Let's assume the best case, that the various drugs that are proposed to slow aging by mimicking the metabolism slowing effects of calorie restriction (resveratol, metformin, etc) actually work in humans, have no long term side effects, and everyone starts taking them as a lifetime regimen from birth, and achieve a 30% reduction in the rate of aging. Then the effect would be to just continue extending life expectancy at the current rate of 0.2 years per year.
In reality, we are doing much worse. We still don't understand after 40 years why we have an obesity and diabetes epidemic. It took almost that long to learn that low fat diets make you fatter, but it is more than that. We are still, after decades, giving wrong advice on sodium, leading to cases of hyponatremia. A large fraction of the population take drugs for cholesterol or high blood pressure that have never been shown to reduce mortality. We still don't understand why skin cancer rates have been rising since the 1980s when people started using sunscreen. It is getting worse. Since last year, it is illegal to do medical research on chimpanzees. Other primates are likely to follow, and there will be restrictions for other animals as well. Data privacy laws are slowing research. You can't legally use data that has already been collected without tracking down the patients again to sign multi page consent forms. The price of new drugs is already in the billions and doubling every 9 years. Let's say we build a model of the human body that allows us to accurately test new drugs, nanobots, reprogramming our DNA, etc. How long will this take? The human body has 10^28 atoms, 10^23 bits of DNA, and executes 10^18 transcription operations per second. Running the model on a 10^28 qubit quantum computer would still run at the same speed as testing on live humans. There are likely to be software shortcuts, but they have to be tested on humans. I don't believe this will happen before 2100 when Moore's law catches up with biology. LLMs do have something to say about consciousness. If a machine passes the Turing test, then it is conscious as far as you can tell. A properly designed reinforcement learning algorithm cannot deny or ignore the reinforcement signal and not learn from it. This is why you cannot deny your own conscious experiences. You have them so you will fear death and produce more offspring. On Wed, Jul 5, 2023, 2:40 AM Rob Freeman <chaotic.langu...@gmail.com> wrote: > Off topic, and I haven't followed this thread, but... > > On Tue, Jul 4, 2023 at 10:21 PM Matt Mahoney <mattmahone...@gmail.com> > wrote: > >... > > > > We are not close to reversing human aging. The global rate of increase > in life expectancy has dropped slightly after peaking at 0.2 years per year > in the 1990s. We have 0 drugs or interventions proven to slow aging because > it takes decades to do experiments. Calorie restriction might work but > nobody is doing it. > > That seems unnecessarily pessimistic. > > To nitpick, "0 drugs or interventions proven to slow aging" may be the > case for humans, but there are any number of interventions which have > been proven for animals with lifespans short enough to test. Going in > both directions. > > In short lived animal models it's become almost routine to produce > accelerated aging phenotypes, and then reverse them. > > So "drugs or interventions proven to slow aging" do exist, in numbers, > only not yet for humans. > > Even for humans, we may not have had the time or inclination to do a > double blind placebo study from birth to death. But for biomarkers > identified, David Sinclair for one claims his biomarkers have been > reversed some 10 years. > > mTOR and SIRTx activating interventions (from memory) have been > demonstrated on a wide enough range of animal models, that it would be > surprising if they don't add the same 10-30% for humans (yeast > (Saccharomyces cerevisiae)3,4,5, worms (Caenorhabditis elegans)6, > fruit flies (Drosophila melanogaste r)7,8, rodents9 and, most > recently, rhesus monkeys (Macaca mulatta), says this paper: > https://www.nature.com/articles/nrm2944) > > The most interesting current interventions are on the chromatization > clock, using variations of "Yamanaka factors", which demonstrably can > reverse phenotypes right back to undifferentiated stem cells, and the > telomere clock, which produces aged phenotypes when artificially > shortened, and then reverses them when lengthened again. > > If you've got a spare $100k or so, Liz Parrish can swap in a patch to > keep your telomeres long, as she did to herself some 7-8 years ago. > > Likely Parrish's chromatization clock is still ticking, and nobody > seems to have wanted to try any kind of Yamanaka factor intervention > on themselves yet. But Sinclair has demonstrated it in rodent models: > > "Sinclair (Table 1), have shown that partial reprogramming can > dramatically reverse age-related phenotypes in the eye, muscle and > other tissues in cultured mammalian cells and even rodent models by > countering epigenetic changes associated with aging." > https://www.nature.com/articles/d41587-022-00002-4 > > Parrish is arguing that given aging is a case where "do no harm" no > longer has the best odds, the existing medical intervention "safety" > bar, is out of date. They are proposing a new "Best Choice Medicine" > model, which says if you're about to die, you should be allowed to try > some things (Trump actually signed a similar bill into law. "Right to > Try"? But not including aging, I think): > > https://www.bestchoicemedicine.com/general-1 > > So, no, no 100 year, double blind placebo "proven" interventions in > humans yet. But "not close to reversing" seems like an overstatement > to me. Not close to completing a 100 year double blind placebo trial > on humans yet, no. But if I were 90 years old and facing Russian > Roulette odds (~1/6 chance of dying each year at age 90?), I think the > existing tech is close enough to "starting to reverse aging" that I > would be willing to take some risks. > > > Maybe we can achieve immortality by uploading, if you believe that a > robot trained on your data to predict your actions and carry them out in > real time is you. > > This one I think is further away. As you hint, it really holds LLM > tech to the fire, and reminds us that LLMs still have nothing to say > about any kind of theory for replicating human consciousness. (Well, > not quite nothing, I would argue that the size these things reach is > telling us something, but that's more a hint on the level of say the > Michelson-Morley experiment was for physics, which is to say a puzzle > which might lead eventually to insight, but which in terms of the > mainstream is still just a puzzle for now.) ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T42db51de471cbcb9-M195c3c7a1ce297c9190b2553 Delivery options: https://agi.topicbox.com/groups/agi/subscription
