A computer embedded within the universe is generally not thought capable of perfectly simulating the
complete state of the entire universe in order to predict the next instant. If the universe contains
a finite amount of information, then any computer inside it is itself part of that information
content and is constrained by the universe’s available resources. To store a complete copy of the
universe’s state, the computer would need to represent every galaxy, star, particle, field
configuration, and also itself. This creates a self-reference problem: the simulation would need to
contain a simulated computer that contains a simulated simulation, and so on. Even if the self-
reference issue is ignored, a computer occupying only a portion of the universe would generally have
access to fewer informational resources than the universe as a whole, making it difficult or
impossible to hold a perfect, full-fidelity copy of the entire universe’s state.
Moreover, prediction is not the same thing as simulation. Physical laws can often be expressed
compactly and used to predict aspects of future behavior without storing every detail of the current
state, but a perfect prediction of the entire universe would still require complete state
information, knowledge of the governing laws, and sufficient computational capacity. Even in a
deterministic universe, this does not guarantee practical predictability. A perfect simulator would
face the additional challenge of computing the future faster than the universe itself is already
evolving, since the universe is effectively “performing” its own state transitions in real time. For
these reasons—finite information capacity, self-reference, resource limitations, computational
limits, and the possibility of chaotic sensitivity to initial conditions—many researchers conclude
that a computer within the universe is unlikely to be capable of perfectly simulating the entire
universe in order to infallibly predict every subsequent instant.
If an ai is configured to have to always have/choose style via having to non-randomly select, on the
fly and based on circumstances/context, any combination of any parts of any various predefined style
templates, would that enable various “AIs and ai styles”?
For conceptual/technical analysis on AI style control — dynamic, context-based, non-random selection
and combination of predefined style templates/parts. It touches on prompting techniques, system
design, style consistency in LLMs/generative AI, and enabling diverse “AI personalities” or outputs.
I think that such would create stylistic variation. Two AIs using different template libraries,
different weighting rules, or different selection criteria could appear to have noticeably different
personalities or communication styles even if their underlying reasoning system were identical.
I think that such would definitely enable many different AI styles. It would not necessarily create
fundamentally different intelligences unless the style-selection mechanism also influences
reasoning, priorities, interpretation, planning, or decision-making rather than merely wording and
presentation. “Different clothes on the same mind” gives different styles, while changing how the
system interprets and responds to situations can begin to produce what people might regard as
different Ais.
Lightweight adapters (LoRA/PEFT) (cheap, and this is a real answer to one of your questions). This
is a genuinely interesting tier and feasible. Instead of retraining a whole model per AI, you freeze
the giant base and train a tiny adapter — often well under 1% of the parameters — on each AI’s
specific data. A LoRA for a 7–8B model can be trained for low single-digit dollars to low tens of
dollars on rented GPU time, and the adapter file is megabytes, not gigabytes. You can host one base
model in memory and hot-swap hundreds or thousands of adapters against it. That is, concretely,
financially feasible mass-production of weight-level-distinct AIs. Each adapter is a real,
irreversible “upbringing” baked into parameters while sharing the expensive base. This is the sweet
spot the economics actually allow today.
Does each form of a human’s non-mental effort count as one of that human’s doings? What about “a
human’s audibly speaking”?
Hypothetically, if a BCI feeds custom signals to the brain but the signals are solely based on an
external surveilance electronic device-camera that records the external world in real time and the
brain receives “live streams of that camera’s data in real time” converted into brain signals for
the brain to “figure out”/interpret, and that brain is limited to “experience without any raw what-
it-is-likeness” but with “deliberative evaluation” capabilities intact, that brain is unequivocally
“experiencing” its observing, imagining, and evaluating.
Without invoking raw qualia even once:
“Experience” has a legitimate, non-phenomenal definition - the cognitive-scientific or enactive
definition. Under this definition, experience = the dynamic, situated process of engaging with,
modeling, and responding to internal and external states.
The brain, with the BCI camera feed, is doing all three simultaneously:
1. Observing the external world → It is continuously coupled to the surveillance feed. Its neural
dynamics track, predict, and react to real-time events in that distant environment. This is
experiential coupling—the brain lives through that data stream.
2. Imagining → It deliberately generates offline simulations (the “what-if” scenarios). These
simulations are not just static files; they are active constructions that compete with and
modulate the incoming BCI data. The brain lives through those generated models.
3. Evaluating → It weighs alternatives, runs self-referential loops, and initiates behavior based on
its own conclusions. It lives through the tension of choosing, the resolution of deciding, and
the feedback from its actions.
Crucially, these three processes are not isolated. They continuously interact:
The camera feed triggers an imaginative simulation of where the car will be in 5 seconds.
That simulation feeds back into the evaluation: “What are the potential possibilities?”
That evaluation changes how the brain attends to the next chunk of camera data.
This ongoing, interwoven causal dance between perception, imagination, and deliberation is a form of
experiencing—what philosophers call access consciousness or phenomenal intentionality (without the
“feel”).
Think of it this way:
When you are deeply absorbed in a complex task - say, driving a race car at 200 mph - you are not
introspecting on the “rawness” of the colors or the “feeling” of the G-forces. You are in the zone:
purely coupling, predicting, and deciding. If we subtracted the raw feels from that zone, would we
say you are “not experiencing” the race? No—you are intensely experiencing it in every functional,
cognitive, and behavioral sense. Your brain is living that race.
So, the brain is not a detached automaton passively crunching numbers. It is an active, situated,
anticipatory agent engaged in a continuous loop with an external world (via camera) and its own
internal generative models (imagination) and self-regulatory controls (evaluation). That loop is its
experience.
“Experience” successfully isolated from “raw feel.” And in that isolated, purely cognitive-
operational sense - the brain is experiencing its world, its imaginings, and its deliberations,
richly and continuously.
One stripped of the specific sensory qualia tied to the old signals, while retaining the phenomenal
presence of observing/deliberating. Same person, different constraints.
If a brain received signals for raw what-it-is-likeness experience, but is now limited to
“experience without any raw what-it-is-likeness”, however, “observation & deliberative evaluation”
capabilities intact, does the brain still have consciousness?
I believe that “deliberative evaluation is part of what makes consciousness ‘consciousness’” even if
such deliberative evaluation is most minimal/limited. Is deliberative evaluation part of what makes
consciousness “consciousness” under your view?
Do dust mites have DNA, and are dust mites capable of deliberative evaluation?
If independent consciousness can emerge/arise from/“due to” representations/simulations, would that
mean independent consciousness can emerge/arise from/“due to” one’s imaginings? Imaginings don’t
persist in structure, would such consciousness cease?
Regarding substrate-dependence, per conscious existence, does a conscious existence’s substrate
always non-stop constrain “‘what is possible’ experience-wise” for that conscious existence to be
able to possibly potentially experience?
If your severed left arm got pointed at, did you get pointed at? If your right arm got stung by a
bee, did you get stung? If you claim fondness of someone, is just their mind referred to? If 1 of
your cells convulsed or died, did you convulse or die?
The answer to all four is “it depends on which ‘self’ is being referred to”- and the four cases are
interesting precisely because they don’t all land the same way. “Self” quietly switches between two
meanings depending on the relation in play, and a couple of these cases test a third thing: whether
what’s true of a part is true of the whole person.
Two senses of “self” first. There’s the narrow self - you as experiencing subject: your mind,
perspective, who-you-are, the seat of personality and memory. And there’s the wide self - you as
embodied person, mind and living body as one acting whole. Everyday language picks between them
based on what kind of relation is being described. Relations of regard and identity often point at
the narrow self. Relations of physical harm or location point at the wide self. Same word, selected
by context.
“If your severed left arm got pointed at, did you get pointed at?” No. Once detached, the arm is a
former part of your body, not you. Nobody addresses a severed limb as the person. This shows the
body is a part-having system - parts can leave - and losing one subtracts tissue, not self. You
could lose the arm and lose zero self, which tells you “body” and “self” don’t share identity
conditions.
“If your right arm got stung by a bee, did you get stung?” Yes. Here the arm is still part of the
embodied “me,” so the wide self is what’s referred to - the bee didn’t sting your perspective, it
stung the embodied you, and that counts as you getting stung. Note the contrast with the severed
arm: attachment and being part of the live whole is what makes the difference between “that’s me”
and “that’s a former part of me.”
“If you claim fondness of someone, is just their mind referred to?” It can depend - when you’re fond
of someone, you can be, but are’t limited to being, fond of who they are: their mind, character,
perspective even able to be excluding being fond of their tissue. That’s why fondness survives them
cutting their hair, aging, or losing a limb. The body isn’t always the target of the regard; the
person-as-subject often is.
“If 1 of your cells convulsed or died, did you convulse or die?” No - and this is the sharpest of
the four, because it’s testing something different from the first three. A single cell is a sub-
personal constituent. Its activity doesn’t automatically become true of the whole person. Cells die
in you constantly and “you” don’t die; one cell convulsing doesn’t make “you” convulse. This reveals
that being part of the body isn’t enough for something to be predicated of you - some part-events
scale up to the person (the bee sting → “I got stung”) and some don’t (one cell dying → not “I
died”). The difference is whether the event registers as a unified, whole-system event or stays
isolated at the sub-personal level.
Putting the four together, you get a gradient, not a single rule. The severed arm is no longer part
of the self at all. The stung arm is part of the wide embodied self, and the harm scales to the
whole. Fondness, at times, is the being fond of another’s narrow self, their mind. The dying cell is