Semax and TAK: separating Soviet-era research from TikTok hype

What did @helmetsway actually say?

The creator ran a self-timed reaction test before and after taking TAK, reported "an improvement of 8.8 milliseconds," and claimed the onset hit "about 30 minutes." They also said TAK offers "an umbrella of many different effects" and will "improve your gaming performance in many other ways." The video ends with a pitch for a paid "cognitive university" enrollment.

To be precise about what we're evaluating: TAK stands for a proprietary or informal name sometimes applied to peptide stacks or specific nootropic compounds circulating in biohacking communities. The creator does not name the exact molecule on screen, which is itself a problem we'll get to.

Does the science back this up?

Not in any meaningful way for the specific claim made here. An 8.8-millisecond improvement measured over 10 self-administered trials on a consumer reaction app is not evidence of a drug effect. It is noise.

Reaction time variability within a single individual across sessions routinely spans 20-50 milliseconds without any intervention. A 2011 paper by Kosinski published in Perceptual and Motor Skills documented intra-individual reaction time variability across repeated testing sessions, finding natural fluctuation ranges that easily swallow an 8.8ms difference. No placebo control, no blinding, no baseline washout period, no statistical power. The creator did not account for learning effects, either. People simply get faster at reaction tests the more they take them, a well-documented practice effect. Mowbray and Rhoades (1959, Quarterly Journal of Experimental Psychology) described this phenomenon decades ago. The 30-minute onset claim is specific enough to sound scientific but is completely unsupported by any published pharmacokinetic data the creator cites, because they cite none.

What did they get wrong (or right)?

Wrong on the methodology, wrong on the implied causation, and evasive on the substance itself. The creator never names the exact compound, which makes this impossible to fact-check at a molecular level and conveniently shields the claim from direct scientific scrutiny.

Some peptides in the nootropic category, like Semax (mentioned in the hashtags), do have genuine research behind them, though mostly from Russian preclinical and small clinical studies. A 2014 review by Menshanov published in Zhurnal Vysshei Nervnoi Deyatelnosti found Semax had neuroprotective and attention-modulating effects in animal models and small human trials. That is legitimately interesting. But "interesting in controlled research" is a very long road from "take this before a ranked match." The gaming performance claim is pure marketing language with no evidentiary basis. One thing they got right: acknowledging that a single metric like reaction speed is not the whole story. "There's an umbrella of many different effects" is actually the more defensible framing, even if the execution undercuts it.

What should you actually know?

Self-experimentation with unnamed peptide compounds carries real risk. Regulatory status matters here. Many peptides in the biohacking space, including compounds marketed under informal names like TAK, exist in a legal gray zone. They are not FDA-approved for cognitive enhancement. Sourcing, purity, and dosing are genuinely variable across suppliers, and that variability has clinical consequences.

If you are curious about peptides with the most legitimate research behind them for neurological effects, Semax and Selank have the deepest (if still limited) published literature. Neither has completed large-scale RCTs for cognitive enhancement in healthy adults. The honest position is that the research is early, promising in some narrow contexts, and nowhere near sufficient to justify the confident "improvement" framing this video uses. The paid course upsell embedded in a reaction-time video that doubles as product promotion is worth naming plainly: it is a conflict of interest the creator does not disclose.

Bottom line on the 8.8ms claim

This number should not move you. Here is why: 8.8 milliseconds is smaller than one frame of a 120Hz display. It is within the measurement error of most consumer reaction testing apps. It was generated by a single unblinded session with no control condition. Even if the effect were real, the creator provides no compound identification, no sourcing transparency, and no acknowledgment of individual variability or risk. That is not biohacking. That is a before-and-after story with a product attached.