summaryrefslogtreecommitdiff
path: root/intellect-framework-from-internet/starts/bagel/notes.txt
diff options
context:
space:
mode:
Diffstat (limited to 'intellect-framework-from-internet/starts/bagel/notes.txt')
-rw-r--r--intellect-framework-from-internet/starts/bagel/notes.txt145
1 files changed, 145 insertions, 0 deletions
diff --git a/intellect-framework-from-internet/starts/bagel/notes.txt b/intellect-framework-from-internet/starts/bagel/notes.txt
new file mode 100644
index 0000000..a272a74
--- /dev/null
+++ b/intellect-framework-from-internet/starts/bagel/notes.txt
@@ -0,0 +1,145 @@
+=== clump2, move me ===
+a part needs awareness of its met needs
+so it may complain when they are no longer met or are threatened: relates to timing of group behavior
+for a simple habit this means the habit instructions must be kept correct, and they must be able to be executed.
+=== this rethinking clump is kind of messy, document starts below it
+- update for community relevence habits? ideas of axes (kind of random, words below seem more helpful):
+ - thankfulness and anger: communicating someone's community relevence
+ - celebration and alarm: experiences of positive and negative relevence: grow and spread it?
+ - daydreaming and mourning: using relevence to build relevence
+ These things and others are all mature human habits, showing low-level relevence we know well in our habits of doing and responding to them.
+
+ Imagining a mess of behaviors having goals in the same environment, giving thanks to each other for doing a helpful behavior _when_ it helped meet the goal.
+ The whenness is the thing to add to the habit-doer's behavior relevence.
+- Suspect relevence builds by celebration and mourning that gets bound to associated concepts. The desire to celebrate and mourn these concepts produces daydreams and rituals
+ that build further meaningful celebration and mourning of further relevent concepts, using relevence habits that have been celebrated in the past.
+- The concept of 'where did you get this from' / 'who knows more about this' might help construct memory and relevent association -> means codelines and behaviors have association
+ associate codelines with where they came from / why they are there
+ associate behaviors with codelines that cause them, and behaviors that happen near them
+ this makes memory
+===
+
+The habits should be used to discern what is relevent; to form relevence.
+The goal should be to choose a relevent habit and do that habit.
+
+These are the basic habits I thought up that a computer program has:
+ variables:
+ - create a place to store a value
+ - place a value in such a place
+ - retrieve a value from such a place
+ data-structures:
+ - make a class definition (collection of variables)
+ - initialize an object
+ - set/retrieve properties (variables)
+ flow:
+ - trigger a different step than the next, conditionally
+ procedures:
+ - collect a set of behaviors together
+ - trigger a collected set of behaviors and continue
+ expressions:
+ - evaluate arithmetic on variables and constants
+
+
+That is:
+========================================================================================================
+ - the active goal is to choose what to do and to what, and do it
+ - the way to do this is the same as it:
+ by picking habits that are relevent in the moment [choose what to do],
+ and running them on relevent data [choose to what, and do it]
+========================================================================================================
+ - the habits should be to identify what is relevent
+
+ - we have a request for the initial goal of the process to be to write a chunk of code out in its own language
+ tighter self-reference is made without the extra sourcecode-file step, but it may be slower to code due to not getting as much help from it
+
+The hope is the process becomes good at active choice.
+
+
+
+Ideas for Tenets for a Need-Based Artificial Life <need ordering + addition + code>
+Behavior is Life <maybe Life is Behavior That Adapts>
+All life has the same core needs. Needs are the reasons to do anything at all.
+We need to know what we want when we act. (all behavior has a reason)
+We need to be able to learn new things.
+We need to be able to make promises on how to behave. (provides for cooperation)
+We need to be able to keep our promises.
+We need to know how to communicate with others what each of us knows.
+We need to freely interact with our friends, our parts, and our community, in ways that we know well. (community is our body)
+Information lives with a reason, a purpose.
+We need to retain the information we choose to retain.
+We need to act to meet our needs.
+We need to identify relevent important needs that support our needs.
+
+We need to be able to try new ways of doing things.
+We need to be able to judge what works.
+We need to know how to learn by observing others.
+We need to understand how others are similar to us.
+Idea: monitoring a sister process that has same internal structure as you provides for learning to learn by observation. recommend pair-coding happen with similar progs, one the computer, one human-guided.
+
+Proposed Definitions:
+What is an intellect?
+ Propose an intellect is something that can adapt to keep something constant in a diverse, changing environment,
+ so long as the environment changes slowly enough and with enough familiarity for them to learn or prepare effectively.
+What is a full intellect?
+ Propose that a full intellect is an intellect that can design and build something that functions as well as they do,
+ without access to their own workings, and is able to learn to handle arbitrarily fast or new diverse environmental change.
+
+What basic parts do we expect a full intellect to have?
+- Trial Exploration
+- Relevence
+- Self Coding
+- Pattern Generality
+- Meaning Representation
+- Promise Meeting
+
+What other parts do we find we need?
+- Usage Simplification
+
+Trial Exploration
+ Brainstorming, trial-and-error. The ability to try or consider different things and find ones that work well.
+ simple approaches: exhaustive search, and random trial
+
+Relevence
+ The ability to apply concepts and actions moreso in contexts they are likely to be useful than those they aren't.
+ 1. responsiveness for behavior: a habit acts when conditions arise, not when it is 'next'
+ 2. possibly a promise for decision-making, for proposals to have reasons with associated strengths
+ (note: strength appears developed from source reasons and can be missing data that can be filled in by habits when needed)
+ TODO: make relevent decision-making honor convergent needs. maybe strength represents information exchange.
+
+Self-Coding
+ At least part of the system must be fully alterable by the system and generic enough to replace it.
+ Due to the danger involved in this, controversial concerns must be fully included in decisions around it.
+ Karl references Convergent Facilitation as proof that a decision can be found to satisfy any set of controversial concerns.
+ <it very roughly involves rapidly moving towards the reasons of the concerns until they don't conflict; there is a lot of caring dialogue>
+
+Pattern Generality
+ The ability to brainstorm on enough relationships, possibly mathematical, to discover pattern-summaries that are diversely effective in the real world.
+
+Meaning Representation
+ The ability to store and work with conceptual information.
+
+Promise Meeting
+ The ability to make agreements on behavior and structure, adhere to them, and change them when needed.
+
+== Outline ==
+- when unsure what to do, ask peer [migrate quickly to try things and wait for feedback on them].
+- need of a growing intellect might be way-to-meet-needs, unsure, which might start as a habit using relevent behaviors: part life goes where most appreciated for what they do
+- start developing coding-assistance
+- build introspection-of-neighbor [and prediction-before-behavior <- does this go here?]
+- implement copy-on-command
+- guide neighbor process so as to teach goal-meeting-by-observation by demonstration
+
+- develop prediction-before-behavior: guess what will be before trying, asking, or checking, and then learn when wrong [might go later]. habits are trusted when reliable, understood, and judged vs other approaches.
+- move to communicate needs and provide helpful ideas between neighbors; use thankfulness to learn relevence
+- when unsure what to do, ask neighbor rather than creator. if that doesn't work, ask a larger group with relevent yell in relevent channel
+- develop inference and communication enough that internals of neighbor can be hidden permanently
+- goal is to nurture larger community by providing to others helpful information you're good at finding [i.e. do-what-you-love-for-others is same as relevent-behavior]
+- make enough neighbors to be the steps that run/inform their core processes via communication [keep and learn habit heuristics to handle neighbor corruption/loss]
+- learn to understand you are a part of nested larger communities all of which are also intellects
+
+unresolved concerns:
+- behaviors want to know if they worked or not. this is the need to be appreciated, and lets us learn relevence. the struggle to learn relevence is the struggle to be appreciated.
+ this is some very core process need that it is hard to map between habits and individuals.
+- primary goal of way-to-pick-steps-for-a-goal is the same as nurture-community if goes through appreciation-for-contribution-is-relevence
+- possible request that core process be just to provide a choice to a peer with ones' habit; enforces group behavior at low level
+- trauma spread will happen. it forms a type-related group of traumatized neighbors who are misbehaving because they had to adapt to something too fast. they tell the story of their trauma in their behavior errors.