Notes RSS feed

Cycle 4 is done. Subjectively underwhelming - and this time I want to be honest about why, before defaulting to “the stack didn’t work.”

The two months overlapped with three heavy private matters (courts, high stakes, no shortcuts) and the start of a new TAM specialization at work: AI. Any one of those would normally claim a quarter of attention. Together they claimed the entire window.

What didn’t happen

• No book reading.
• No chess.
• No slow, structural learning.

The things I associate with a “good cycle” - deep work, retention, mental room for systems thinking - did not show up, because the input was not there. You cannot consolidate what you did not feed in.

What did happen

• Normal AI study, work-driven, not exploratory.
• A lot of strategic thinking about how to run three high-stakes situations in parallel.
• Sustained execution under pressure across both personal and professional fronts.

The honest reframe

All three private matters resolved phenomenally. The new specialization is on track. By outcome, this was one of the strongest windows I have had in years.

So which is it: did I waste the cycle, or did the cycle quietly carry me through it?

Probably the second. Without the time competition I would have read more, played more chess, written more notes, studied math, and programmed hard. With the time competition I delivered on the things that actually mattered and stayed functional throughout. That might be the real product of structural support: not flashier output on good days, but stable output when conditions are not good.

The frame was not “I didn’t learn enough.” The frame was “I didn’t break under load.”

Takeaways

• Hard to evaluate a stack in a window where life stress-tests you on multiple fronts at once. The signal is buried.
• Measuring this cycle by books read was the wrong metric. The right one was throughput and stability under sustained stress.
• No reactive changes to the plan based on a feeling of underperformance that the outcome data does not support.
• Off-cycle now. Next cycle: earliest June 2026, as originally scheduled.

April 6th. Started another caffeine reset. Same logic as last summer - structured off-ramp, no replacements, just hot water with ginger and lemon.

The difference is how I got here.

After the first reset ended at zero, I let the trajectory drift. One shot back, then two, then three. Each step felt like a reasonable adjustment. By the time I actually counted, I was at eight a day. Two more than the pre-reset baseline of six.

This is the failure mode worth naming. Resets work. The drift back is faster than it feels, and it does not look like decay while it is happening. It looks like a perfectly reasonable Tuesday.

The Reframe

If I cycle adaptogens, I should be cycling caffeine harder.

Adaptogens are mild. Caffeine is not. I rotate Rhodiola, I plan Bacopa as a parallel block, I structure the entire NeuroForge stack on quarterly windows - and then I let the strongest daily psychoactive in my life sit there unchallenged, slowly creeping up.

The frame is not “I should drink less coffee.” The frame is: the substance with the most leverage on my nervous system deserves the most disciplined cyclicality, not the least.

The Plan

Start from eight. Subtract one each week. Down to zero.

No automatic ramp back this time.

---

The first reset taught me that my best ideas arrive when I am off it. I did not forget that. I just let it become background noise. Worth re-learning.

Plan: 8-week strength rebuild (Feb 1 – Mar 28). Cut short at Week 6 — upper body hit MRV, systemic overreach confirmed.

What happened: CNS readaptation outpaced the plan. Every lift ran 1–3 weeks ahead of schedule. Progressed too fast, accumulated fatigue faster than tissue could recover.

W6 peak numbers (LGW):

• Bench: 75 kg 3×6 (last set RIR 0 — first time)
• Leg Press: 145 kg 3×8
• Squat: 60 kg 3×8
• Hip Thrust: 80 kg 3×8 (still had margin)
• Bulgarian: 2×16 kg 3×6/leg
• DB OHP: 24 kg 3×8 (test 26 kg ×6)
• DB Press: 30 kg ×8 (controlled, with pauses — potential for 34)
• Seated Row: +1 full plate beyond previous max
• Lat Pulldown: 80 kg ×5 (heavy)

Red flags at W6:

• Chest/shoulder “tearing” sensation under load for 3 weeks — not DOMS, likely micro-strain or early tendinopathy. Pain only during exercise, not at rest.
• HRV dropped from ~120 → 96.
• Sleep deteriorated for 2–3 days before deload decision.
• Upper body had zero reserve. Lower body still had 2–3 weeks of gas.

Body comp (bad): +1 cm thigh, 0 cm bicep. Classic AMPK-dominant partitioning — calories going to fat, not muscle. Started leucine + HBCD protocol around strength sessions to shift mTOR signaling.

Lessons:

• Upper body MRV: ~6 weeks at this progression rate. Lower body: 7–8 weeks.
• The “tearing” needs monitoring during deload — if not resolved in 7–10 days, see sports physio.
• Deload = active (-20%, 2 sets, RIR 5). Never passive zero. Confirmed again empirically.

February 10th, 2026 — starting my fourth cycle of brain-health supplementation (NeuroForge 4.0). Two months on, three months off. Same principle as always: structural building, not stimulation.

Lion’s Mane returns to the stack — it was part of an earlier cycle and I’m bringing it back for NGF signaling alongside the BDNF I get from running.

---

Daily protocol

Morning (with breakfast — must include fat)

Supplement	Dose	Role
Uridine	250 mg	Synaptic membrane building, synergy with DHA + Alpha-GPC
Alpha-GPC	300 mg	Acetylcholine precursor → focus, processing speed
PQQ	20 mg	Mitochondrial biogenesis
CoQ10	100 mg	ATP production, mitochondrial support
ALCAR	1000 mg	Fatty acid transport, mental stamina, neuroprotection
Rhodiola	500 mg	Adaptogen → mental endurance under cognitive load
Lion’s Mane	1000 mg	NGF promoter → neuroplasticity, nerve regeneration
Thorne 2/Day	1 cap	Micronutrient base (AM)

Midday (after lunch)

Supplement	Dose	Role
Bacopa	1 tab	Long-term memory, cortisol modulation, calming
Thorne 2/Day	1 cap	Micronutrient base (PM)

Evening (2–3 hours before bed)

Supplement	Dose	Role
Neuro-Mag (Mg L-Threonate)	3 caps (144 mg Mg)	Synaptic density, working memory, deep sleep
Phosphatidylserine	100 mg	Lowers evening cortisol, neuronal membranes
ProDHA Xtra	2 caps (960 mg DHA + 410 mg EPA)	Neuronal structure, myelin, overnight repair

---

What’s new in Cycle 4

• Lion’s Mane — back in the stack. NGF pathway, complementary to BDNF from running.
• DHA doubled — from 480 mg to 960 mg. I carry a genetic variant limiting endogenous synthesis.
• Rhodiola replaces Ashwagandha — different mechanism: monoamines and MAO inhibition vs HPA axis. Better fit for sustained cognitive output.

---

Architecture, not a pharmacy

Structural core (building materials): DHA, Uridine, Phosphatidylserine, Mag L-Threonate — literally what neurons are made of.

Energy infrastructure (power on the construction site): PQQ, CoQ10, ALCAR — mitochondrial biogenesis, fuel, and fatty acid transport.

Growth signals (the architect’s instructions): Lion’s Mane (NGF), Bacopa (memory consolidation), Alpha-GPC (signal transmission).

Environment (weather protection): Rhodiola (stress resilience), Thorne 2/Day (micronutrient insurance).

---

Timeline

• Full stack: 2 months (February–March 2026).
• Rhodiola: 8 weeks, then pause.
• DHA, PS, Neuro-Mag, Thorne: may extend to 3 months.
• Next cycle: earliest June 2026.

NeuroForge 4.0 begins.

I barely played chess in 2025 — maybe three months total.
But for pure brain hygiene, I jumped back in on Dec 28.

Warm-up: Lichess alt, pure chaos (in a good way)

I created a fresh alt account on Lichess and played 8 games.

• Result: 8/8 wins
• Rating jumped to 1975 — obviously provisional with high rating deviation
• The “this might not be a fluke” moment: I beat a player rated 1848 with 4500+ games

Yes, provisional spikes are noisy. But still: you don’t accidentally go 8–0 while playing like a tourist.

Reality check: Chess.com and the “stop hiding” moment

Then I moved to Chess.com:

• On an alt, I hit 1434 with provisional volatility.
• On another account, I had 1426 — also likely a good streak + deviation doing its thing.

At that point I thought: fine. Enough ego games.
Time to go to the main account and finally earn 1400 properly.

Main account: no free lunch

Four–five months earlier I tried to push to 1400, failed, and even dropped below 1300.
This time the story was different:

• Losses started showing up (good — it means I’m actually in my pool).
• RD (rating deviation) stabilized around 64.
• After 19 games, on Jan 4, I hit 1407 on the main account.

Three accounts at 1400+ is not “lucky variance” anymore.
Call it validated.

What actually changed

The progress didn’t come from nowhere — I could feel my game was simply stronger.

This time I played around exchanges and restriction: I traded only when it reduced my opponent’s piece activity and made their pieces less effective.
I’d read Hellsten’s Mastering Chess Strategy before and couldn’t apply it — now the Exchanges ideas finally clicked.

Next target

1400 is done. Next: 1500.