tab can't see the live response, your stop button conflicts with resume logic, or your orchestrator is drowning in relay code. This skill is the clear winner—and the only relevant option—for AI streaming and agent-infrastructure problems.

Choose the FoundMyFitness Exercise Intensity Optimization Skill if you want to structure or fix a weekly workout for longevity, cardiovascular health, metabolic health, or VO2 max—especially if you're unsure how much Zone 2 versus vigorous work to do, or you've plateaued on moderate-only training. It is the clear winner for exercise-programming decisions.

If you happen to need both—an engineer optimizing both their product and their health—run them independently. They will never conflict because they never touch the same problem. Do not try to blend them; the value of each comes from its domain-specific precision.