When you try someone else's "broscience", you're not really experimenting with the unknown, so it's unlikely you're going to stumble into a "result". They know it doesn't work. If it did work, they'd have patented it and licensed it to Merck or Pfizer.
Choosing quackery is not experimenting.
I think your argument is that developing cancer treatments is really hard and the clinical trial process has problems. I can believe it. Do you think the mainstream epistemological process produces more false negatives or more false positives? What's the proportion of cases where a chemist finds a compound that really works, but the process incorrectly rules it out, vs cases where something that doesn't really do much still comes out the other end as a marketable drug?
I'm not in the industry, so what do I know? But I kind of doubt there are actual, effective treatments just sitting there unmonetized in Merck's basement library because the company is slow and the process makes things difficult. Especially the kinds of things that get mentioned in HN threads that don't require any chemistry research and are pretty straightforward to test, like fasting, meditation and yoga.
TBC I have no reason to believe that Big Pharma is deliberately withholding effective treatments it knows about so much as making (understandable) decisions not to investigate them in the first place. The total addressable market of a potential therapy is a central consideration for target selection-- you're roughly trying to maximize patient population times marginal willingness to pay over the standard of care. I don't think there's anything nefarious about this btw, you have to do what makes the greatest difference.
But that does mean that if you have a hyper-specific rare disease, the person who is willing to spend the most time thinking about it may very well be you. Or if you're living far from a major hospital and getting treated by a generalist with a heavy case load, you might be the most invested person within 500 miles of you, which is almost the same thing but better because you can still read the literature.
I wouldn't encourage the average patient to try doing rational drug design in their garage, but one could ask: "are there things that look promising that are still a year out from clinical trials?" and think about how to DIY some approximation to that.
At the extreme end of this spectrum, you have people like Beata Halassy who did just that, treating her own cancer with DIY viral therapy (https://www.nature.com/articles/d41586-024-03647-0, and please do note all the finger-wagging about how terrible and irresponsible of her it was to save her own life). Why did she have to do that instead of just going to a doctor? Because the route to the clinic is too slow. Why is the route to the clinic too slow? Because FDA has the institutional incentives that it's better that ten thousand patients die for lack of a cure than one die of quackery. Why's that? Because the FDA gets penalized for bad treatments but not for treatments that don't exist. But I say that dying of lack of a cure is not much better than dying of quackery, so we might as well minimize total deaths.
At the somewhat less extreme end, you have ideas like trying to treat GBM with Zika virus, which has a sketch of a mechanistic explanation and some support in animal models, but afaict no clinical trials yet (https://clinicaltrials.gov/search?cond=GBM&intr=Zika). Is this a cure for GBM? Complete BS? Something in between? I have no idea, which is kind of my point-- no one knows yet. Regulators probably aren't that jazzed about signing off on giving neurotropic viruses to immunocompromised patients without a lot of prior evidence that will give them cover for making that call if something goes wrong (which it totally might!). But an individual patient might look at that question with a different set of incentives.
Or consider psychiatric treatment of various mental illnesses. The best and most honest psychs I know will tell you that, past a certain point, responsiveness to a given drug is idiosyncratic and the state of the art is really just "try a bunch of stuff until something works", without much concern for hypotheses about underlying mechanisms. Is that rational medicine or bro science? Something in between, I think. And is it possible that there are behavioral protocols that help one particular schizophrenia patient to manage their quality of life better than the standard of care as defined for the entire population? Given that no one really agrees about what schizophrenia even is, this seems not totally impossible.
And then at the other end of the spectrum you basically have lifestyle interventions, as you note. Almost everyone agrees that certain of these are good for you, but some are really difficult to implement and adhere to. It seems reasonable to self-experiment with those things.
Elsewhere in the thread I argue why efficient market hypothesis arguments are unlikely to fully apply in this case.