Why AI-Generated Cardio Beats C25K for Lifters (With Data)

Concurrent Training Interference: What the Research Actually Says

The concurrent training interference effect was first described by Robert Hickson in 1980. Subjects who added 6 days/week of endurance training to a strength program plateaued in leg strength after ~7 weeks and then regressed, while pure strength-training subjects kept progressing. Something about combining the two modalities was antagonizing the strength adaptation.

Thirty years later, Wilson et al. (2012) meta-analyzed 21 studies to quantify the effect. The headline findings: (1) concurrent training reduces hypertrophy and strength gains compared to strength training alone, in a dose-dependent way; (2) interference grows with the frequency and duration of cardio sessions; (3) running interferes more than cycling, likely due to the eccentric component; and (4) high-intensity intervals interfere more per minute than zone 2 work.

The mechanisms aren't fully settled, but two well-supported contributors are: acute residual fatigue (your squat session is limited if your hamstrings are still recovering from yesterday's intervals), and molecular signaling competition (AMPK-driven endurance signaling attenuates mTOR-driven hypertrophy signaling when the two are layered close in time).

The practical takeaway: cardio isn't the enemy for people who lift, but how and when it's programmed matters a lot. A generic plan that doesn't know what you lifted this week is flying blind into the exact variable that determines whether the cardio helps or hurts.

Why Fixed Templates Break (Seiler and the Polarized Model)

Couch-to-5K, Nike Run Club base-building plans, generic marathon plans — they all share a structural flaw: the week N session is predetermined, regardless of your state on the day. Week 4, Day 2 might say “run 5 min, walk 3 min, repeat 5x.” It doesn't know you squatted heavy yesterday. It doesn't know you slept 4 hours. It doesn't know you're fighting a cold.

The second problem is intensity distribution. Seiler (2010) reviewed the training logs of elite endurance athletes across multiple disciplines and found a consistent pattern: ~80% of total training time sits in Zones 1–2 (low intensity, aerobic base), and ~20% sits in Zones 4–5 (high intensity, threshold and VO2max). Zone 3 — the “tempo” or moderately-hard zone — is deliberately avoided.

Why avoid Zone 3? Because it accumulates fatigue nearly as fast as threshold work but produces smaller specific adaptations. For elite athletes, time spent in Zone 3 is time that could have been either truly easy (bigger aerobic base, minimal fatigue) or truly hard (bigger VO2max stimulus). The polarized distribution extracts more adaptation per unit of fatigue than a middle-heavy distribution.

Most generic templates violate polarization. Couch-to-5K running intervals land squarely in Zone 3–4 for most beginners; the “long run” in many beginner marathon plans is often run too hard, drifting into Zone 3. Laursen & Jenkins (2002) also documented how well-structured high-intensity interval training (HIIT) produces disproportionate aerobic adaptations when properly programmed — but the operative phrase is properly programmed, not dropped into a fixed calendar.

When you add lifting on top of a non-adaptive plan, the mismatches compound. The plan doesn't know to make a Zone 2 day shorter after a heavy deadlift session, or to shift intervals to a different day, or to substitute cycling for running to reduce eccentric load.

HR Zones 1–5 and Why Zone 2 Isn't Just “Low Heart Rate”

Heart rate zones partition aerobic effort into 5 bands. A common model uses percentages of maximum heart rate, though lactate-threshold-based models (used in more precise contexts) map slightly differently.

Zone 2 is the zone that's been getting most of the attention lately, largely because of Peter Attia's work with Iñigo San Millán. The claim isn't that Zone 2 is mysteriously special — it's that it drives mitochondrial density and fat oxidation more effectively (per unit of fatigue) than higher-intensity work, and it minimizes interference with strength training.

A key detail: Zone 2 isn't just “low heart rate.” It's the upper edge of fully aerobic work — just below the first lactate threshold. You should be able to hold a conversation but notice breathing slightly deeper than at rest. Many beginners run their “easy” runs in Zone 3, which both accumulates more fatigue and provides less of the mitochondrial stimulus. This is a failure mode that fixed templates can't fix because they don't read HR at all.

For lifters specifically, Zone 2 has one more practical benefit: it's the lowest- interference aerobic modality. Low-intensity cycling or incline walking in Zone 2 barely registers on AMPK-mediated interference pathways, and the cardiovascular adaptations carry over to recovery between lifting sets.

How Arvo's AI Balances Strength and Cardio

The Arvo AI Cardio Coach treats each cardio session as a prescription written the morning of — not a calendar entry locked in three weeks ago. The prescription reads multiple inputs:

• Strength load (last 72h). What muscles were trained, at what volume and intensity? Heavy squats in the last 48h push leg-dominant cardio (running, hard cycling) further out.
• Current HR zone targets. Based on observed resting HR, recent session HRs, and (if available) lab or field testing. Zones recalibrate as fitness changes.
• Readiness signal. Sleep (from wearable integration if enabled), subjective readiness, and any fatigue flags from recent sessions.
• Adaptation goal. Zone 2 base-building vs. threshold work vs. VO2max development, chosen based on the overall mesocycle phase and upcoming strength block demands.

The output isn't just “do 30 min cardio today.” It's a specific session type (e.g. “Zone 2 cycling, 40 min, 125–140 bpm”), with rationale (“legs fatigued from yesterday's squats — cycling chosen over running to reduce eccentric load”). The same engine will shorten the session, drop the intensity, or substitute an active-recovery walk if readiness signals indicate it.

On heavier-readiness days, the same engine can prescribe the high-intensity work. The key is that the timing of interval sessions is dynamic: HIIT lands on days when legs are fresh and the next lifting session is 24+ hours out, not on a fixed Tuesday regardless of what happened Monday. For lifters managing concurrent training, this dynamic placement is where the largest gains over a fixed template are realized.

Example: What AI Prescribes the Day After Heavy Squats

Imagine two plans, side-by-side, the day after a heavy squat session (5 sets of 5 at high RPE, near-max effort).

The difference isn't that one plan is “harder” and the other “easier.” It's that the AI-prescribed session is correctly timed. The same athlete, three days later with fully recovered legs, might be prescribed a Zone 5 interval session — because now the physiology supports it.

This is what “AI cardio that compounds with strength” actually looks like in practice: reading the real inputs (load, recovery, HR, goal), choosing the modality that fits the day, and adjusting intensity to land in the right zone for the right duration. A generic template can't do this because it can't see the inputs. A coach can, but at cost. Software that integrates your strength log with your cardio prescription can do it continuously.

For more on how Arvo's recovery model feeds into these decisions, see the recovery and sleep resource, or explore the live Recovery Map in the product.

Research references: Hickson (1980) on the original concurrent training interference finding; Wilson et al. (2012) meta-analysis on concurrent training effects; Seiler (2010) on polarized intensity distribution in elite endurance athletes; Laursen & Jenkins (2002) on high-intensity interval training adaptations. Zone 2 training framework popularized by Peter Attia and Iñigo San Millán.