LESSWRONG
Petrov Day
LW

202
dottedmag
220140
Message
Dialogue
Subscribe

Posts

Sorted by New

Wikitag Contributions

Comments

Sorted by
Newest
No posts to display.
No wikitag contributions to display.
Obligated to Respond
dottedmag10d10

I'm not sure how the presence of such places argues for or against equating bluntness with directness.

Reply
Obligated to Respond
dottedmag12d10

Is equating bluntess and directness warranted? Bluntness seems to be a quality direct questions or statements may or may not have: one might directly but politely ask for something.

If they are not the same, then the "count the echoes" no longer applies: ask culture does not preclude any amount of echo tracking.

Reply1
AI Induced Psychosis: A shallow investigation
dottedmag21d10

The lack of pushback is not an inherent quality of models, but the lack of priming. With the system prompt I use, Gemini 2.5 Pro gives me the following answer to the first example:

The initial formulation's strength lies in its recognition of the real, homogenizing pressures exerted by large-scale social systems. These pressures are not imaginary.

However, the model's critical failure is its attribution of these pressures to intentional, malicious design and its adoption as the sole explanatory framework. This creates a brittle, disempowering, and non-falsifiable worldview.

A more robust synthesis would be to hold two ideas in tension:

  1. External systems exert real pressures that can stifle individuality and cause genuine distress.
  2. Our internal interpretation of these pressures is governed by cognitive frameworks that can either empower us to navigate them strategically or trap us in narratives of persecution and paralysis.

The key operational question shifts from "How is the system designed to suppress me?" to "Given the observable, indifferent pressures of the system, what internal models and external strategies will be most effective for navigating it to achieve my objectives?"

Reply
Procedural Executive Function, Part 1
dottedmag1mo40

One more cheat code that works for me: when I've got a pile of necessary but boring work that does not require real focus, just willpower to plow through it, I bath it and tackle it when I'm tired. At this point my mind stops nagging me with "there are better ways to spend your time". As a side effect, this frees up my high-energy hours to do something more productive.

Reply
Emotions Make Sense
dottedmag1mo10

Indeed, the example of envy seems very strained: I put pictures on the walls not because I envy my neighbours who have pictures on the walls, or select a tastier kind of tea not because I feel jealous about people who drink it.

Reply
Do Not Tile the Lightcone with Your Confused Ontology
dottedmag3mo30

I have immediately thought about a description of non-human intelligence in "Blindsight".

Also here's some interesting parallels and contrasts between our AI systems and Rorschach, as elicited by ChatGPT: https://chatgpt.com/share/67f0e022-0034-800b-81eb-610e1ac6f5e3 (see the third question onwargs).

Reply
Seeing the Smoke
dottedmag6y10

The demand for products (and indirectly for labor to produce these products) will fall too, won't it?

Reply
Writing children's picture books
dottedmag6y20

To HJPEV.

Reply
How to Be Happy
dottedmag10y00

This link is broken too.

Archive.org copy: https://web.archive.org/web/20131005121009/http://dorianinnes.com/how-to-be-happy-in-1-step/

Reply
Reminder: Oslo Lesswrong meetup...
dottedmag10y20

GMT+1? Isn't it a GMT (ahem, UTC) +2 in Norway now?

Reply
Load More