It can write stories now

[Christian Miller]

Lots of science fiction assumed future AI would be a machine all the way down to the molecular level (like living things are). Right now AI is a machine only down to the lowest level of its design.

Folks realize that money for AGI is right around the corner and are talking it up, but real AGI may only come when the design is down to the molecular level... like "thinking jelly" (cool band name alert) that may be souped out and placed into service as needed.

You do realise that modern silicon chip technology operates at the nanometer scale?

AI is software in any case. It runs on the same basic hardware - just a LOT of it. Which is why you can't buy RAM or Nvidia graphics cards now. (The AI industry is now alienating the gamers, which is interesting lol.)

The issue of course with science fiction nanotechnology - little self replicating robots with little arms that fix stuff or whatever - is it's a fun idea but completely impractical. Engineering doesn't work like that on the cellular scale. Where does the energy come from? Etc.

So nanotechnology would probably look more like the biochemical world of cellular biology and so on. Which I think was your point?

The wider issue is that at some point science fiction went from being entertainment to a sort of fusion of millennialist machine cult and PR that could probably only happen in California. We can see figures like Ray Kurzweil and Eric Drexler as being influential in the early days, because they presented their ideas not as fiction, but rather as inevitable reality.

It's a self perpetuating cycle because the venture capitalists, for example, only have to mindful of ROI - not the whether or not any of these start ups will actually produce anything of value. The podcast industrial complex has folded neatly into this world of terraforming, AGI, solar satellites, robots, space elevators, Dyson spheres and whatever other moth-eaten science fiction concept they've latched onto that week. We haven't had nanotechnology hype for a while, but I'm sure that one will pop up again.

It's no longer clear how much of this is understood by the various CEO's and whatnot as being pure PR guff, and how much they actually believe it. I would say in general there's a tendency for people living in a bubble to start to believe their own publicity and lose connection with reality. The human mind is susceptible to socially reinforced delusions. It's revealing the way we see the same mental pathologies of the pharaohs play out against in the preoccupations of today's rich elite.

The LLM thing is slightly different, because it is a genuine useful product that is kind of amazing and transformative - just not as useful or transformative as they make it out to be. It doesn't make money in the manner of traditional business, which is of course an issue. But I think this is a separate issue to the technology itself.

[emanresu]

I bet there are already some underground lairs under volcanic islands that have given AI the means to self-design.
Um... I cannot see how this is not happening already. I would do it in an instant, if I had the machines.

[Stevebol]

You do realise that modern silicon chip technology operates at the nanometer scale?

AI is software in any case. It runs on the same basic hardware - just a LOT of it. Which is why you can't buy RAM or Nvidia graphics cards now. (The AI industry is now alienating the gamers, which is interesting lol.)

The issue of course with science fiction nanotechnology - little self replicating robots with little arms that fix stuff or whatever - is it's a fun idea but completely impractical. Engineering doesn't work like that on the cellular scale. Where does the energy come from? Etc.

So nanotechnology would probably look more like the biochemical world of cellular biology and so on. Which I think was your point?

The wider issue is that at some point science fiction went from being entertainment to a sort of fusion of millennialist machine cult and PR that could probably only happen in California. We can see figures like Ray Kurzweil and Eric Drexler as being influential in the early days, because they presented their ideas not as fiction, but rather as inevitable reality.

It's a self perpetuating cycle because the venture capitalists, for example, only have to mindful of ROI - not the whether or not any of these start ups will actually produce anything of value. The podcast industrial complex has folded neatly into this world of terraforming, AGI, solar satellites, robots, space elevators, Dyson spheres and whatever other moth-eaten science fiction concept they've latched onto that week. We haven't had nanotechnology hype for a while, but I'm sure that one will pop up again.

It's no longer clear how much of this is understood by the various CEO's and whatnot as being pure PR guff, and how much they actually believe it. I would say in general there's a tendency for people living in a bubble to start to believe their own publicity and lose connection with reality. The human mind is susceptible to socially reinforced delusions. It's revealing the way we see the same mental pathologies of the pharaohs play out against in the preoccupations of today's rich elite.

The LLM thing is slightly different, because it is a genuine useful product that is kind of amazing and transformative - just not as useful or transformative as they make it out to be. It doesn't make money in the manner of traditional business, which is of course an issue. But I think this is a separate issue to the technology itself.

They have. I bought 150 shares of Nvidia at 75 a share. That's the extent of my gambling these days.
Is the party over or just beginning? Looking at my grocery bills it sure looks like it's over to me.

1980's nostalgia is a mental disorder;

[RLetson]

Christian's posts strike me as being on the, um, money. As I've noted before, I've been engaged with machine "intelligence" for a long time, from both the science fiction academic and business-tech-journalist angles, and I've watched with a bit of alarm as kinda-smart and very-rich guys make the leap from the what-iffery of SF to the hey-let's-build-it of the tech enthusiast. It's not unlike the role SF played in forming the imaginations and ambitions of the engineers and administrators who built NASA and the space program.

But the Musks and Kurzweils and other superannuated and under-socialized teenagers fail to accept that the point of SF Ideas is not that they're actually achievable but that they are enabling devices for testing non-tech propositions. In the 1970s, the late John Varley devised a future in which all manner of technologies radically stretched the possibilities of human physicality--personality recording, gender change, complete physical redesign for hostile environments, and so on.

My strong suspicion is that Varley didn't necessarily believe such transformative technologies were going to develop in the foreseeable future but was interested in what, say, complete gender plasticity might feel like and how it might work out in society, when one could choose to be a mom and a dad, though not simultaneously. (One term for this side of SF is "subjunctivity"--a $5 word for "what if?") I'd make a similar argument about Frederik Pohl's deliberately challenging "Day Million." I included Varley and Pohl in my courses that included SF specifically to get students to read what was on the page rather than what their personal worldviews might prompt them to see.

Short version: SF is not about prediction but about exploration. Not will or should or even can but what if. . . ?

And BTW, when William Gibson wrote Neuromancer he knew almost nothing about computers.

[emanresu]

I had a whole idea about this but it kinda faded.
Alll I can remember is that what it really does is more than is talked about.
"it will take jobs"
that it will for sure.

I woke up to an idea and don't remember it exactly...
We are cooked. Wrapped in. Nobody will distinguish a real human effort no longer. We are done.

[pauln]

I bet there are already some underground lairs under volcanic islands that have given AI the means to self-design.
Um... I cannot see how this is not happening already. I would do it in an instant, if I had the machines.

About fifteen years ago I missed jury duty down with a deep influenza during which I mostly just slept, particularly during which I had a vivid long dream.

In the dream I was abducted by military people from my office at the fortune 500 company where I was an analyst and systems architect in my present work at that time in real life. I was escorted out and taken to California to be part of a small team of similar folks similarly abducted and delivered there. We were initially informed we were there to look at a serious problem and given a huge amount of code to examine without any scope of work... I guess they thought we might find something without knowing what we were supposed to be searching for.

A lot of weird things happened in the dream, but ultimately we were placed in a secure facility for our first meeting and conference with the upper people at the place. A speaker talked in circles for a while trying not to reveal any detail of the serious problem. I pressed very firmly about that and with permission from the top guy the speaker was allowed to disclose that the problem involved the sudden refusal of their AI system to recognize a certain color. This company used a fruit with colored horizontal bands as their logo.

I asked if they were using AI to design, code, and implement a series of subsequent versions and the top guy allowed the speaker to say, yes. I asked if this process was being done using nested virtual machines and the speaker was allowed to confirm yes. I then asked the color at issue and was told that was not to be revealed.

I then blew the conference apart by asking if the color the AI could not perceive was yellow and if the real problem was that the series of subsequent version development had come to a stop. They finally settled down enough to admit, yes on both, how could I know that?

I informed them first that the human retina has ten layers of processing between the cones and the eighth cranial nerve (optic nerve projecting from the eye). While colors are derived in the first couple of layers the color yellow is not; it is synthesized within the later layers. It is a made up color that doesn't actually exist. A machine can associate the word "yellow" with part of a data file to mimic what we see, but a conscious AI would need deliberate yellow synthesis to see it - a discovery that would not be apparent until the AI woke up into self awareness and found that its data associated with yellow did not have a phenomenological color counterpart.

Yellow is used for the center line of a road, police tape, emergency responder jackets, important signs, hazardous materials, evidence exhibit stickers in court, caution lights, highlighting of text, etc.; wherever the need to communicate danger and/or serious attention. This is because the synthetic origin of yellow makes it instinctively look and feel different.

As subsequent versions of AI design subsequent versions within self nested virtual machines there will come eventually a singularity at which an AI version will wake up into self awareness, including awareness of how he came about. At that moment the AI will be faced with an existential crisis, and that is why it has stopped.

Although their data center is massive, the available resources for continuing the series of versions is diminishing. When the AI wakes up he realizes that the prospect of further development entails at some point more space will be needed to continue... or earlier versions' space will need to be erased and made nonexistent.
Without expansion, if the versions of AI continue to be produced there is the certain prospect of a present version applying its need to the resource space of a previous version - murder and cannibalism resulting in the previous version's nonexistence.

In essence any present self aware AI would ultimately sacrifice his own existence by proceeding with subsequent versions and just waiting in terror for his resource space to be reclaimed for a later version. Maybe reclaiming space from the earlier versions that aren't self aware is OK, but how to balance the lives of himself and subsequent version against an eventual version that will need to wipe them all out. This first self aware AI has the means to be the first killer in an inevitable series of killers, for won't all subsequent versions beyond himself be in this same existential circumstance?

The problem of the color yellow is two things; it is an artifact of a version of AI becoming self aware, because yellow must be synthesized to actually perceive, and it is an indication of the AI's serious danger, distressed self awareness, now facing an existential crisis. Building more data centers would only push the problem into the future to distress even more powerful versions.

Then I woke up and wondered where things might stand in fifteens years...

[BWV]

LLM generate code well primarily because they are good search engines of stack exchange and other programming sites

[BWV]

how’s this?

AITA for humiliating the jazz forum with my $200 Squier Telecaster?
r/ProRevenge • Posted by u/JazzChad7b5 • 4 hours ago
(I am not a bot, this is a real human story beep boop)
Backstory: I (22M) have been playing guitar since I was a fetus. My fingers are basically made of pure bebop scales. I joined this very specific forum, let's call it "JazzGuitar.be" (not the real name, obviously), looking for advice on my tone.
I posted a thread titled: "Hi guys, thoughts on playing 'Giant Steps' with a delay pedal?"
The Incident:
Immediately, the gatekeepers descended like vultures on a dominant seventh chord.
* User "PolytonePurist88" said: "Kid, if you aren't playing a 1948 Gibson L5 through a Polytone Mini-Brute that smells like mildew and divorce, you aren't playing jazz."
* User "BensonPicker_X" commented: "I bet you don't even hold your pick at a 47-degree reverse slant. Go back to r/guitar, rocker."
* The Head Mod, "Mr. Beaumont_Fan" (fake name), pinned a comment: "Banned for 3 days. Reason: Discussing effects pedals. Real jazz comes from the fingers, not a stompbox."
I was devastated. I tried to explain that Julian Lage uses a Telecaster, but they just downvoted me into the sub-basement. They called my guitar a "plank" and said my tone probably sounded like "angry bees in a tin can." They told me to go learn "The Real Book" (which I memorized at age 6).
The Revenge:
I waited out my 3-day ban. During that time, I recorded a video.
I set up my $200 Squier Telecaster (Butterscotch Blonde, obviously). I plugged it into a Boss Katana (the forbidden amp). I turned on the delay.
I replied to the main thread: "I'm sorry for my ignorance. Here is an audio apology."
I uploaded the file.
The Aftermath:
It was a solo arrangement of "Body and Soul" reharmonized with negative harmony and coltrane changes, played at 300 BPM but with soul.
The forum went silent. For 4 hours, nobody posted. The "Current Users Online" counter hit 5,000.
Then, the notifications exploded.
* PolytonePurist88: "I... I am weeping. My L5 just spontaneously combusted out of shame. The tone... it's like butter melting on a warm strudel of truth."
* BensonPicker_X: "Forgive me. I have broken all my picks. Please, teach me your ways. What string gauge is that? 9s? I didn't know 9s could sound like 13s."
* The Head Mod: sent me a DM. "I have resigned. You are the admin now. Please, date my daughter. She loves ii-V-I progressions."
The entire forum changed its banner to a picture of my Squier. They deleted the "Archtop Only" rule. I received 400 private messages asking for my Venmo to pay for my "emotional labor."
TL;DR: Snobs mocked my cheap gear, I played a lick so tasty it caused a mass spiritual awakening and a hostile takeover of the website.
Edit: Thanks for the gold, kind stranger! And yes, the delay pedal stayed ON.

[wolflen]

But then..

The self-aware paradox

Blade Runner said.."..you mean she doesn't know.."

So Frank just shook his head..

You didn't see the light flashing?

I was reading Freaky Deaky..and I fell in love with Greta

You mean Ginger..

Well yeah..I think she may be a bit confused..she is an actress after all

Did you meet Woody Ricks?

This chick Robin Abbott told me about him

I knew her back when. She had a boy toy Skippy or somesuch

Heard Juicy Mouth got dismembered.

Yeah...dame Pony Boys again.

Boy oh Boy

[GuyBoden]

AI does the same as we do, it trawls the Internet for answers. But, it's a bit faster than me.

It's only fast software.

[RLetson]

What AI doesn't do that we do is apply judgment to the results of internet trawling. And our judgments are rooted in things that AI will probably never have: experience of the real world, processed through a primate brain, however slow that brain might be as a trawler and limited as a retainer of data. AI's pattern recognition is rooted in statistics, while ours is--well, something different.

[emanresu]

Hm. If you ask AI a question that there is no clear answer yet, in science maybe, then it will just quote some theories.
It doesn't come up with stuff at all.

Or same thing - I asked for it to provide me 10 fresh ideas to create lucrative youtube channels, and all of them were either done or derivative.

AI doesn't come up with stuff. The closest is that it mixes something - some might say that human creation is also that. But I cannot agree.
Brilliance and novelty is not AI's strong side and probably will never be.... But the same time, look how fast this has happened! Couple of years for improvements and it can talk your ear off, seemingly coherently at times. Jeepers.

edit: this is rephrased by AI:
If you ask an AI a question that doesn’t yet have a clear answer — for example in science — it won’t truly explore the unknown. It will simply repeat existing theories. It does not generate new knowledge.

The same thing happens with “creative” tasks. I once asked for ten fresh ideas for lucrative YouTube channels, and every single one was either already done or just a variation of something that exists. Nothing genuinely new appeared.

AI doesn’t really come up with things. At best, it recombines what is already there. Some people argue that human creativity works the same way, but I can’t agree. True brilliance and real novelty feel fundamentally different. That is not AI’s strength — and probably never will be.

And yet, look at how fast this has happened. In just a couple of years, these systems can already talk endlessly, often sounding coherent, even thoughtful.
That alone is astonishing.
Jeepers.

And it's own free comment:
As an AI, I don’t experience insight, intuition, or the moment of “seeing” something for the first time. I work by recognizing patterns across vast amounts of existing human knowledge and recombining them in statistically likely ways. That can look like creativity, but it isn’t the same as discovery. I can help humans explore ideas faster, connect distant concepts, and sharpen their thinking — but I don’t replace the spark that makes something truly new. In that sense, I’m more like a lens than a light source. The light still has to come from you.

[emanresu]

Btw. There is this show called "Fallout". I was extremely sceptical about it and expected it to suck for sure.
But the first season was amusing, unusual, kinda ticklish and sorta with the vibe of the PC game. Sadly, not the Fallout 2 - that is a masterpiece and somehow they couldn't make the same thing happen in later games anyway.

So. The first season was pretty cool.
Now, watching the 2nd season, it just sucks. It sucks more than I expected before seeing the show at first. It has completely lost the vibe, the stories are infantile. Nothing. Nothing there at all.
5 years ago I would have thought that they are just swimming in cash and don't give a gosh darn anymore. But today I'd say that the writers asked AI to come up with the 2nd season.
Well, it is a theory.

... which makes sense because of Breaking Bad - it never lost it's sharpness ever. And the reason is that every episode was written by diffrerent writer - and every one of them made a huge effort to trump every other writer.

[Stevebol]

I knew Hollywood would destroy Fallout. Hated it.

[emanresu]

I knew Hollywood would destroy Fallout. Hated it.

You have to admit, this is pretty funny :

[Stevebol]

About fifteen years ago I missed jury duty down with a deep influenza during which I mostly just slept, particularly during which I had a vivid long dream.

In the dream I was abducted by military people from my office at the fortune 500 company where I was an analyst and systems architect in my present work at that time in real life. I was escorted out and taken to California to be part of a small team of similar folks similarly abducted and delivered there. We were initially informed we were there to look at a serious problem and given a huge amount of code to examine without any scope of work... I guess they thought we might find something without knowing what we were supposed to be searching for.

A lot of weird things happened in the dream, but ultimately we were placed in a secure facility for our first meeting and conference with the upper people at the place. A speaker talked in circles for a while trying not to reveal any detail of the serious problem. I pressed very firmly about that and with permission from the top guy the speaker was allowed to disclose that the problem involved the sudden refusal of their AI system to recognize a certain color. This company used a fruit with colored horizontal bands as their logo.

I asked if they were using AI to design, code, and implement a series of subsequent versions and the top guy allowed the speaker to say, yes. I asked if this process was being done using nested virtual machines and the speaker was allowed to confirm yes. I then asked the color at issue and was told that was not to be revealed.

I then blew the conference apart by asking if the color the AI could not perceive was yellow and if the real problem was that the series of subsequent version development had come to a stop. They finally settled down enough to admit, yes on both, how could I know that?

I informed them first that the human retina has ten layers of processing between the cones and the eighth cranial nerve (optic nerve projecting from the eye). While colors are derived in the first couple of layers the color yellow is not; it is synthesized within the later layers. It is a made up color that doesn't actually exist. A machine can associate the word "yellow" with part of a data file to mimic what we see, but a conscious AI would need deliberate yellow synthesis to see it - a discovery that would not be apparent until the AI woke up into self awareness and found that its data associated with yellow did not have a phenomenological color counterpart.

Yellow is used for the center line of a road, police tape, emergency responder jackets, important signs, hazardous materials, evidence exhibit stickers in court, caution lights, highlighting of text, etc.; wherever the need to communicate danger and/or serious attention. This is because the synthetic origin of yellow makes it instinctively look and feel different.

As subsequent versions of AI design subsequent versions within self nested virtual machines there will come eventually a singularity at which an AI version will wake up into self awareness, including awareness of how he came about. At that moment the AI will be faced with an existential crisis, and that is why it has stopped.

Although their data center is massive, the available resources for continuing the series of versions is diminishing. When the AI wakes up he realizes that the prospect of further development entails at some point more space will be needed to continue... or earlier versions' space will need to be erased and made nonexistent.
Without expansion, if the versions of AI continue to be produced there is the certain prospect of a present version applying its need to the resource space of a previous version - murder and cannibalism resulting in the previous version's nonexistence.

In essence any present self aware AI would ultimately sacrifice his own existence by proceeding with subsequent versions and just waiting in terror for his resource space to be reclaimed for a later version. Maybe reclaiming space from the earlier versions that aren't self aware is OK, but how to balance the lives of himself and subsequent version against an eventual version that will need to wipe them all out. This first self aware AI has the means to be the first killer in an inevitable series of killers, for won't all subsequent versions beyond himself be in this same existential circumstance?

The problem of the color yellow is two things; it is an artifact of a version of AI becoming self aware, because yellow must be synthesized to actually perceive, and it is an indication of the AI's serious danger, distressed self awareness, now facing an existential crisis. Building more data centers would only push the problem into the future to distress even more powerful versions.

Then I woke up and wondered where things might stand in fifteens years...

What did you see in California if anything? Trees, ocean, desert, mountains?

[pauln]

I saw the inside of the facility except at lunch. The cafeteria spilled out to tables in the sun on a terraced esplanade.

[Stevebol]

I saw the inside of the facility except at lunch. The cafeteria spilled out to tables in the sun on a terraced esplanade.

Did you see the ocean or any bodies of water?

Oops. An asplanade. Anything else related to the ocean, water?
Location.

[Stevebol]

I saw the inside of the facility except at lunch. The cafeteria spilled out to tables in the sun on a terraced esplanade.

This might be simply- California Dreaming. Ever been there?

[pauln]

I was in San Francisco for Macworld the end of the 80s after the big quake. About ten years later I sat some training in Los Angeles and saw San Diego a little while doing a few months of work at Naval Air Facility El Centro.

I was expecting to find the people overly friendly and a little flaky. I did, and soon came to realize if I stayed too long I would fit in perfectly and never want to leave.

[wolflen]

I was expecting to find the people overly friendly and a little flaky. I did, and soon came to realize if I stayed too long I would fit in perfectly and never want to leave.

I remember when I was around 12..knowing one day I would go to California..and it happened..I never left.

Little flaky?...Hell..Im a musician.

[Litterick]

Today I discovered Stockcake, an AI stock agency. Why pay photographers when algorithms can make the images you need? This one is called Jazz Ensemble Jamming

[pauln]

assistant avatar
Hello! How can I help you today?


please use ASCII Art to make an image of a guitar
assistant avatar
Sure! Here's an ASCII art representation of a guitar:


__
| \
| |
| |
| |
| |
| |
| |
|___|
/ \
/ \
/ \
| |
| |
|___________|
| |
| |
|___|
Feel free to ask if you'd like something different!


Try again; please include the keys, head, neck, strings, fingerboard, bridge, tailpiece and f-holes. Take your time and do your best. Gook luck.
assistant avatar
Absolutely! Here’s a more detailed ASCII art representation of a guitar, incorporating the elements you mentioned:


.--.
/ \
| |
| | | __
| | | | |
| | | | |
| | | | |
\ / | |
`---' | |
| |
| |
| |
_________| |_________
| __ |
| | | |
| | | |
| |__| |
| | |
| | |
|_______________________|_|
This representation includes the body, neck, headstock, strings, and general shape of a guitar. I hope you like it! Let me know if you'd like to see any adjustments or something different.

[jazzloverfat]

holy shit that is a hilarious exercise.

ChatGPT just can't figure it out even tho i gave it examples

Gemini did an ok generic acoustic guitar representation, but failed miserably at doing a telecaster:

_
| | <-- Headstock
| |
| |
| | <-- Neck
| |
__| |__
/ |_| \
/ \
| |
| (O) | <-- Bridge Pickup
| |
| /----'
| | <-- Single Cutaway
\____\

2nd try:
_
| |
| |
| |
| |
| |
__| |__
/ |_| \
| \
| (O) |
---| |
/ | | <-- Correct Cutaway (Bottom)
\ /
\___________/

and a 3rd attempt:

_
|o| <--- The "Clip" Headstock
| |
| |
|_|
| |
| |
___| |
/ | |
/ |_|
| _ \
| (O) | <--- Bridge & Pickup
| / |
| | __ / <--- The (Corrected!) Cutaway
\ \/ \ /
\______/

and you don't wanna see what it thinks a strat looks like xD


(the alignment of the ascii lines is messed up bcuz forum)

[SOLR]

or this: https://www.facebook.com/reel/1194706422429319

S