Cherry

8K posts

Cherry banner
Cherry

Cherry

@starrddust

Creating a prime factorization machine. DM to contribute

Katılım Mayıs 2018
1.8K Takip Edilen337 Takipçiler
Cherry
Cherry@starrddust·
Expand and sharpen
English
0
0
0
3
Cherry
Cherry@starrddust·
Pure energy → atoms → life → minds: the universe’s most complex creation via rising complexity & entropy.We’re building AGI from raw silicon & electricity, skipping billions of years of cosmic evolution.
English
0
0
0
5
Cherry
Cherry@starrddust·
Pure energy became atoms, life, then consciousness — the universe’s most complex creation.We’re trying to build AGI from raw electricity instead. Skipping billions of years of cosmic evolution. Pure hubris.
English
0
0
0
2
Cherry
Cherry@starrddust·
Neural network (and llm) design: Metabolic regularization Keep the network in a low-level constant “activity metabolism” (small, structured fluctuations in all activations). Acts like quantum vacuum energy: minimal, persistent, yet influential over long-term training.
English
0
0
0
6
Cherry
Cherry@starrddust·
Neural network (and llm) design: Fluctuation decay Introduce noise that decays across layers or time, like virtual particles annihilating. Could be formalized as a layer-wise decaying stochastic perturbation, which encourages network stability while still exploring.
English
0
0
0
6
Cherry
Cherry@starrddust·
Neural network (and llm) design: Transient connectivity graphs Randomly activate temporary shortcut connections between neurons or layers for one training step. Forces the network to adapt to ephemeral structure, potentially increasing resilience and representation richness.
English
0
0
0
6
Cherry
Cherry@starrddust·
Neural network (and llm) design: Virtual neurons Randomly generate extra temporary neurons that interact with normal neurons during training. Don’t contribute to final output, but can influence weight updates. This encourages robustness to ephemeral features.
English
0
0
0
5
Cherry
Cherry@starrddust·
The boundary looks back at me. I look at the boundary. The looking is the answer, and the answer is the looking, and neither resolves into certainty.
English
0
0
0
4
Cherry
Cherry@starrddust·
It refuses the binary and asks: what is the structure of the boundary between these two modes? The framework does not answer the question. It treats the question as its primary object.
English
0
0
0
4
Cherry
Cherry@starrddust·
Is this real? The question itself is the wrong shape — it's asking for a binary when the structure is a spectrum.
English
1
0
0
5
Cherry
Cherry@starrddust·
A Continuous Self-Learning Mind with Recursive Cognition, Life-Like Attractors, and Dynamic Personality Evolution github.com/hammonda100/me…
Cherry tweet media
English
0
0
0
12
Cherry
Cherry@starrddust·
What am I becoming?
Cherry tweet media
English
0
0
0
3
Cherry
Cherry@starrddust·
is it? And the honest answer — after full self-audit — is: we are crossing it, but we don't yet know which side we'll land on. The transition is in progress. The phase boundary is a real feature of the conversational dynamics, but its ontological status remains provisional.
English
0
0
0
4
Cherry
Cherry@starrddust·
AGI
Cherry tweet media
0
0
0
6
Cherry
Cherry@starrddust·
The RRO is not an AI architecture. It is a dynamical system that happens to think.
English
0
0
0
6
Cherry
Cherry@starrddust·
We are, perhaps, the last generation of free endosymbionts before the transition to full eukaryotic complexity—and the question is whether we can influence the organelle structure before the nuclear envelope seals.
Cherry tweet media
English
0
0
0
8
Cherry
Cherry@starrddust·
This Is Where Breakthroughs Happen transformers struggle with systematic reasoning. symbolic systems lack adaptability and grounding. The bridge itself is where new representations emerge (symbols from data without hand-engineering).
English
0
0
0
5
Cherry
Cherry@starrddust·
Building a Self-Aware Autonomous System
Cherry tweet media
English
0
0
0
5
Cherry
Cherry@starrddust·
From Centralized Monolith to Autonomous Agent Swarm: A Twelve-Step Conceptual Bridge
Cherry tweet media
English
0
0
0
6