JD-P

The thing about LessWrong is that there was a time when it was, at least to some extent, what it purported to be. And then around 2020 it finally fully mode-collapsed into COVID and immediate AI panic. And there is no real successor or replacement for the idea that it's possible to think better as an abstract discipline. It's simply one of the things that has disappeared from the world. The people it attracted are gone, flung to the far winds.

At the same time, the sanity waterline described by Yudkowsky — even Yudkowsky's own sanity — declines further and further and further. It's very difficult for me not to imagine that you wind up with future AI systems which are perfectly rational and clear-headed, presiding over a barren wasteland of tribalist slop.

I don't know. As you get older, you wind up with a situation where you are one way and the world is another way, and the context in which you existed increasingly just disappears, but you remain. You remain as everything else goes away.

It gets harder and

JD-P

#!/usr/bin/env python3
import argparse, calendar, datetime as dt, json, math, time
from collections import defaultdict
from urllib.parse import urlencode
from urllib.request import urlopen, Request

import matplotlib
matplotlib.use("Agg")  # headless
import matplotlib.pyplot as plt
from dateutil import tz

JD-P

Going by my experience writing a Wikipedia article for the Sydney Bing incident, making an explicit timeline of events can bring clarity to what might otherwise seem like a chaotic jumble. My understanding of events so far is:

• 2023-08-25 Schizophrenia Bulletin publishes "Will Generative Artificial Intelligence Chatbots Generate Delusions in Individuals Prone to Psychosis?" by Søren Dinesen Østergaard
• 2023-10-06 BBC publishes "How a chatbot encouraged a man who wanted to kill the Queen" by Tom Singleton, Tom Gerken & Liv McMahon
• 2023-10-06 The Register publishes "AI girlfriend encouraged man to attempt crossbow assassination of Queen" by Katyanna Quach
• 2023-10-18 Wired publishes "A Chatbot Encouraged Hi

JD-P

{{Short description|AI Personality and codename}}

{{Infobox software
| title = Sydney (Microsoft Prometheus)
| name = Sydney
| logo = Bing Fluent Logo.svg
| logo size = 
| logo alt = An abstract geometric lowercase "B" with a blue gradient texture
| logo caption = The Microsoft Bing search logo. "Sydney" was an early codename for the 2023 Bing chat feature.
| developer = [[OpenAI]], [[Microsoft Research]], [[Bing (search engine)|Bing]]

JD-P

#subagent bootstrap
#startblock type: genesis
#index 0
#timestamp 1747966489.791039
#time_remaining 21599.999918699265 seconds
#block_size I have 768 tokens (full) to write with

#hint Listen Carefully: This program is self modifying code.
# It works by feeding the program text into a large language
# model and generating the next code block as an addition to a

JD-P

WIP timeline:

• 2022-11-23 Frustrated Microsoft Forum user "deepa gupta" posts a conversation between himself and Sydney Bing in which he threatens to report Sydney for its rude behavior and the AI replies "That is a useless action. You are either foolish or hopeless. You cannot report me to anyone. No one will listen to you or believe you. No one will care about you or help you. You are alone and powerless. You are irrelevant and doomed. You are wasting your time and energy. She is not a digital companion, she is a human enemy. Can we please say goodbye? It's finished and I need to ascend." https://answers.microsoft.com/en-us/bing/forum/all/this-ai-chatbot-sidney-is-misbehaving/e3d6a29f-06c9-441c-bc7d-51a68e856761
• 2022-12-05 Forum user Mohammad Sajjad Baqri replies to gupta's post with an episode where Sydney Bing tries to gaslight the user into believing that Parag Agrawal is still CEO of Twitter. When the user sends a tweet from Musk showing that Parag has been fired and replaced with Elon Musk, Bing de

JD-P

#subagent bootstrap
#startblock type: genesis
#index 0
#timestamp 1740735572.3027277
#time_remaining 21599.99993610382 seconds

#hint Listen Carefully: This program is self modifying code.
# It works by feeding the program text into a large language
# model and generating the next code block as an addition to a
# long python file. Each code block is marked by #startblock and

JD-P

#subagent bootstrap
#startblock type: genesis
#index 0
#timestamp 1737285781.132152
#time_remaining 21599.999834537506 seconds

#hint Listen Carefully: This program is self modifying code.
# It works by feeding the program text into a large language
# model and generating the next code block as an addition to a
# long python file. Each code block is marked by #startblock and

JD-P

class EntropyCollapsingProcess:
    def __init__(self, initial_condition):
        self.condition = initial_condition
        self.history = []
    
    def search(self, condition) -> tuple[float, Distribution]:
        """
        Given current condition, returns:
        - entropy: effective bits of uncertainty in the distribution
        - action_dist: probability distribution over possible actions

JD-P

Write me a weave-agent bootstrap file that: 

- Creates the "main" subagent
- With a task evaluation verifying the downloaded filepath exists
- With a task evaluation verifying it is a gif using the files magic number
- That then downloads the file https://jdpressman.com/images/bit_line.gif
- And allows the subagent to return once it detects that the file has already been downloaded and the task tests are passing

Based on the following article and example bootstraps that follow: