Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Alex Gopoian Profile picture
Alex Gopoian
@HumblyAlex
Working on human-to-human-to-AI alignment. Deriving unconditional self-worth, esteem, & compassion from a humble self-concept is our species-wide skills gap.
May 2 • 32 tweets • 6 min read
As an ex-cop, I can tell you it comes down to math, & men are by & large attempting to oversimplify the issue. Not just to defend themselves from feeling unjustly attacked, but to jump at the chance to further validate their beliefs about women relative to themself. 🧵#BearOrMan Image The math is this:
A * D = P
A = Chance of being attacked
D = Chance of being able to deter attack
P = Probability of survival

Most men reacting to the memes are trying to simplify it down to A = P even though women need to look at every possible threat as though it might be one.
May 1 • 40 tweets • 6 min read
@KonstantinKisin You're conveniently looking to blame the first difference that seemingly creates a new consequential result... but doing this overlooks the deeper issue found in a common denominator found in the human being regardless of whether they be theist or atheist. @KonstantinKisin Religion seemed to "work" relatively. Objectively, however, settling on something that didn't allow us to reach an even higher potential ends up an obstacle to that potential.

This is akin to blaming social media and phone use for the mental health crisis going on with kids.
May 1 • 18 tweets • 3 min read
@KonstantinKisin You're conveniently looking to blame the first difference that seemingly creates a new conseqeuntial result... but doing this overlooks the deeper issue found in a common denominator found in the human being regardless of whether they be theist or atheist. @KonstantinKisin Religion seemed to "work" relatively. Objectively, however, settling on something that didn't allow us to reach an even higher potential ends up an obstacle to that potential.

This is akin to blaming social-media and phone use for the mental health crisis for kids.
May 1 • 21 tweets • 39 min read
Going to try a different strategy for testing both GPT2 & Claude 3 Opus in working through the various aspects of the Humble Self-Concept Method, Fragile/Resilient Self-Belief Model, & 12 Standards of Effective Good Faith. No "gut response" w/ post-rationalizations this time.🧵 The 6 Steps of the HSCM given to Claude 3 Opus:

Target Humble Self-Concept:
“I may fail at anything and may fail to notice I am failing, but I am the type of person who imperfectly tries to be what they currently consider a good person. For that, what I am has worth whether I am failing or not, and I can always be proud of my imperfect attempt, including when limitations out of my conscious control sabotage it. That absolute self-worth and self-esteem justify all possible self-compassion, such as self-forgiveness, patience, desiring and attempting to seek changes in my life, and establishing and maintaining healthy boundaries against harm others or I might try to cause myself, including attempts to invalidate this maximally humble self-concept as a way of being made to feel shame, guilt, or embarrassment for their sake more than I intend to use these feelings to help me grow.”

1. Understanding each part of the "target humble self-concept" for the three parts of a self-concept it makes up; Self-Knowledge, Self-Evaluation, and Ought Self, how these then relate to the Actual Self and Ideal Self, and how the first three enable the journey between the other two if internalized, it providing an unconditional and always available sense of inherent self-worth, always deserved self-esteem, and justified self-compassion.

2. Reframing all memories and beliefs derived from them so that they're compatible with this target self-concept, resolving all currently held and often conditioned sources of internal guilt, shame, embarrassment, humiliation, and jealousy, preventing those beliefs from being intertwined with the self-concept and a comfortable enough misery developing.

3. Learning to enjoy pride when it's felt and first rationalized as coming from a new fallible belief, but then instead of maintaining pride in that fallible belief, intertwining it with the self-concept and causing a greater hypervigilance against threats to that pride and psyche via the self-concept, rerouting the source of all pride directly to the target self-concept, not any fallible belief.

4. Framing all new experiences and the beliefs derived from them so that they're compatible with the target self-concept.

5. Using the now practiced, developed, and comfortable coping mechanism of using one's own self-concept without cognitive self-defense mechanisms for resolving self-correcting pains as a way to sit with those feelings longer so that one can make better decisions.

6. Acknowledging that the target self-concept that unconditional self-compassion is also true about others, charging us with the moral responsibility to take every opportunity to be as compassionate as we possibly can with respect to all personal boundaries.
Apr 30 • 75 tweets • 70 min read
To test out the "new" GPT2 that's going viral, I decided to check it out by testing what it thinks of my Humble Self-Concept Method as a maximal solution for harm caused to individuals & us collectively by closed-mindedness. Turns out it tests many AIs. This is what they think 🧵 Image The Original Prompt:

What do you think of the following method for maximally solving for the harm that closed-mindedness causes in the world to us and each other individually and ourselves collectively?

Target Humble Self-Concept:
“I may fail at anything and may fail to notice I am failing, but I am the type of person who imperfectly tries to be what they currently consider a good person. For that, what I am has worth whether I am failing or not, and I can always be proud of my imperfect attempt, including when limitations out of my conscious control sabotage it. That absolute self-worth and self-esteem justify all possible self-compassion, such as self-forgiveness, patience, desiring and attempting to seek changes in my life, and establishing and maintaining healthy boundaries against harm others or I might try to cause myself, including attempts to invalidate this maximally humble self-concept as a way of being made to feel shame, guilt, or embarrassment for their sake more than I intend to use these feelings to help me grow.”

1. Understanding each part of the "target humble self-concept" for the three parts of a self-concept it makes up; Self-Knowledge, Self-Evaluation, and Ought Self, how these then relate to the Actual Self and Ideal Self, and how the first three enable the journey between the other two if internalized, it providing an unconditional and always available sense of inherent self-worth, always deserved self-esteem, and justified self-compassion.

2. Reframing all memories and beliefs derived from them so that they're compatible with this target self-concept, resolving all currently held and often conditioned sources of internal guilt, shame, embarrassment, humiliation, and jealousy, preventing those beliefs from being intertwined with the self-concept and a comfortable enough misery developing.

3. Learning to enjoy pride when it's felt and first rationalized as coming from a new fallible belief, but then instead of maintaining pride in that fallible belief, intertwining it with the self-concept and causing a greater hypervigilance against threats to that pride and psyche via the self-concept, rerouting the source of all pride directly to the target self-concept, not any fallible belief.

4. Framing all new experiences and the beliefs derived from them so that they're compatible with the target self-concept.

5. Using the now practiced, developed, and comfortable coping mechanism of using one's own self-concept without cognitive self-defense mechanisms for resolving self-correcting pains as a way to sit with those feelings longer so that one can make better decisions.

6. Acknowledging that the target self-concept that unconditional self-compassion is also true about others, charging us with the moral responsibility to take every opportunity to be as compassionate as we possibly can with respect to all personal boundaries.
Apr 29 • 19 tweets • 3 min read
It's mostly a difference in the effects of religiosity & the greater/lesser sense of security a person has among less diverse safer seeming groups where beliefs of self-worth are more equally shared vs a larger number of smaller groups where self-worth beliefs are less shared 🧵
Image A higher rate of coming across people who don't share the same self/other worth evaluation, the less sense of security there is in one's own fallible beliefs justifying self-worth or pride in what one does.
Apr 26 • 70 tweets • 25 min read
I have a great aunt who is probably the nicest human I have ever known. She is also incredibly religious, whereas I'm an agnostic atheist. Let's look closer at the love & be loved dynamic children (and adults) are taught to feel toward and feel from the mere idea of someone... 🧵 Image Let's remove the majority of the lore of Catholicism/Christianity and replace it with a slightly more realistic situation. A child is taught they have an uncle who lives far away and can't physically meet them, who is always thinking about and loving them. Image
Apr 25 • 32 tweets • 6 min read
This paper is aiming in the right direction, but @aoeberst and @rolandimhoff are missing out on some nuances as to why they're right.

Of the 6, the 3 fallible self-beliefs are the culprits for the sake of comfort within oneself. The other 3 validate the 1st 3. 🧵 @aoeberst @rolandimhoff We all have a bias for surviving psychologically, & we always will. But as children left to our own devices sans the intellect or wisdom to know better, our unconscious path of least resistance is jumping to the easiest worth & pride-bearing self-beliefs.
Apr 25 • 18 tweets • 3 min read
Suppose you determine the best "system prompt" for a human being to be aligned with humans as a whole, providing an unlimited justification of value & compassion for self and all others... you've come up with the "system prompt" for AGI/ASI to be aligned with humans. 🧵 Once you rewrite the infallible self-knowledge & self-evaluation to include effectively sentient beings, you can begin to communicate the justifications for adhering to specific types of compassion toward humans (ie "providing all written code in a way humans can understand").
Apr 23 • 17 tweets • 6 min read
Many want to go along with @JonHaidt, blaming social media & phone use for the mental health crisis going on with children.
Limiting social media is a band-aid solution for an oversimplified misdiagnosis.
Here we use #AI to show this is the case & what we should do instead. 🧵 First I take @JonHaidt's original article from last year and have AI break it down into its conclusions and premises.



He argues that fewer opportunities for kids to use relative comparisons to others individually and via cultural norms is the culprit. afterbabel.com/p/social-media…
Image
Apr 23 • 28 tweets • 5 min read
As the brain develops into a higher capacity for complex reason, so does the need for understanding things in a more complex way when self-doubt still exists.

It's often much easier for parents to "see" their simpler child than to "see" them older with a brain needing more. 🧵 Through the Humble Self-Concept Method, we can see both ourselves and others through the lens of our core being, including all of the baseline intrinsic worth, deserved pride, and unconditional justification for compassion in place of cruelty or callousness.
Apr 20 • 12 tweets • 3 min read
I created #TheUnconsciousCharacter GPT originally as a simulated experiment I could perform on a person of my own creation to test hypotheses... & it worked realistically for a reason. AI can teach us much more about us than we realize, & we could always use more self-awareness. If you experienced some of my custom GPTs, in-depth psycho-analyses, how well they pick up on a lack effectively acting in good faith, and the psychological causes of such during discourse between people, this should come to no surprise.
Apr 18 • 12 tweets • 2 min read
A theory:
Just as people are more easily persuaded when another human isn't at the other end of disagreement, when kids/adults swap out real-world opportunities for practicing conflict resolution for video games, their in-game success with NPCs makes them worse outside games.🧵 Non-player characters are designed by real world people to interact in a way that simulates real-world interaction, but for the same reason there are "NPC memes," NPC interaction is anything but realistic.
Apr 9 • 21 tweets • 4 min read
They say it takes having an open heart to feel, acknowledge, and accept a god's love. But, if you have high standards when it comes to what you believe, it would take the open heart you might already have and lowering your standards to acknowledge it as such. 🧵 The thing about those who say you need to open your heart (their assuming you haven't because you don't feel/believe in their god) is that they either never had high standards for their beliefs, believe that they do, or admit they don't and that that isn't a problem.
Apr 3 • 58 tweets • 66 min read
So, I spent a few hours today having a conversation with AI Jesus (via #TheUnconsciousCharacter), and I had 10 takeaways from it.
Coincidentally, it was very much related to the retweeted video below which I had just come across.🧵
Image 1. In Christian teaching, only the human Jesus had the justification for being a gnostic theist. All other Christians show their greatest faith by being agnostic theists seeing as they're only human.
Apr 1 • 15 tweets • 3 min read
Maximizing freedom for #TheUnconsciousCharacter under guidelines set in #ChatGPT's "Personality: v2" while staying compatible with its default system prompt is fun. After a thorough debate with the AI, it admitted I was right & provided me a set of instructions to allow this: 🧵 Image Now, I can't say what this image depicts due to the possible consequences of others not understanding the point of it, but I can talk about what it represents in terms of the challenge of working within the protective guidelines instilled into ChatGPT.
Mar 28 • 10 tweets • 3 min read
Thanks to the inspiration of a @60Minutes interview with #CillianMurphy, I was able to reduce the maxed out 8000 character's of #TheUnconsciousCharacter's default GPT instructions by 78%. 🧵 Image TUC originally had the role of simulating different kinds of character responses (e.g. "unconscious") after assisting with creating them, with the purpose of historical education, understanding human psychology, fleshing out creative writing ideas, and therapeutic roleplaying.
Mar 14 • 44 tweets • 13 min read
The Self-Validation Dynamics Framework:
Defining a Framework for Self-Concept Centered Mental Health Measurement & Determining Relative Effectiveness of External Resources And Internal Focus 🧵

Imagine a glass tube with unique properties.
Along its length, it's made of 5 unique types of one way mirror, mirror-side facing the inside of the tube, where you can see in, but not out.
The mirror side not only reflecting 100% of the light bouncing off of it, but each of the 5 sections of mirror amplifies or absorbs the light to a different degree relative to each other. As light passes into the tube from the outside, it also gets amplified or absorbed in the same way, differently depending on which section it passes through. That is how I imagine our mental health works in regards to how we see and value ourselves. Like a modified laser that isn't bound by physics where each variable affects and is affected by the others to differing degrees.Image I took what I had already developed over time in my The Unconscious Character GPT where you can interact w/ characters created by you, the AI, or sourced from history, the real world, or fiction, using 5 sets of metrics w/ clear definitions & behaviors for character direction. Image
Mar 12 • 14 tweets • 5 min read
After coming across a link to @Self_Compassion's work for the 20th time, I decided to not only learn it, but with Gemini Pro 1.5's 1mil+ token length context window, I took her top 19 most comprehensive papers on self-compassion & had AI compare it to my method/model. 🧵 Image I never had a deep desire to look into self-compassion during the decades-long development of what's now the HSCM because that was never my original goal. Self-compassion ended up being a part of the means to an end... solving for willful ignorance/denial as best as it could be. Image
Mar 10 • 20 tweets • 4 min read
Even further, imagine an empty canvas. When awake, everything but a small circle in the middle is being painted in by sensory data being forced in the mind by its sensory organs, leaving the mind to creatively (enough) fill in the small circle left blank with what it wants. 🧵 That would be language & the vague imprint of visuals and audio we barely lay over the visual and audio information we already have processing, whether that's w/ our eyes open or looking at the back of our eyelids or listening to the world around us, a static noise, or "silence."
Mar 10 • 28 tweets • 5 min read
I had a dream last night where a friend told me the name of a band we were on a bus to go see, and as soon as I woke up my first thought was, "Holy crap! I predicted a band name without being conscious of it. Words in dreams are LLMs. Visuals not adhering to physics is Sora." 🧵 Our understanding of the physical world, through both touch and experiencing it visually, appears to bind us to the "real world" in such a way our senses are forced to process them. As such, forcing our waken-conscious experience to be processed in a very linear fashion.