#876: Dr. Andrew Huberman — Peptides, Performance, and Protocols

Why does it matter? Because every biological catalyst you're chasing is only half the intervention.

Huberman's through-line in this conversation cuts across exercise, psychedelics, peptides, and AI: a catalyst opens a window, but directed effort is what determines what forms inside it. Most people evaluate the catalyst. The people getting results are working the window. That distinction — simple, easy to miss, almost never spelled out — reframes how to think about nearly everything covered here.

• Quadruple your circulating BDNF through exercise and nothing neurologically changes unless you direct hard, specific learning during the elevated plasticity window that follows • High morning cortisol and low evening cortisol is the single unifying principle behind every health protocol; sunlight, cold, caffeine, and exercise are interchangeable tools for hitting that one bookend • Psilocybin can reach grief for things that never happened — absences, unlived experiences — that conventional therapy is structurally incapable of surfacing • The coming AI health breakthrough won't come from LLMs trained on internet text; it will come from models trained on actual neural recordings, and sleep is where that writing-to-the-nervous-system capability arrives first

You could quadruple your BDNF and nothing would change — the window after the catalyst is where plasticity actually forms

"You could quadruple the amount of circulating BDNF in your brain. Nothing's going to change. I know this for a fact because for years I studied neuroplasticity." That is Huberman's bluntest correction to the way most people think about exercise, nootropics, and psychedelics alike. The catalyst opens a window. What you do inside it is the intervention.

The mechanism: neuroplasticity requires precisely timed neural firing, pre- and post-synaptically. Alertness helps. Elevated neuromodulators help. But you have to go do the thing you want to rewire — learn the skill, confront the pattern, attempt the hard cognitive work — or the elevated chemistry closes without reforming anything. "Just getting the chemicals increase doesn't do it. You need alertness. You need these chemicals circulating." And then you need to go try something.

Facilitators who have supervised thousands of psychedelic sessions have long treated the two to three weeks following a session as a golden window. Huberman says Gül Dölen's research at UC Berkeley on MDMA-reopened critical periods now validates exactly that framework. Modern science reinforces what experienced guides already knew.

The same logic scales to brain stimulation. Using D-cycloserine to prime neuroplasticity before accelerated TMS, Huberman has achieved a 90% reduction in his generalized anxiety disorder and OCD lasting three to four months from a single day of treatment. Some patients report 18 months of durability for major depression.

The practical reframe: stop evaluating a nootropic or psychedelic session by how you feel during it. Evaluate what directed, intentional work you did in the days after.

Spike cortisol high in the morning, keep it low at night — every health protocol is just a tool for hitting this one bookend

"The real principle here is if your morning cortisol is high and you can make it higher doing any number of the things I just described, it ensures that your evening and nighttime cortisol is low." That sentence, Huberman argues, is the master key hidden inside every discrete health recommendation he has ever given.

Cortisol is widely framed as a stress hormone to suppress. The better frame: it is an energy-deploying hormone, and the goal is a high morning peak followed by a sustained evening floor. Morning sunlight in your eyes raises the cortisol awakening response by roughly 50%. Exercise triples or quadruples it. Cold exposure spikes it. Caffeine amplifies it. Every one of these is a different tool for hitting a higher morning peak — which in turn drives a lower evening baseline, which sets up melatonin rise and high-quality sleep.

The corollary matters too. A high-intensity workout at 9 p.m. creates a cortisol blip on top of an already elevated baseline instead of on top of a floor that is trending low. The timing of the spike is the whole game.

Once you understand the principle, protocol rigidity dissolves. "It's not about morning sunlight per se. It just turns out that morning sunlight in your eyes is the best way to ramp up that morning peak." On a plane, in Scandinavia in January, in a life that doesn't permit a morning walk — "choose a different tool, but the principles are what count." The bookend is the goal. Everything else is implementation.

Psilocybin surfaced grief for things that never happened — a category of pain no therapist had ever helped him reach

The hardest session Huberman describes wasn't MDMA alone or psilocybin alone. It was the combination, and it took him somewhere entirely outside the frame of what he thought he needed to process. "I'd never thought about the problem of what hadn't happened as a viable option to explore. It's so obvious now."

High-dose psilocybin layered onto MDMA several hours in produced a different quality of confrontation. Not cataloguing things that occurred. Sitting with absence — the experiences that didn't happen, the things he was in no position to create for himself, unlived chapters. "There is no way whatsoever that a therapist was going to sit down and go, 'Hey, let's talk about what didn't happen.' No one's ever done that for me."

This is a structural observation, not a complaint about therapists. Talk therapy is anchored in events. It works with what is present and retrievable. The absence of formative experiences is invisible to that frame — but can be a dominant shaping force. Psychedelics, in this account, dissolved the frame entirely: "it revealed both the thing that needed work and allowed me to do a lot of the work in real time."

The combination also notably eliminated the MDMA comedown. Psilocybin in the mix seemed to blunt it, which Huberman describes as common.

The caveats are real: clinical support, trained guides, hard exclusionary criteria (bipolar disorder, family history of schizophrenia or borderline personality disorder). Cannabis near these compounds carries underestimated risk. But the category of grief only reachable through dissolved framing — that is the finding worth sitting with.

The science caught up to Dorian Yates — but you think you're training to failure and you're not

After training legs with Dorian Yates in Marbella — the sorest he's ever been — Huberman returns with a correction to how most people read the hypertrophy literature. "The science caught up to Dorian." Low volume, high intensity, one or two all-out sets per exercise: the peer-reviewed data now validates what Yates was doing to win six Olympias.

But the validation comes with a catch that almost nobody accounts for. "You think you're training to failure. You're not training to failure." Huberman's illustration: "If somebody really put a gun to your head and said, 'Curl that bar one more time or I'm going to kill your family,' you'd curl the bar." Real failure demands a mental intensity that most people never generate in a gym.

The second catch is methodological. Virtually all hypertrophy research is conducted on subjects who do no cardio. When the literature says 10–20 work sets per week maximizes muscle growth, that number was generated in a context with zero conditioning. If you also run three days a week, your quads are getting stimulus on the assault bike, your lats are getting hit on pulling days, your entire system is under a different load than any study subject. Training each muscle directly once per week and indirectly a second time through cardio and compound overlap may be more than enough — and dramatically better suited to recovery.

Huberman's current split: legs, chest/back, and shoulders/arms once each, plus three cardio sessions and one long slow outdoor movement day. At 50, he's still getting stronger every session.

Competition is what destroys people — Dean Potter showed that seeking beauty outperforms seeking victory

Four episodes into The Dark Wizard — a documentary about free soloist, wingsuiter, and highline pioneer Dean Potter — Huberman arrived at something he frames as close to a principle of energy. "When you are in competition with other people, it's a misuse of the energy that can harm you. But when you are seeking what he calls just seeking beauty, like really trying to chase the feeling, the pure experience... that's when the really impressive feats land."

Potter's life illustrates both directions. When energy went toward competing — toward position, toward proving something against another person — things went sideways. When it went toward the experience itself, toward the sensation of committing to a line with no rope or flying a wingsuit through a gap in the rock, the feats became extraordinary. "Competition is, it turns out, is what destroys people."

Alex Honnold appears throughout the documentary as a counterpoint and a confirmation. Two climbers at the outer edge of what is humanly possible — and the one who has consistently transcended limits describes motivation in terms that have nothing to do with beating anyone.

"There does seem to be something almost spiritual about whether or not we're funneling energy into things that really reflect us at our best." The competitive frame feels productive because it generates urgency. Huberman's read is that urgency is real but misdirected — borrowed energy that eventually turns against you. The people reaching the ceiling are usually chasing a quality of experience, not a position on a ranking.

Most peptides won't do what you're hoping — the short list of ones with real effects is very short

"If you're expecting Cipionate HGH type results from peptides, you ain't going to get them." Huberman clears the expectation before it can contaminate the evaluation, then sorts the landscape into what actually does something and what is biology theater.

GLP-1 agonists (Ozempic, tirzepatide, the incoming retatrutide from Lilly, which also hits GIP and glucagon) produce robust weight loss because they push GLP-1 to a thousand times physiological levels. Human trial data exists. Effects are real. Side effects are real. This is a distinct category.

Growth hormone secretagogues — sermorelin, ipamorelin, CJC variants — spike GH three to eight-fold and improve deep sleep. Huberman tried sermorelin, watched his PSA spike significantly, came off it, and never found a mechanistic explanation. His verdict: "If you could get a good script for real GH and take a low dose versus taking a bunch of unknown secretagogue peptides, that seems like an obvious one for me."

BPC-157 has genuinely impressive animal data — cartilage regrowth, fibroblast migration, nerve regrowth. Zero published human trials. His position: use it if you have an injury, get it from a compounding pharmacy rather than the gray market, inject locally near the site. "Why would you take BPC? If you don't have an injury, it's not going to accelerate your recovery that much."

Pinealon — which does not come from the pineal despite the name — is the one he finds genuinely surprising: dramatic REM sleep increases, no suppression of slow-wave sleep, carryover into subsequent nights. "To my knowledge, there has not been a reported adverse event that can be tacked back to one of these non-FDA approved peptides. And I'm not saying that means they're safe." Absence of evidence is not evidence of safety when the reporting mechanism barely exists.

The real AI health breakthrough won't come from internet-trained models — it will come from AI built on neural recordings

"AI is trained on the internet. It's not trained on human brain activity." Eddie Chang, the UCSF neurosurgeon Huberman has known since childhood, framed the frontier more sharply than most AI commentary manages. The current LLMs are processing the outputs of human brains — text, images — not the source. The next step is training models on actual neural activity recorded directly from individuals.

The current wave is already medically powerful. Huberman names four or five people — well-resourced, with access to elite physicians — who found materially actionable health insights by uploading genomes, wearable data, and lab reports and letting an LLM mine the interrelationships. Insights no doctor had surfaced, not from negligence but because no human could hold that much data in mind simultaneously.

But that is still the internet-trained phase. Chang's argument: AI built on neural recordings would let individuals understand and shape their own brain states with precision that is currently impossible. What does your neural activity look like in your highest-focus state? What does a given compound actually do to it? What combination of sleep, behavior, and pharmacology gets you there reliably?

"The real holy grail of AI... is AI built on neural activity, not based on the internet, which is just the output of our neural activity." Writing to the nervous system — not just reading from it — is where it heads. Sleep is the first domain to watch: AI-controlled temperature environments, vestibular rocking mechanisms, eye-movement induction devices that can compress sleep onset below six minutes. The companies building hardware that measures and writes to the nervous system during sleep are prototyping the architecture for everything that follows.

The gap between stimulus and adaptation is where everything actually happens — and soon we'll be able to see inside it

What connects Dorian's training philosophy, psilocybin-accessed grief, cortisol bookending, and neural AI is a single, underappreciated gap: the space between the catalyst and the change. Right now, most people can't see inside that gap. They feel the stimulus, assume the adaptation, and move on. What Huberman is describing — in every domain here — is a future where that gap becomes legible: where the AI trained on your neural recordings can tell you not just that you opened a plasticity window, but whether you actually did anything with it. That changes the game entirely. The catalyst was never the point.

---

Topics: neuroscience, peptides, psychedelics, strength training, sleep, cortisol, neuroplasticity, MDMA, psilocybin, GLP agonists, BPC-157, Dorian Yates, AI health, performance, biohacking