SilverFlame
  • Computer scientist
  • Fairly deep experience with self-programming and modification of intuition/reflexes
  • Personal jargon/nomenclature was developed in isolation and seldom matches what other people use

Wiki Contributions

Comments

  • Learning about the trigger conditions for serotonin, oxytocin, dopamine, and cortisol, which allowed for more direct optimization away from cortisol activations

This idea started when I read this article I was pointed at by a coworker in 2020: The DOCS Happiness Model. I then did some naturalist studies with that framing in mind, and managed to reduce cortisol activations that I considered "unhelpful" by a significant degree. I consider this of high value to people who have enough control over their environment to meaningfully optimize against cortisol triggers.

  • Using method acting and other mimicry skills to more quickly learn from experts I was already trying to learn from

This was mostly learned via self-experimentation.

This is a large part of what I call my "skill stealing" skill tree, which nowadays mainly focuses on training an IFS "voice" that possesses knowledge of the skill or skill set in question. The stronger forms of these techniques tend to eat a lot of processing cycles and make it hard to maintain other parts of a "self image" while you use them, so be wary of that pitfall.

If you do want to pursue it, remember to focus on aligning as many parts of your thought process in that field to the expert's thought process as seems appropriate instead of just becoming able to sound like them. There are a lot of layers and details to be mastered in this process, but even lesser forms can start showing value quickly.

  • Applying operating system architecture knowledge to my internal thinking patterns to allow more efficient multithreading and context switching

This was mostly learned via self-experimentation.

This is performed by analyzing where there seems to be bottlenecks in my personal processing speed, and then doing some tests to see if I can nudge things towards a slightly different architecture to reduce the constraint. Which changes are needed and when seems to be pretty individual-specific, but here's some things I did:

  • Practice switching between commonly-used headspaces to make such transitions more reflexive (and thus cheaper in both energy and time)
  • Train a "scheduler" and figure out how to let it cut off trains of thought that aren't a priority at the moment (there are many pitfalls to doing this poorly, approach carefully)
  • Start grouping my IFS "skillset voices" into semi-specialized "circles" I can switch between to partition which ones are "active" at different times, which saved processing resources during active calculations and reduced signal noise during queries; picture having different predefined parties to choose from in an RPG

In general, your guiding creed should be "know your constraints and know your capabilities".

The mimicry and OS architecture applications have borne a lot of fruit over the years, but they both can take time to yield their first fruit, even if they are a good fit for your setup. The mimicry tactics are probably useful to anyone who wants to benefit from cooperative experts, but the OS tactic doesn't seem as widely useful.

  • Learning about the trigger conditions for serotonin, oxytocin, dopamine, and cortisol, which allowed for more direct optimization away from cortisol activations
  • Using method acting and other mimicry skills to more quickly learn from experts I was already trying to learn from
  • Applying operating system architecture knowledge to my internal thinking patterns to allow more efficient multithreading and context switching

The two failure modes I observe most often are not exclusive to rationality, but might still be helpful to consider.

  1. Over-reliance on System 2 processing for improvement and a failure to make useful and/or intentional adjustments to System 1 processing
  2. Failing to base value estimations upon changes you can actually cause in reality, often focusing upon "virtual" value categories instead of the ones you might systemically prefer (this is best presented in LoganStrohl's How to Hug the Query)

The decision was generated by my intuition since I've done the math on this question before, but it did not draw from a specific "gut feeling" beyond me querying the heavily-programmed intuition for a response with the appropriate inputs.

Your question has raised to mind some specific deviations of my perspective I have not explicitly mentioned yet:

  • I spent a large amount of time tracing what virtues I value and what sorts of "value" I care about, and afterwards have spent 5-ish years using that knowledge to "automate" calculations that use such information as input by training my intuition to do as much of the process as is reasonable
    • I know what my value categories are (even if I don't usually share the full list) and why they're on the list (and why some things aren't on the list)
    • My "decision engine" is trained to be capable of adding "research X to improve confidence" options when making decisions
      • If time or resources demand an immediate decision, then I will make a call based on the estimates I can make with minimal hesitation
    • This system is actively maintained
  • I do not consider lives "priceless", I will perform some sort of valuation if they are relevant to a calculation
    • An individual is valued via my estimates of their replacement cost, which can sometimes be alarmingly high in the case of unique individuals
    • Groups I can't easily gather data on are estimated for using intuition-driven distributions of my expectations for density of people capable of gathering/using influence and of awful people
    • My estimations and their underlying metrics are generally kept internal and subject to change because I find it socially detrimental to discuss such things without a pressing need being present
  • Two "value categories" I track are "allows timelines where Super Good Things happen" and "allows timelines where Super Bad Things happen"
    • These categories have some of the strongest weights in the list of categories
    • They specifically cover things I think would be Super Good/Bad to happen, either to myself or others
  • I estimate that skilled awful people having an unlimited lifespan would be a Super Bad Thing, therefore timelines that allow it are heavily weighted against
    • Awful people can convert "normal" people to expand the numbers of awful people, and given a lack of pressure even "average" people can trend towards being awful
    • The influence accumulation curves over time I have personally observed and estimated look to be exponential barring major external intervention and resource limitations, and currently the finite lifespan of humans forces the awful people to each deal with the slow-growth parts of their curves before hitting their stride

Ok.  So remember, your choices are:

  1.  Lock away the technology for some time
  2. Release it now

 

You are choosing to kill every living person because you hope that the next generation of humans is more moral/ethical/deserving of immortality than the present, but you get no ability to affect the outcome.

Even with this context, my calculations come out the same. It appears that our estimations of the value (and possibly sacred-ness) of lives are different, as well as our allocations of relative weights for such things. I don't know that I have anything further worth mentioning, and am satisfied with my presentation of the paths my process follows.

I'm not sure your position is coherent.  You, as a SWE, know that you can keep producing turing complete emulations and keep any possible software from the past working, with slight patches.  (for example, early game console games depended on UDB to work at all).

Source code and binary files would qualify as "immortal" by most definitions, but my experience using Linux and assisting in software rehosts has made me very dubious of the "immortality" of the software's usability.

Here's a brief summary of factors that contribute to that doubt:

  • Source code is usually not as portable as people think it is, and can be near-impossible to build correctly without access to sufficient documentation or copies of the original workspace(s)
  • Operating systems can be very picky about what executables they'll run, and executables also care a lot about what versions of libraries they want are present
  • There's a lot of architectures out there for workspaces, networks, and systems nowadays, and information about a lot of them is quietly being lost to brain drain and failures to document; some of that information can be near-impossible to re-acquire afterwards

 It's irrelevant if it isn't economically feasible to do so.

I do not consider economic infeasibility irrelevant when a problem can approach the scope of "a major corporation or government dogpiling the problem might have a 30% chance of solving it, and your reward will be nowhere near the price tag". It is possible that I am overestimating the feasibility of such rehosts after suffering through some painful rehost efforts, but that is an estimate from my intuition and thus there is little that discussion can achieve.

While I found your careful thought process here inspiring, the normal hypothetical assumption is to assume you have the authority to make the decision without any consequences or duty, and are immortal.  Meaning that none of these apply.

First, I make a point of asking those questions even in such a simplified context. I have spent a fair amount of time training my "option generator" and "decision processor" to embed such checklists to minimize the chances of easily-avoided outcomes slipping through. The answer to the first bullet point would easily calculate as "your role has no obligations either way", but the other two questions would still be relevant.

But, to specifically answer within your clarified framing and with the idea of my choice being the governing choice in all resulting timelines, I would currently choose to withhold the information/technology, and very likely would make use of my ability to "lock away" memories to properly control the information.

The rest of your response seems reasonable enough when using the assumption that software is immortal, so I have nothing worth saying about it beyond that.

Do you think that some future generation of humans (or AI replacements) will become immortal, with the treatments being widely available?

I do not estimate the probability to be zero, but other than that my estimation metrics do not have any meaningful data to report.

Assuming they do - remember, every software system humans have ever built already is immortal, so AIs will all have that property - what bounds the awfulness of future people but not the people alive right now?

First, I'm not sure I agree that software systems are immortal. I've encountered quite a few tools and programs that are nigh-impossible to use on modern computers without extensive layers of emulation, and I expect that problem to get worse over time.

Second, I mainly track three primary limitations on somebody's "maximum awfulness":

  • In a pre-immortality world, they have only a fixed amount of time to exert direct influence and spread awfulness that way
  • The "society" in which we operate exerts pressure on nearly everyone it encompasses, amplifying the effects of "favored" actions and reducing the effects of "unpopular" actions. This is a massive oversimplification of a very multi-pronged concept, but this isn't the right time to delve into this concept.
  • Nobody is alone in the "game", and there will almost always be someone else whose actions and influence exerts pressure on whatever a given person is trying to do, although the degree of this effect varies wildly.

If immortality enters the picture, the latter two bullet points will still apply, but I estimate that they would not be nearly as effective on their own. Given infinite time, awful people can spread their influence and create awful organizations, especially given that people I consider "awful" tend to more easily acquire influence than people I consider "good" (since they have fewer inhibitions and more willingness to disrespect boundaries), so that would suggest a strong indication towards imbalance in the long term.

Why do you think future people will be better people?

I don't necessarily think future people will be better people. I don't feel confident estimating how their "awfulness rating" would compare to current people, but if held at gunpoint I would estimate little to no change. I am curious what made you think that I held such an expectation, but you don't have to answer.

If you had some authority to affect the outcome - whether or not current people get to be immortal, or you can reserve the treatment for future people who don't exist yet - does your belief that future people will be better people justify this genocide of current people?

There would be several factors in a decision to use such authority:

  • If I gained the authority through a specific role or duty, what would the expectations of that role or duty suggest I should do? This would be a calculation in its own right, but this should be a sufficient summary.
  • Do I expect my choice to prevent the spread of immortality to be meaningful long-term? The sub-questions here would look like "If I don't allow the spread, will someone else get to make a similar choice later?"
  • Is this the right time to make the decision? (I often recommend people ask this question during important decision-making)

The first and third factors I feel are self-explanatory, but I will talk a bit more on the second factor.

I would expect others given the same decision to not necessarily make the same choice, so by most statistical distributions even one or two other people facing the same decision would greatly increase my estimation of "likelihood that someone else chooses to hit the 'immortality button'". Therefore, if I expect the chance of "someone else chooses to press the button" to be "likely", I would then have to calculate further on how much I trusted the others I expected to be making such decisions. If I expected awful people to have the opportunity to choose whether to press the button, I would favor pressing it under my own control and circumstances, but if I expected good people to be my "competition", I would likely refrain and let them pursue the matter themselves.

... does your belief that future people will be better people justify this genocide of current people?

I do not currently consider myself to have enough ability to influence the pursuit of immortality, but I have consciously chosen to prioritize other things. I also prefer to frame such matters in the case of "how much change from the expected outcome can you achieve" rather than focusing upon all the perceived badness of the expected outcome. I've found such framing to be more efficient and stabilizing in my work as a software engineer.

 

As a general note to wrap things up, I prefer to avoid exerting major influence on matters where I do not feel strongly. I find that this tends to reduce "backsplash" from such exertions and shows respect for boundaries of people in general. As the topic of pursuing immortality is clearly a strong interest of many people and it is not a strong interest of mine, I tend to refrain from taking action more overt than being willing to discuss my perspective.

First, a brief summary of my personal stance on immortality:

- Escaping the effects of aging for myself does not currently rate highly on my "satisfying my core desires" metrics at the moment

- Improving my resilience to random chances of dying rates as a medium priority on said metrics, but that puts it in the midst of a decently large group of objectives

- If immortality becomes widely available, we will lose the current guarantee that "awful people will eventually die", which greatly increases the upper bounds of the awfulness they can spread

- Personal growth can achieve a lot, but there's also parts of your "self" that can be near-impossible to get rid of, and I've noticed they tend to accumulate over time. It isn't too hard to extrapolate from there and expect a future where things have changed so much that the life you want to live just isn't possible anymore, and none of the options available are acceptable.

Some final notes:

- There are other maybe-impossible-maybe-not objectives I personally care more about that can be pursued (I am not ready to speak publicly on most of them)

- I place a decent amount of prioritization pressure to objectives that support a "duty" or "role" that I take up, when relevant, and according to my estimations my stance would change if I somehow took up a role where personal freedom from aging was required to fulfill the duty

- I do not care strongly enough to oppose non-"awful" (by my own definitions) people from pursuing immortality; my priorities mostly affect my own allocations of resources

- I mentioned in several places things I'm not willing to fight over, but I am somewhat willing to explain some aspects of my trains of thought. Note, however, that I am a somewhat private person and often elect silence over even acknowledging a boundary was approached.

1:15 with the use of some distraction and breathing techniques. Mid-20s male in decent health but asthma.

I remember pushing to 90 seconds at one point when experimenting with some body control techniques, but that was a couple years ago and I'd probably have to take some unhealthy measures to repeat that nowadays.

Circling back a few months later, I have some observations from trying out this idea:

  • I found myself tossing ideas to friends and acquaintances more often, which tended to improve my relationships with them somewhat
  • I noticed that some of the ideas I was preparing to hand off to someone else had glimmers of concepts I could use for other things, which had obvious benefits
  • I didn't notice any impact to my normal ideation/processing bandwidth as a result of the change in operating method
  • Sometimes ideas I handed off to someone else would circle back later and benefit one of my own projects, although I suspect the success rates for such second-order results will vary wildly

Overall, it seems to have been worth trying, and I'll probably keep it going.

Load More