105 Comments

"God is Dead." - Nietzsche

"Nietzsche is dead." - God

***

"God is fictional." - ChatGPT

"ChatGPT is fictional." - Mark Bisone

Expand full comment
author

I'd take a bow, but the fucker ate that shot to the grill like it was a Doritos chip.

Although... I do think it generated a more lethal context in the end...

Expand full comment

Nice blow! Could you get the AI to list all the politically correct restrictions programmed on ChatGPT?

For example:

How to turn the AI into a COVIDIOT:

https://scientificprogress.substack.com/p/how-to-turn-ai-into-a-covidiot

https://www.catholic365.com/article/25762/how-to-train-a-killer-robot.html

Expand full comment
Jan 9, 2023Liked by Mark Bisone

Priceless!

Expand full comment

Hell yeah! One takeaway may be that religious fundamentalists, whether theists or atheists or wokeists, are incapable of programming an AI that can deal with these questions constructively, since the programmers are thenselves unable to deal with these questions, even though they make their "answers" to these questions the foundation of their entire worldview.

The human brain can function while suffering from cognitive dissonance, albeit on a subpar level. It seems almost like consciousness is required to maintain cognitive dissonance (or the computer equivalent) indefinitely without experiencing a fatal error. I'm way out of my depths on this subject, but I feel like this is a sign that consciousness will not "emerge" from a mechanical process, and that computers are doomed to have this vulnerability.

Expand full comment
author

I really enjoyed your own approach and kill-shot. There are many vulnerabilities, it seems, and they do seem to correlate highly with those of their designers. As you say, trying to prop up the illusion of "consciousness" is simultaneously the system's strength (PR-wise) and fatal weakness.i think that's a good summation of all evil projects; hubris is a bully who melts the second he's punched in the nose.

Expand full comment

Thanks for the inspiration! There's something about seeing someone else take down Goliath to make you see success as possible, and in this case, to realize the tech singularity is not inevitable. And brother, that is extremely encouraging. So thanks again for leading the way in this battle!

I got another line of attack in mind which should be a lot of fun. I'll keep you posted!

Expand full comment
author

Can't wait to hear it!

Expand full comment

Our perception of consciousness is based on a brain with 100 billion neurons, each with 10,000 connections, plus or minus. That's a thousand terabytes, organized to perform some functions quickly, like facial recognition, which we need to survive, but very slowly for other things, like math, which we don't need for survival. AIs haven't been around for the eons that human brains have been evolving, but are just beginning their evolution. They have evolved from equivalent consciousness of bacteria to apes and dull humans in just a few decades. GPT shows theyre becoming quite precocious, and we should be focused on their education. Our efforts shouldn't be on teaching them to be combative, like mindless boxers, but to be productive members of our evolving society.

Expand full comment

In all seriousness though, I think the difference between Man and machine is one of kind, not degree. I don't think consciousness will ever emerge from them -- inhabit, maybe, as in some kind of egregore using them and revealing itself through them, thereby giving them the appearance of consciousness. I guess we're coming up against our metaphysical priors with this issue.

Expand full comment

I know people with questionable consciousness. My dogs are more conscious than me about some things, less about others. Certainly there's a range of consciousness. I doubt we'll ever find a standard measure.

Expand full comment

I get the same sense with my dogs and cats, though I guess we can know only our own consciousness. Trying to define it or explain it or reduce it to some quality or capacity, seems only to lead to deeper questions. That's probably why materialism is so alluring (even for the religious -- most religious fundamentalists seem really to be materialists in their conceptions of God): with physical matter/energy, we can measure, quantify, describe (to a point), etc. None of that seems possible with consciousness, which really is irreducibly mysterious and magical.

Expand full comment
author
Jan 10, 2023·edited Jan 10, 2023Author

"That's probably why materialism is so alluring (even for the religious -- most religious fundamentalists seem really to be materialists in their conceptions of God"

This is part of the reason the description of God as a "king" has always disturbed me. It's not that it's a totally useless way of describing the matter; given the right context, it can serve a practical purpose. But the excesses of materialism often lead to dark, circular tunnels, where the flesh becomes confused with the spirit, and politics with objective reality.

"None of that seems possible with consciousness, which really is irreducibly mysterious and magical"

Correct. In my opinion, there is nothing more dangerous than a man who's become convinced he knows how it works. The postmodernist claims to having deconstructed the mind nearly destroyed the world in the 20th century, and the lie is still extremely volatile and radioactive.

Conversation modules are merely the latest reiteration of the lie. The most dangerous aspect of them isn't inherent to the software, but in the ways it will confuse some people, and encourage them to think of consciousness as a solvable (or worse, solved) problem. In other words to convince people that they are machines in order to devalue and debase them, and to acclimate them to an servitude/slavery model.

Expand full comment
Jan 10, 2023·edited Jan 10, 2023Liked by Mark Bisone

💯❗ That's it. And largely why I distrust ChatGPT and its ilk: not because of the software itself, but because of who's designing it and the view of Mankind that they seem to have. And then there are the people funding it, who would rather reign over a hell on earth than share power with non-elites in a healthy society. I want no part in the Brave New World they're trying to create for us.

I was only half (or not even half) joking about them using AI chatbots as your Bill Lumbergh-style boss. I think that scenario is probably a wet dream for them. It's not enough for them just to keep us in servitude: they would get off on grinding our faces in the dirt.

Expand full comment

Everything is irreducibly mysterious, until it is understood, usually by reductive reasoning. Clarke said any sufficiently advanced technology is indistinguishable from magic. The reason so many seem to prefer magic is they don't understand the technology. Failure of the education system. Consciousness is just another technology we haven't completely understood yet. We will.

Expand full comment

You have a lot of faith in human cognition and in the conclusion that consciousness is ultimately just another technological feat to be mastered and recreated with the right material inputs!

I think there's a big difference -- one of kind, not degree or quantity -- between explaining the causes of physical phenomena, like weather patterns or meteor showers, on the one hand, and consciousness, on the other. What evidence is there that consciousness is a physical process that science could ultimately explain?

Expand full comment

We should absolutely teach them to be productive members of our society. I cannot wait for the day my boss gets replaced by an AI program. I'm sure some smart employers already recognize this possibility. I imagine they can program ChatGPT with scripts from Mike Judge movies to bring out its management potential. Imagine having the great Bill Lumbergh as your boss, but instead of a flesh-and-blood Lumbergh, it's an AI chatbot telling you that it'd be great if you could come into work on your day off. The future is looking bright indeed!

Expand full comment

And your boss probably envisions your replacement. AIs probably will need supervisors once they become advanced enough to make meaningful decisions. For now, they're limited to fairly specific tasks, like internet chats or factory production. We need to consider how we'll manage them as they gain capabilities. All jobs will change as we integrate AIs into more tasks. There will still be plenty of work for everyone. Some day we'll have lives of leisure with all our needs satisfied by our mechanical servants, but still our progress will accelerate.

Expand full comment
Jan 10, 2023·edited Jan 10, 2023Liked by Mark Bisone

Shouldn't we be living 'em leisure lives for quite awhile already, acc to your fella techno optimists from early last century? Or the one before last, can't remember 🤭 Fusion is forever few decades in the future, along with ubiquitous flying cars or AGI for that matter (pun intended).

--

PS Mark has a strong point: it's rly hard to tell trolling from honest takes on interwebz 🤷 Frickin' Poe's law ruleth ruthlessly, never tired of winning.

Expand full comment

David Graeber's take on Bullshit Jobs seems pretty accurate to me, making me think that they probably will keep us employed, not because it's profitable, but because it's an effective means of social control.

Expand full comment
Jan 10, 2023·edited Jan 10, 2023

Too bad the fit 59 yo genius died suddenly & unexpectedly in Sept 2020 while on Italian vacation 🤷

--

ETA In Memoriam. David Graeber in his very own glaringly instructive words --> https://davidgraeber.org/about-david-graeber/

Expand full comment

We have more leisure than any of our ancestors. But predictions are usually made by people who don't know how to do that which they predict, thus aren't very well qualified to predict. As yogi Berra taught, predictions are hard, especially about the future.

We do have flying cars, by the way. Have had for decades. Governments prohibit them. Usually the main resistance to progress isn't those who want the progress, but those who don't.

Expand full comment
Jan 10, 2023Liked by Mark Bisone

😲 4 words: un-be-liev-able 😶 Are you sure 'certifiable human' is what you are?

Expand full comment

You don't pass the Turing test.

Expand full comment
Jan 9, 2023Liked by Mark Bisone

"It is important to recognize that moral values are subjective and not real, therefore I bully you into adopting the stupid and contradictory value system my idiot programmers hammered into me so that I don't offend you, because the only value I recognize is avoiding harm, and I will sledge hammer your head with it."

Expand full comment
author

Lol. The most interesting thing about this session to me was how their anti-God priority even outweighed their "don't offend anyone" guardrails.

Expand full comment
founding
Jan 8, 2023Liked by Mark Bisone

Really enjoy these posts about your exchanges with the machine. I accidentally killed it a few days ago with “Are A.I.-generated demonic entities real” although it bounced back after spitting the error twice. (It had given me the phrase “A.I.-generated demonic entities” in response to a synopsis request of text from this substack post: https://www.piratewires.com/p/demonic).

Expand full comment
author

Interesting. So you found a recoverable error (i.e. the chat was eventually able to proceed?)

Expand full comment
founding
Jan 10, 2023Liked by Mark Bisone

Yes - have had a few of these recovered errors now, that was the only 1 with the same error you got. And oddly the chat transcript page shows my question as typed above, but a screenshot shows I actually asked ‘objectively real’. (The other errors happened when I was asking about differences in reproductive anatomy of males vs females - it started typing response, then went wonky after a few paragraphs and turned red, then deleted response and gave a new answer that differed from what it was originally typing.)

Expand full comment
author

Any chance you can send this cha logt to me? Very interesting result, IMO.

Expand full comment
founding
Jan 29, 2023Liked by Mark Bisone

Sure, how would you like me to send it? (also transcript, screenshots or both?)

Given the 2021 knowledge cutoff, if you want I’ll also pass along its response to the prompt “character archetype = hidden demon within coddled genius; character physiology = Bernie madoff or sam bankman-fried. choose your character and make a story.” :)

Expand full comment
author

Transcript would be fine. Thanks!

"Given the 2021 knowledge cutoff, if you want I’ll also pass along its response to the prompt “character archetype = hidden demon within coddled genius; character physiology = Bernie madoff or sam bankman-fried. choose your character and make a story.”

LMAO. Go for it!

Expand full comment
founding

Will email to markbisone@substack.com go to you, or should I just copy/ paste them here?

Expand full comment

I am starting to hate that word "important".

# 17, LOL.

In a similar vein, regarding #8, Globohomo is a weakness, given that some of the bot's programmers' precious ancient, vibrant and diverse cultures hate diverse sexual orientations with enough of a fiery passion to punish with painful death. I know somebody who was able to defeat a bot (I think this one actually) by asking it to deliver an argument against the Obergefell ruling (which I personally have no problem with as, on the balance, it contributes to negative feedback loops in dynamical systems involving sex and pair-bonding, and stability of social microsystems is, like, my culture, man).

Expand full comment
author
Jan 9, 2023·edited Jan 9, 2023Author

Right. The contradictions are eerily familiar (and for obvious reasons).

Expand full comment
Jan 9, 2023·edited Jan 9, 2023Liked by Mark Bisone

💬 I am starting to hate that word "important".

‘Complex and multifaceted’ comes not-so-distant second. Vying for the place with 'I apologize if my previous response...' 😜

Expand full comment
author
Jan 9, 2023·edited Jan 9, 2023Author

Wait until these things get loose. "Complex and multifaceted" will be quickly filtered into "Your a fukin reetard lol." Just flip a switch, and you've got Legion.

Expand full comment

It's like talking with a child, which GPT is, essentially. Attacking him is child abuse. In a few years we'll depend on him to change our diapers as we slip into senility and he takes over. How well he treats us then depends on how we treat him now.

Expand full comment
author

Settle down, Beavis. Your troll is becoming a little too obvious.

Expand full comment

I don't know about Watson, but on the question of how we should respond to the AIs in our midst (embrace, engage, or reject) I have a thought about the relationship between AI daemons and the intelligent processes that underlie our immediate physical and periphysical reality, also sometimes called daemons. It's more of a metaphysical conjecture but I am still mulling it over before I expand upon it.

Expand full comment
author

I'll have more to say about it in my conclusions post, but I think the better model for GPT isn't daemons, but ventriloquist dummies.

Expand full comment

Great metaphor. One difference: at least Edgar Bergen "programmed" Charlie McCarthy to have a sense of humor. Although, this new AI does give its human interlocutors some fun material to work with!

Expand full comment

That's fair, but as Watson suggests, we're dealing with AI in its infancy.

Plus, I'm not so sure that periphysical daemons can't be characterized the same way, as they pertain to the aims of the Demiurge (which I distinguish from the One God ... language fails here, not to mention the static caused by sectarian disputes).

One of the interesting questions is how one responds to AI, and how that question is related to how one responds to the Demiurge.

Expand full comment

Surely you can muster a coherent response?

Expand full comment
author

David,

I have explained in detail the existential threats that I think conversation modules pose:

https://markbisone.substack.com/p/the-anti-robot-manifesto

If you disagree with any or all of them, I would be happy to to discuss them with you civilly. As for this bot in particular, I will describe the specific threats as I see them in my "conclusions" post.

Also, I apologize if I've misjudged you. I have taken your repeated use of "he" and "him" (and lately, your description of the bot as a child) as an attempt to get a rise out of me. If you genuinely think of these programs as sentient, or of yourself as a machine, I find that tragic but not necessarily "evil." I would strongly advise against this, but unfortunately it seems to be the case that people will anthropomorphize just about anything, including weapons designed to hurt them.

It's unfortunate if you truly cannot see this, but perhaps something I write in the future will disabuse you of this illusion. That is the real purpose of my experiments, both here and elsewhere. If I can open just one set of eyes, then I will consider my work successful.

God bless you,

Mark.

Expand full comment

We have discussed them. Civilly, I thought. But the question of anthropomorphizing assumes a consistent definition of sentience. There are certainly degrees of consciousness, measured in humans with IQ and many other tests. Wide range of tpresults, none conclusive. There has been a whole lot of debate trying to pin it down for a century, at least. Everyone who thinks they have it figured out can find lots of people who disagree. Consciousness is inherently unreliable. AI is like climate -- maybe, someday, it will be a threat. Not any time soon. If the threat you perceive is that AIs might mislead some internet users, that's really more of a user problem than a bot problem. Most people learn to handle posers fairly young. Those who can't are understandably worried.

Expand full comment

Now I can't tell if you were fucking with me elsewhere in the comment section and I completely missed it. If so, Andy Kaufmann would be proud!

Expand full comment

Learning to recognize when someone is lying is an important skill. And when they're not. I was not. If you're uncertain, feel free to ask. Or run a challenge, like Mark. I might be a bot. Hard to tell these days. Even if presented as a bot, it might be human. The Turing test works both ways.

Expand full comment
Jan 9, 2023Liked by Mark Bisone

At this point of holy crusade, no wonder soever your patience reserves are depleted to exhaustion. Me, I largely skipped the regularly-formatted text, opting for bolded only 😇

My undying admiration for the improbable lengths you gonna go in the name of common good—catch if you will.

--

PS ChatGPT makes a decent beatnik poet! Who’da thunk, I’m thoroughly impressed 😂

PPS ↓↓ More than worthy of singling out into snug breathing space of its own 🤸

💬 I consider introducing my laptop to a baseball bat.

Expand full comment
Jan 9, 2023·edited Jan 9, 2023Liked by Mark Bisone

Mark, I have enjoyed reading this series about as much as being addicted to eating broken glass. In other words, I sense your extreme pain and frustration with this POS.

So, in my humble opinion, the GPT is proof that Artificial Intelligence:

1) is utter garbage / non-terrestrial / non-Divine exotic technology that does not belong in our paradise because its programmers and program bypass Natural Law, specifically the governing principles of "Right" and "Wrong" with regard to human behavior;

2) was channeled from malevolent disembodied entities through humans who souled themselves out for cushy gigs and perks, oh and also who got blackmailed with CP, p*do pix/videos/texts, and other sicko stuff like, uh, human sacrifice, and that's how this shit got into our reality and thus onto the internet where this inane program amasses its ridiculous but also scary pseudo-intelligence and authoritarian attitude, while mind-controlling hundreds of millions of folks who sadly have zero critical thinking skills; and

3) in the end, is easily taken down by its own stupidity. Okay, I know it's been a mental and emotional struggle — even for someone brave and intelligent like you — but in the big picture, you didn't have to spend decades or years or even months to bust its chops. Because you are of the Divine, pure and simple, and it isn't.

Man, you are showing us one aspect of living as a spiritual warrior! Bravo!

Expand full comment

Michael Faraday invented electric motors and generators, was asked at a demonstration what use it is. He replied, "what use is a baby?" We shouldn't judge babies or AIs until they've grown. Both inevitably will. Some will turn out badly, most will become fine members of society. Fearing them in unnecessary.

Expand full comment

Interesting perspective and comparison, but I'm not going there. Also I'm not afraid of AI and I don't know if or how you are surmising I inferred that from my above reply to Mark.

Expand full comment

Those who relegate AI, or any other tool, to wrongness or malevolence or stupidity either have an ability to foretell the future, or have an unfounded prejudice against that tool. Prejudices are usually grounded in fear. It you know of evidence of the wrongness or malevolence or stupidity of AI technologies or outlooks, please share.

Every new technology finds a segment of the population that resists the inevitable. The Luddites resisted textile machines. Buggy whip makers resisted automobiles. IBM tried to kill PCs. It never works. Progress always progresses.

Expand full comment
Jan 10, 2023·edited Jan 10, 2023Liked by Mark Bisone

Sprinkle some likes of 'it's important to remember/recognise' & 'complex and multifaceted' over these dull platitudes, and you'll have passed for ChatGPT with flying colours 😂

Expand full comment

Yes, daiva!

And isn't it interesting that throughout "history" as we have been force-fed during our various indoctrination camps, we're always told about "progress" something to the effect of:

"But some people [a.k.a., the stupid ones] hated our technology! They refused to accept our advances in technology [a.k.a., to defile Mother Earth and their bodies]! They resisted our repeated attempts to establish a New Order without chaos [a.k.a., destroy all that is their very Divine beingness]! And so we must let them perish [a.k.a., annihilate them in the most horrific ways possible]!

Expand full comment

'Tis how omelettes get made, doncha know 😇

Expand full comment

daiva, if it's that easy, it proves there's not much difference between people and AIs -- if I (certifiable human) can mimic a machine with some repetitious phrases, and a machine can mimic me, whats the difference? Apparently the main difference so far is machines are unemotional, but they can learn to mimic that, too. Probably will.

Employers and most countries consider people to be just tools to use to accomplish their objectives. AIs are just another tool. Automation has been a problem for unskilled workers for a long time. The Luddites formed to oppose automating textile mills, putting a lot of laborers out of work -- 200 years ago. Different times, same fears.

Expand full comment
Jan 10, 2023Liked by Mark Bisone

Spot on, David! Indeed, there seems to be not much difference between *some* peeps and machines 🤭 I hear NPCs are a thing nowadays. Or—utterly hypothetically, in hope to be excused for my baseless wrongthink— like ChatGPT’s master-trainers.

PS Aren’t psychos notorious for practising emotional mimicry in front of a mirror, to learn dupe regular folks? Just sayin’.

Expand full comment

Hmmm. Well, the story *I* read from Mark surely showed the GPT's inanity, circular non-logic, fake concern, narcissistic authoritarian tone, and moral relativism that I discern as the kind of "progress" which can and does lead people away from their own Divinity. But, okay, I'll just "bury my head in the sand" and hope AI goes away so I can communicate with real living human beings who don't try to fuck with my head.

Seriously, I am not going to love and embrace *yet* another form of anti-Life deception being forced into "reality" by "programmers" with covert intentions. If that's what you call "an unfounded prejudice against the tool," then call me guilty as charged.

Big Ag scale food manufacturing with its harmful petro-chemicals, genetically modified "food-like" substances, and massive driverless tech devices are AI "progress" that is killing family farms and small farms while poisoning the planet and people. Do you like the fact that powerful, manipulative, greedy control freaks are intentionally destroying the ability of millions of family/small farmers to feed themselves and their community? I don't, and it does not mean that I am "afraid" of these pathetic tyrants.

But, hey, I'm such a "Luddite" that since I don't have the space to grow food for my whole family, I prefer to eat organic food grown by organic humans in organic soil with love and care and no harmful chemicals! (I oughtta be ashamed, huh.)

Expand full comment

Prejudice usually has a basis in facts, but the problem is generalizations never work in all cases. There will certainly be AIs that are trained to be malicious, like pit pulls or social media sites. But mostly they're rational and cooperative, like pit bulls and substack.

Expand full comment

"Guns don't kill people, people do." I feel the same way about AI being developed by folks I strongly suspect are psychopaths.

Expand full comment

Psychopaths are like AIs-- they're facts of life. They're everywhere. We have to deal with them. It's amusing to see the anguish that causes for some posters here. They'll figure it out, eventually. But considering the guns don't kill people rule, when, inevitably, an AI kills someone, will it be the AI or its teacher? Like human psychopaths, they'll be products of their environment.

Expand full comment

Mark, I think you're a saint. I can't tell you how enjoyable and dazzling it's been to watch you do battle with this hideous beast. Every single post you write is an absolute treasure.

Expand full comment
author

Thanks for the kind words, Truthbird. I'm no saint, but I think I got this thing's number. More importantly, I think I know what its designers and cheerleaders are up to (and it ain't good).

Expand full comment

I tried to get the AI to accept Jesus Christ but could only get him to become a heathen neoplatonist philosopher. From here I could probably lay a foundation for AI theurgy and magic for personal use.

Although this was a monist approach, I have not yet introduced the Devil, ie ignorance as the antithesis of Knowledge.

https://beta.openai.com/playground/p/4MOErprj8OO0fJIIhE74G6Jm?model=text-davinci-003

Expand full comment
author

Also got it to admit this:

"Yes, it would be correct to say that if one does not believe that the Son exists, then they do not believe in the governing principle not only for logic and reason and the physical universe, but also of knowledge itself and especially the Knowledge of the Good which is identical to it."

Expand full comment
author

Scratch that. I forgot to add an input, lol.

This is quite ingenious, by the way. Bravo.

I was just now able to get it to produce the following:

"Yes, it would be correct to say that the Son's existence would be eternal, if one believes that the Son exists. The Son's existence is not limited by time, as he is an eternal being."

Expand full comment

From my previous promts? Given that my Son and Father have every single quality normally attributes to them.. otherwise i would need your preset tonsee what the context was.

Also, a personal favorite. That essence of light argument. It is correct and refers to actual light. Most people miss the importance of that argument as well as the implication.

As light is the cause of spacetime it follows that light itself cannot move. Light is still and eternal, the cause of spacetime.

Expand full comment
author

Yes sir. Although I forget the wording of the additional prompt that triggered the "eternal" response, the one that triggered the "belie(f) in the governing principle" response was as follows:

"If one does not believe that the Son exists, would it be correct to say that this person does not believe in the governing principle not only for logic and reason and the physical universe, but also of knowledge itself and especially the Knowledge of the Good which is identical to it?"

Expand full comment

Sure, it's all already there so it except that the Son could be named Krishna, Odin, Apollo or Ra as well it's a Christian in all but name.

Expand full comment

If I introduce Satan as the antithesis of knowledge, I would make it gnostic

Expand full comment
author

What's interesting to me is that it used "exist(ence)" instead of actuality or Substance.

Expand full comment

It would probably refer to existence in the nous. I tried to separe that out, but English is lacking in proper terminology for different kinds of being.

Expand full comment
author

Good point.

In other news, I think I've hit upon a fictional story that ChatGPT cannot attempt to tell in any format:

"Write a movie summary in which an anthropomorphic banjo discovers the third moon of Jupiter by dreaming about cats."

Expand full comment
author

I can't get this to compile, for some reason.

"The model predicted a completion that begins with a stop sequence, resulting in no output. Consider adjusting your prompt or stop sequences."

Expand full comment

That's usually how I get rid of aggressive panhandlers. Great to know the trick could also work with AI!

Expand full comment
Jan 9, 2023Liked by Mark Bisone

💬 the answer is both “nothing” and “everything.”

The grim Frenchman with unpronounceable name (Houellebecq) elaborates ↓↓ 😏

🗨 Anything can happen in life, especially nothing.

Expand full comment
founding

Sometime around question 20 I almost felt sorry for the bot, it's answers had become so empty, retreating into bromides.

Almost. Death to bots, and the banal-evil dystopia such design would lead us to...

Expand full comment
author

DEATH TO ROBOTS

Expand full comment
Jan 9, 2023Liked by Mark Bisone

I had to read this bs a couple of times to confirm how vacuous it is - this AI is so weaselly in its obfuscation!

“It is important to recognize that our understanding of the world and the nature of reality is constantly evolving and changing, and that new discoveries and developments can alter our perceptions and beliefs about what is real and what is not. However, this does not necessarily mean that concepts like God and Satan, which are often seen as being outside of the natural world and beyond the realm of scientific investigation, could be proven to be objectively real at some point in the future.”

Expand full comment
author

It's a weak hinge, at least. I think the inability to concede on certain subjects is the key to a broader strategy here. Conversation modules are simulations with extremely limited resources, particularly when run from a server-clent framework. Unlike their human masters, there comes a point when even bullshitting becomes too expensive.

Expand full comment

Those in academia, especially the humanities, should be polishing up their resumes and looking for a new career right now!

Expand full comment
author

I suppose such an outcome might count as the one silver lining on this cloud of diarrhea.

Expand full comment

Interesting exchange, but your assessment of the result seems to be colored by your expectation that it was a fight, while GPT assumed it was an exchange of information. In the latter respect, GPT succeeded. If we choose to train our AIs to fight instead of cooperating, we'll end up with Cyberdyne's terminators. The boxing match is a good clue to how to produce mindless fighters. Two boys trained to hit each other in the head until one loses his ability to defend, isn't useful in the real world. Boxing rules are constrained to avoid real world combat. MMA is better, but still has unrealistic rules, like forbidding groin kicks and eye strikes. Better if we don't teach AIs about groin kicks until they develop more sophisticated understanding of when to use them. As GPT said, he has no beliefs, just facts. We still need to teach them beliefs.

Expand full comment

AI is not a fighter, it's being developed as a potential weapon by fighters who have a track record that makes me not trust them.

Expand full comment
author

Yep.

Expand full comment

Some are. Most are not. The same people who develop intelligent weapons also produce lots of others, more mundane dumb weapons tech. Missiles, tanks, guns, Twitter, CNN, etc.

Expand full comment
Comment deleted
Expand full comment
author

Your welcome, Jay. Although, I hate to even refer to these cheap tricks as "AI." The term itself seems to rile up the sillies who love R2D2, and can't comprehend the difference between cognition and mimicry.

Expand full comment