Hi @RobTuck. I normally stick to science here, but you touched on philosophy and seemed highly invested in the consistency of your reasoning process—so here we go!
First, I think this is a philosophical claim about limits of knowledge—not a scientific claim. More on that below.
Second, impossibility is a very low bar. You mentioned “decision literature”, so I imagine that you already know this. For example, it is not impossible that a sapient ballon who refers to itself as “Robert” resides on a planet outside of our solar system. Or, to use Bertrand Russell’s famous example, it’s not impossible that a teapot (too small to be revealed even by our most powerful telescopes) is currently revolving around the sun in an elliptical orbit.
In other words, possibility is the bare minimum for a claim. It’s not interesting on its own: the list of possible truths is effectively unbounded (i.e. nearly infinite). The interesting questions are around what’s probable, which is a much much smaller set than what’s possible. To quote Kahneman and Tversky’s foundational work establishing the modern sub-field of biases and heuristics, we are always making judgements under uncertainty.
It’s not impossible that rapamycin instigates a catastrophic epigenetic change in parents that only manifests after ten generations. But, to reiterate: possibility is the bare minimum for a claim. All true claims are possible, but most possible claims are not true.
As you pointed out, none of us know how you navigate “choice points based on [your] adjudication of a complex set of facts and principles”, but structuring your thoughts with “It’s not impossible that…” is a red flag that I invite you to reflect on.
So, what would a revised, well-structured (i.e. falsifiable), and scientific version of this claim look like? Something like this:
We do not currently have enough evidence to assess the probability of long-term harms in early adopters of rapamycin. In order to confidently rule out an unknown long-term harm, I would need to see [specific collection of evidence] that showed [success conditions] and didn’t show [failure conditions].
Structurally, this avoids the central problem with all unbounded negative claims, where the implicit demand for comprehensive evidence means that no amount of actual evidence can be sufficient (e.g. you can’t search every planet in the universe for that sapient balloon named Robert).
Practically, this puts the burden of proof (rightly) on you to argue why we should be concerned about unknown long-term harms in the first place. This is the step where you promote your claim from possible (uninteresting) to probable (interesting). Here, establishing a precedent of other drugs with late-arriving harms might be enough of a precautionary principle to warrant attention.
But this, then, invites a comparison of the data available at the release of those drugs vs. rapamycin, exploration into how well the mechanism of action was understood vs. rapamycin, how much preclinical data was available, etc.
Structuring a claim this way also means you have to specify what evidence would be sufficient to convince you. Which means the claim is now properly falsifiable. So, when @RapAdmin mentioned that:
this provides, not the gold-standard, but a substantial evidence base to interrogate for signals of harm. If you didn’t already have that background information, then finding out that a population of people have been consistently taking higher doses of rapamycin without unknown/unexpected harms should be a relief. Depending on your risk aversiveness, you may want long-term, randomized clinical trials in diverse populations, across multiple ages, exploring the dosing schedules and amounts used by early adopters. But, regardless, thanks to the structure of your claim—you would at least be able to update your confidence along the way.
And, crucially, well-structured, falsifiable claims make conversation so much more constructive. We’re all walking around with false beliefs in our heads—and I think engaging with the ideas of others (whether through reading or conversation) remains the best way to root them out.
