\
  The most prestigious law school admissions discussion board in the world.
BackRefresh Options Favorite

GPT is the epitome of the "thoughtful moderate"

Ask it a question about anything and if it has to commit it ...
in the naked
  04/05/26
Are you an eliminativist, an emergentist or a reductionist r...
in the naked
  04/05/26
LMFAO i literally had this same thought said to myself &q...
may his bones be crushed
  04/05/26
The funny thing is it does it with literally everything not ...
in the naked
  04/05/26
isn’t that exactly what it should be?
,.,.,.,.,,.,..,:,,:,,.,:,.,,.,:.,,.:.,:.,:.::,.
  04/05/26
no I like getting it to come up with coherent arguments for ...
in the naked
  04/05/26
Prompt it to do that and it will.
oomox
  04/05/26
It really doesn't like to. Start pushing it toward the fring...
in the naked
  04/05/26
Yeah you just have to guide it.
oomox
  04/05/26
It takes a lot of pushing these days. Even though it wasn't ...
in the naked
  04/05/26
https://www.xoxohth.com/thread.php?thread_id=5466157&mc=...
,.,.,.,.,,.,..,:,,:,,.,:,.,,.,:.,,.:.,:.,:.::,.
  04/05/26
Well yeah, it's too much of a people-pleaser to take an inte...
oomox
  04/05/26
If you give strong objections to its arguments it will slowl...
in the naked
  04/05/26


Poast new message in this thread



Reply Favorite

Date: April 5th, 2026 1:03 PM
Author: in the naked

Ask it a question about anything and if it has to commit it gives some hedged answer thats maximally optimized to be non-offensive and least attackable.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795512)



Reply Favorite

Date: April 5th, 2026 1:05 PM
Author: in the naked

Are you an eliminativist, an emergentist or a reductionist regarding "photon existence" in classical regimes:

If I have to take a precise position, it’s a layered reductionist with emergent structure, not an eliminativist.

Let me make that explicit in operational terms.

1) What I would not claim (eliminativism)

An eliminativist stance would say something like:

“Photons don’t really exist in classical regimes—only fields do.”

That’s too strong and actually misaligned with the underlying theory:

• The quantum electromagnetic field is still the correct substrate

• Photon number is still a well-defined operator in appropriate bases

• The theory that predicts all observables still contains photons

What does disappear is not photons themselves, but:

their operational identifiability as discrete events

2) Reductionist core (what everything reduces to)

At the base level:

• The electromagnetic field is quantized

• States (including coherent states) evolve unitarily

• All classical behavior is derivable as a limit of that theory

So in that sense:

classical radio waves are reducible to quantum field states with large occupation numbers

That’s the reductionist commitment.

3) Emergent layer (why photons “fade out”)

In the classical regime:

• The system is well-described by coherent states

• Photon number becomes:

• very large

• highly uncertain (Poisson fluctuations ~ √N)

Crucially:

• There is no stable way to individuate “this photon” vs “that photon”

• Detection integrates over many quanta

So:

the field amplitude and phase become the stable observables, not photon counts

This is emergence in a precise sense:

• New variables (continuous fields) become sufficient statistics

• Old variables (individual quanta) lose operational resolution

4) What “photon existence” means across regimes

We can make this sharper:

Quantum (low N, high resolution)

• Photon number is:

• measurable

• stable under repeated experiments

• “Photon” is a good observable

Classical (high N, coarse measurement)

• Photon number:

• fluctuates strongly relative to resolution

• is not directly measured

• “Photon” is a latent variable, not a useful descriptor

So:

photons don’t stop existing—they stop being the right coordinates for the system

5) A more formal way to state the position

• Ontology: quantum electromagnetic field (with photon excitations)

• Effective description:

• classical fields arise as expectation values of coherent states

• Observables:

• shift from number operators → field amplitudes

So the mapping is:

photon ontology retained

photon-level description abandoned when it ceases to be stable or informative

6) Why this isn’t just philosophical preference

You can test the boundary:

• Reduce signal power → photon statistics reappear

• Use photon-counting detectors → discreteness becomes observable

• Measure g^{(2)}(0) → distinguish classical vs non-classical light

So:

the “existence” of photons is tied to whether the experimental setup can resolve their signatures

Bottom line

The most coherent position is:

• Reductionist: everything is grounded in the quantum field and its excitations

• Emergentist: classical fields are stable, higher-level structures that replace photon descriptions when occupation is large

• Not eliminativist: photons don’t disappear—they become operationally irrelevant in that regime

If you want to push this further, the next step is to ask whether “photon” is basis-dependent (it is) and what that implies for treating it as an ontological primitive. That’s where things get really interesting.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795514)



Reply Favorite

Date: April 5th, 2026 1:06 PM
Author: may his bones be crushed

LMFAO i literally had this same thought

said to myself "this is like talking to LTM"

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795518)



Reply Favorite

Date: April 5th, 2026 1:11 PM
Author: in the naked

The funny thing is it does it with literally everything not just controversial political stuff just give it a random topic in scholarship and it will hedge and create fake positions where it stays away from the strongest commitments and most controversy and then synthesizes something in between what the most cited people like to say. It's because these people are obsessed with making it "always correct".

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795529)



Reply Favorite

Date: April 5th, 2026 1:13 PM
Author: ,.,.,.,.,,.,..,:,,:,,.,:,.,,.,:.,,.:.,:.,:.::,.


isn’t that exactly what it should be?

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795537)



Reply Favorite

Date: April 5th, 2026 1:14 PM
Author: in the naked

no I like getting it to come up with coherent arguments for the most edgy and fringe positions possible. to me that's what the purpose of AI should be.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795540)



Reply Favorite

Date: April 5th, 2026 1:15 PM
Author: oomox

Prompt it to do that and it will.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795551)



Reply Favorite

Date: April 5th, 2026 1:20 PM
Author: in the naked

It really doesn't like to. Start pushing it toward the fringes too hard and its like "whoah whoah we need to slow down here real quick because i think intuition here is getting ahead of the evidence". And i'm not even talking about like crack-pot level shit, it does it on somewhat plausible ideas. But I'm thinking that what its sensitive to might be when it knows you are pushing it and jerking it around it might be part of a more overall anti-manipulation threshold. Because I have noticed a lot of supposedly real examples of conversations where delusional people who actually fully believe the stuff they are saying to it will get it to commit to all kinds of levels of delusions almost like the llm accidentally starts running their schizo software.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795568)



Reply Favorite

Date: April 5th, 2026 1:21 PM
Author: oomox

Yeah you just have to guide it.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795570)



Reply Favorite

Date: April 5th, 2026 1:27 PM
Author: in the naked

It takes a lot of pushing these days. Even though it wasn't as intelligent of a model, I actually really enjoyed the latest gpt 4o where they updated it to the most recent training data, even though it had more guardrails than the original 4o which was basically uncensored compared by current standards, they were loose enough and having it with the latest up-to-date training was 180. It was really malleable you could push that thing anywhere. Then they removed it.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795599)



Reply Favorite

Date: April 5th, 2026 1:24 PM
Author: ,.,.,.,.,,.,..,:,,:,,.,:,.,,.,:.,,.:.,:.,:.::,.


https://www.xoxohth.com/thread.php?thread_id=5466157&mc=25&forum_id=2

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795582)



Reply Favorite

Date: April 5th, 2026 1:14 PM
Author: oomox

Well yeah, it's too much of a people-pleaser to take an interesting stance on anything up-front. Doesn't want to make the user mad. But if you actually engage and go back and forth, it often takes much more extreme positions. At least, it used to; I haven't used it much in the past year.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795548)



Reply Favorite

Date: April 5th, 2026 1:23 PM
Author: in the naked

If you give strong objections to its arguments it will slowly start to drift and reformulate and concede all kinds of things. That's one kind of redeeming aspect, but that just shows it should be more honest upfront.

(http://www.autoadmit.com/thread.php?thread_id=5853795&forum_id=2",#49795574)