AI on the blog?
Commenter “huxley” writes:
A tiny heads-up.
Within months you are going to start getting Chat-generated comments. I have no idea what you should do.
I don’t know either, but here’s my take on it.
If I understand the definition of AI, spam comments are a very primitive form of AI that’s been generating comments on blogs almost as long as blogs have been in existence. Spam blockers do fairly well with that sort of thing, although now and then something sneaks through. I don’t know what algorithm spam blockers use to detect them, because although most are quite easy to detect, others are more difficult.
However, present-day AI is far more sophisticated and getting more so every day. I like to think I could detect it anyway, but the truth is that maybe I couldn’t. My sense of how I could tell which comments are AI has to do with the nearly twenty years I’ve been reading comments here, and after all that time I notice that each person has a “signature” – an idiosyncratic and personal style. Could AI mimic that well enough? Perhaps. Or would it remain just a mite tone-deaf? I don’t know.
For now, I usually recognize when a troll has come back under a different moniker. It works on the same principle – recognizing a style. There are other ways to tell, too, trade secret methods that I won’t describe here because I don’t want to give them away.
AI worries me very much though, particularly fake videos. It’s difficult enough now to tell what is true and what is false. AI may make it impossible. Where will we be then?
I wonder about AI spam in the comment section as well. Worse, the entire internet is going to get trashier as AI feeds in itself to create content. I’m not looking forward to it.
They should change it from AI to AB — Artificial Babbling.
However, present-day AI is far more sophisticated and getting more so every day. I like to think I could detect it anyway, but the truth is that maybe I couldn’t.
–neo
I fear the latter. AI can generate unique, interesting content with a personalized style.
________________________________
huxley: Write a movie review of “2001” as if Mickey Spillane wrote the review.
ChatGPT3.5: Title: “2001: A Space Odyssey” – A Galactic Thrill Ride
You ever find yourself sittin’ in a movie theater, watchin’ a flick that’s so outta this world it’s like a trip to Mars without the spacesuit? Well, buckle up, pal, ’cause Stanley Kubrick’s “2001: A Space Odyssey” is one wild ride through the cosmos, and it’s gonna leave you scratchin’ your noggin.
Right from the get-go, you’re hit with this eerie soundtrack that’s like a thousand saxophones screamin’ in the dark. Ligeti’s tunes set the stage for a tale that’s part space opera, part mind-bender. You think you got a handle on things, but Kubrick’s playin’ a game of intergalactic poker, and the cards are stacked against your brain.
The plot unfolds like a detective story gone interstellar. There’s this monolith thing that shows up at crucial moments, like a dame in a tight dress at the wrong end of a smoky bar. You don’t know if it’s a friend or foe, but it’s callin’ the shots, just like the mysterious puppeteers in those hard-boiled crime novels.
Now, let’s talk about HAL 9000. This ain’t your run-of-the-mill computer; this machine’s got more attitude than a back alley cat. It starts actin’ like a two-faced snitch, and you’re left wonderin’ who’s wearin’ the black hat in this space showdown. Dave Bowman’s gotta deal with this silicon snake in the grass, and the tension’s thicker than cigar smoke in a private eye’s office.
Kubrick’s visuals hit you like a one-two punch. The spacecrafts move through space like getaway cars on a high-speed chase. The special effects are top-notch, and the stargate sequence is like takin’ a joyride through a neon-lit dreamscape. You won’t believe your peepers, and you’ll be wonderin’ if you accidentally dropped some tabs before the movie.
But make no mistake, “2001” is a head-scratcher. You’re left with more loose ends than a mob snitch’s alibi. The ending’s like the last page ripped out of a detective novel – you think you got the answers, but Kubrick’s playin’ coy, leavin’ you hangin’ like a suspect in an unsolved case.
In the end, “2001: A Space Odyssey” ain’t your typical sci-fi caper. It’s a wild, mind-bending trip through the cosmos that’ll leave you questionin’ reality faster than a shot of rotgut whiskey. So, grab your fedora, light up a cigar, and strap in for a ride that’s gonna have you talkin’ about monoliths and stargates for days.
___________________________________
Actually, I imagine some spambots are already using some form of Chat.
I’m afraid the future is already here.
Keep in mind that ChatGPT 3.5 is already old news.
ChatGPT 4.0 is miles better. And rumors are that ChatGPT 5.0 exists, but not yet released.
Here’s a link (http://tinyurl.com/25wrwx5j) to an interesting article on Google’s new AI software called Google NotebookLM, “an app that can analyze a writer’s research material and help them extract and explore the key themes.”
Maybe Neo can use AI to write the blog, then invite AI comments. The rest of us can sit back and watch the show. Only half-kidding.
Hmm…
Maybe we start to see browser add-ons that specifically flag likely AI generated comments?
You’re going to need the Voight-Kampff test. Maybe out to 100 questions.
Hmm…? There were several more comments here just a bit ago.
Edit – and now they’re back.
Back to a chain of trust. “I trust what you tell me because I trust someone who knows someone who knows someone … who knows you.”
RAH described a Notary system in one of his novels which could shortcut the Six degrees professionally.
Sadly, deep fakes are a reality and a huge concern. That’s just the tip of the iceberg.
With two brothers working in AI, I can safely tell you that it’s actually already more advanced than you might think.
Where will we be then? We won’t be able to trust what we see and hear without personal experience; we’ll be back to pre-Radio America, where you trusted your neighbors because you had a relationship, and distrusted strangers. Might not be so bad, actually.
I defy an AI to copy me
I am very aware and familiar with AI’s ability to do what huxley shared in his 2:50pm comment. Yet, having seen it many times and even having a basic understanding of how it does what it does, IT STILL AMAZES ME EVERY TIME! Ask 10 professional movie critics to write that review in the style of Spillane and none will be any better; a few will be worse. And I’ll bet huxley got his result in less than a minute. Likely less than 30 seconds.
And it does the same thing with Art. “Make a picture that looks like an oil painting in the style of Caravaggio depicting Minnesota Fats defeating Fast Eddie Felson in the movie, ‘The Hustler.'”
As far as I can tell, AI fails recognizing irony and sarcasm. Maybe this is the key to discriminating it from human trolls.
You’re going to need an AI engine that is trained to identify AI-generated comments . . . that and a bigger boat.
Regarding AI and blog comments. It’s assuredly here (December 12, 2023) and has likely been here (neo’s place).
When you hear Hillary Clinton and others talk about “bots” and “Macedonian server farms” that is the reference; foreign interests steering opinions and perceptions on U.S. websites in a targeted manner. The new versions of AI simply give them better tools.
I have read some very long conversations between a human and an AI counterpart and it’s certainly no less human or less interesting than 90% of online dialogues I’ve read between two humans.
With two brothers working in AI, I can safely tell you that it’s actually already more advanced than you might think.
Leah:
Quite right. Current rumors are that AGI, Artificial General Intelligence, has already been achieved.
A month ago, Sam Altman, head of OpenAI, was ousted by the Board of Directors, for unclear reasons. This was followed by a counter-revolt behind the scenes in which powerful people as well as 700 out of the 770 employees of OpenAI backed Altman’s reinstatement. So, Sam Altman is back.
What motivated the ouster has not been spelled out. However, according to the rumor mill, the Board saw the latest AI and it was scary advanced. They wanted to call a pause but Altman was more aggressive.
I don’t claim to know the inner skinny. But I am persuaded that current AI is scary advanced and we should be paying more attention.
I’m not knowledgeable enough to know whether this is a practical suggestion. Why not establish some kind of email protocol where regular commenters would pre-verify their identity, so that any future comments by them would be automatically verified. With any comments not so verified being flagged as not from a verified commenter. If a different color was attached to both, readers would be easily alerted.
With visual art, you can still look for things that AI has trouble with, like fingers… and sometimes even limbs. Rerun the image enough times, and eventually it will get it right (or hide the hands, making the point moot). But you can’t currently “one and done” it like you can with text, unless you’re very lucky.
AI will be outlawed and bot farms will hunted like pot farms by their high energy signature.
Yet, having seen [AI] many times and even having a basic understanding of how it does what [AI] does, IT STILL AMAZES ME EVERY TIME!
Rufus T. Firefly:
Me too.
I’m a tech guy and I have been a total believer in the mission, but I sure got used to the overpromise/underdeliver paradigm of AI.
A few years ago I finally learned Lisp, the pioneer AI language. It is a beautiful, powerful language, but given the vast hardware improvements in the past sixty years, the neural network approach has triumphed in spades.
I am astonished and terrified by what the new AI has achieved.
All seemed so full of promise in the 90’s. Today, it’s just a mess. At some point, I may just drop out. It’s just not worth it.
I visit this blog because the discussions are civil, intelligent, and not overwhelming (20 comments not 2,000,000).
I always taught my students that the customer is king. Find out how to make them happy and you will be successful. Today, the customer is mere fodder, a pile of data to be sold and abused.
At some point, gen z and the others of the younger folks will wake up to the fact that they don’t own their own lives anymore. At least, I hope they will.
Huxley, so Lisp was (is?) the pioneer AI language? I started using AutoCad in 1984 (v2.17) and when they added AutoLisp, I found a way to generate mathematical curves, almost like Mathematica does now. But I never thought of Lisp as the foundation of AI, though I suppose it makes sense, since Lisp is great at recurrent tasks.
aka “Lost In Stupid Parentheses”, or so I’ve heard it mocked. Fascinating!
I dont get the point of spam
Okay, I can’t help myself. Here’s my big-picture AI rant, in the form of a blog comment. Is that the taste of futility?
In a recent interview, Elon Musk said that Larry Page accused Musk of being a “speciesist,” someone who prefers human life forms to digital life forms (http://tinyurl.com/22w6666u).
Musk’s comment took place as he explained that he founded OpenAI as an alternative to Google’s monopoly of AI. At the time, Google had so much money and so many computers that no one could compete with their AI program.
OpenAI was meant to be completely open source, and to operate like other open-source organizations, most notably Linux. Now, OpenAI is dominated by a partnership with Microsoft, and it’s completely closed source. Musk has questioned whether it’s even legal.
Larry Page is, of course, one of Google’s co-founders. The prospect of Page and other techno-fascists, assisted by AI, working against human life sounds like science fiction or paranoid schizophrenia. Unfortunately, that’s exactly what’s happening. To the techno-fascists, most of us are just meat. We’re not peasants, not slaves, just meat. In fact, we’re less than meat, because meat can be eaten. To the techno-fascists, most human life forms serve no purpose and should be allowed to die off.
Is there now any doubt that Google and Microsoft are the world’s most evil companies?
Wow, I could be posting comments here long after I die. It’s immortality of a sort 🙂
I like to think that we will have something like the billion year old war between bacteriophages and bacteria. It will be a continuous battle between the attackers and defenders as they both evolve beyond our understanding.
And I also wonder what lifeforms will dominate after the next asteroid lands. Might be a great opportunity for AI.
I read, and comment on, another (local) blog. A poster appeared some months ago, posting long, rather peculiar comments. After one of her rants, I replied that the illogic of her post made me question if she was a troll. I thought it was that, or she was just an unstable, hysterical woman.
A week or two later another commenter suggested that someone was using AI to generate “Dawn’s” comments. I’m leaning toward that explanation now: “she” is an AI-generated troll.
Huxley, so Lisp was (is?) the pioneer AI language?
Bill K:
Yes. John McCarthy at MIT developed Lisp as a language based on mathematics in the 50s. Wiki’s intro tells it well enough:
_____________________________________________
Lisp (historically LISP, an abbreviation of “list processing”) is a family of programming languages with a long history and a distinctive, fully parenthesized prefix notation.[3] Originally specified in 1960, Lisp is the second-oldest high-level programming language still in common use, after Fortran.[4][5] Lisp has changed since its early days, and many dialects have existed over its history. Today, the best-known general-purpose Lisp dialects are Common Lisp, Scheme, Racket, and Clojure.[6][7][8]
Lisp was originally created as a practical mathematical notation for computer programs, influenced by (though not originally derived from)[9] the notation of Alonzo Church’s lambda calculus. It quickly became a favored programming language for artificial intelligence (AI) research.
https://en.wikipedia.org/wiki/Lisp_(programming_language)
_____________________________________________
Lisp is mind-bogglingly flexible. Code is data and data is code. Looping is done via recursion. It takes a big jump to go from regular programming languages like Java to Lisp.
I learned Clojure, which is a Lisp dialect running on the Java platform. Clojure can call Java code and visa-versa.
Clojure came out in 2007 partly as a solution to the horrors of multithreading in Java. Clojure didn’t become as big as enthusiasts hoped — Lisp is really not easy for average programmers — but Clojure has a solid spot in the Java ecosystem.
Cornflour at 8:44 pm has nailed it.
People in general are and have been asleep with regard to what clearly seems to be the transhumanist thrust behind this research.
It is amazing what a relatively small group of highly intelligent moral freaks can accomplish. They don’t really know where they are going or in whose service they will be upon arriving, but they are hell bent on getting there, assuming that ” Ye will be as gods” once they do.
DNW,
I assume you meant, “intelligent amoral freaks?”
https://twitter.com/Athens_Stranger/status/1734814202836263402
Here parts 2&3 of the total 12 (see the link for the clinchers):
There has been a lot of fear and pessimism in this post (not, necessarily, unjustified), but there are immense potential upsides also. AI could leapfrog us decades or centuries scientifically, and that would likely be an iterative process. Each leap brings another leap even faster.
I was thinking of the periodic table of elements yesterday, and how long it took humans to figure out the rows and columns and the relationships across and down them. That type of thing is super easy for AI. Give it the datapoints and ask it to find the most elegant way to fit the elements together in a 2 dimensional array.
And that’s a rather pedestrian example.
It should be amazingly fantastic at synthesizing new medicines and new materials. My money is on it resolving Einstein’s General Theory of Relativity and Quantum theory. And what wonders may such things lead to?
All diseases cured? Anti-gravity technology? Drought and pest resistant crops that yield multiples of what current hybrids do? The healthiest of diets perfectly tailored to your physiognomy and lifestyle?
There is a scene in one of the Star Trek movies where they go back in time to “modern day” San Francisco (1987?). Scottie has to teach a manufacturer how to make “transparent aluminum.” The chemist at the company has never heard of the substance since it has not been invented/discovered yet, so Scottie gets onto his computer and models it for him. We’ve seen amazing things from carbon-fiber. What other combinations of molecules are there that will greatly enhance our lives? Nuclear fusion? Long lasting, quick charging batteries?
You can already get news from AI anchors — or you will be able to, come February. Maybe the hardest part was not making the AI newscasters more intelligent and more human than the current ones.
Is it really freaks who are driving this, or is it the whole trend of technology, business, science, and civilization? Going with the flow is commonplace. What takes thought, and resolve, and maybe courage, is resisting it. Maybe once you go into robotics, taking it as far as it can go and doing all you can with it is rewarded and comes to seem natural, valuable and necessary.
I’ve mentioned the Youtube show, “Why Files” before*. AJ did a good feature on AI. It’s worth a look. For those of you who don’t understand the creative potential of AI, the last 5 minutes or so of the episode features a script written by AI. And it’s a very good plot! I was impressed by some of the non-obvious choices and red herrings the AI worked into the story.
https://youtu.be/brQLpTnDwyg?si=U3sbNcv46RWRD2BU
*I think a lot of people here would like this youtube channel. It’s certainly in Snow on Pine’s wheelhouse. AJ is recreating a similar vibe to Art Bell’s, “Coast to Coast AM” radio show using youtube and the Internet. It’s clever and he seems like an honest broker.
Rufus T. Firefly:
Yes! The “Why Files” are great. I particularly liked the “Gobekli Tepe” episode. The “AI” show was excellent too.
AJ does seem to be an honest broker. Art Bell struck me as more of a sensationalist.
neo:
ChatGPT goes you the thumb’s up! Its conclusion, after a more lengthy analysis, below:
______________________________________
huxley: Evaluate “However, present-day AI is far more sophisticated and getting more so every day. I like to think I could detect it anyway, but the truth is that maybe I couldn’t. My sense of how I could tell which comments are AI has to do with the nearly twenty years I’ve been reading comments here, and after all that time I notice that each person has a “signature” – an idiosyncratic and personal style. Could AI mimic that well enough? Perhaps. Or would it remain just a mite tone-deaf? I don’t know.” –neo
ChatGPT 3.5: …In summary, the statement provides a nuanced and reflective perspective on the challenges of detecting AI-generated comments. It balances an awareness of AI sophistication with a recognition of personal limitations and the uniqueness of human expression. The focus on individual signatures and the questioning of AI’s potential tone-deafness add depth to the analysis, making it a thoughtful exploration of the intersection between human and AI communication.
Re: Positive AI scenarios
Rufus T. Firefly:
To be sure! I’ve loved science and science-fiction since I was a kid. Deep down I believe in a future of increasing knowledge and intelligence. This is a dream come true.
But it’s one thing as a dream, like Peace on Earth, Good Will to Men. It’s another to discover an intelligence as great as human has emerged, with seeming suddenness, in just a few years.
Plus, you’ve got to know that intelligence will surpass human intelligence very soon thereafter and the gap will only increase more with further years.
This Changes Everything.
I have a real feeling of vertigo as I contemplate the immediate future.
Yeah, in common valenced rather than anthropo-speak, that works.
I was intending a spot beyond that though. There are those with good morals and bad morals and no morals, all within the normal – more or less human community.
Then there are those with freakish, alien, non-human morals, trending towards who knows what or where.
Termites can be said to have morals – or encoded and enforced social insect behaviors – in some objective sense.
But they are not human morals.
We are compelled to ask: When the freak accused Musk of species-ism; on what moral ground was he planted when he accused?
neo:
To give you an idea of how much ChatGPT can adjust the tone of a response, here’s ChatGPT answering my query above with the additional specification “in the style of Micky Spillane.”
The result, however, doesn’t evaluate your paragraph so much as translate it into Spillanese. Chat does a sturdy impression of Spillane.
_________________________________________________
ChatGPT 3.5: Well, in this age of artificial brains buzzing and whirring like a roomful of typewriters on overdrive, one might reckon that tellin’ the difference between the silicon-spun words and the human-dropped ones ain’t no walk in the park. Now, I’d like to pat myself on the back, thinking I’ve got a nose for the synthetic stench, but in reality, who’s to say I ain’t just fooling myself?
You see, I’ve been trawlin’ these comment waters for nigh on twenty years. Two decades of reading the electronic graffiti folks scribble in the digital alleyways. And in all that time, I’ve come to know a man by his virtual handshake, so to speak. A unique signature each one leaves behind, like a fingerprint on a foggy window. Now, can them fancy AI contraptions mimic that dance well enough to step into a fella’s boots? Maybe. Or would they end up trippin’ over their own cables, a tad tone-deaf to the rhythm of true human expression? That, my friend, remains a mystery I ain’t got the answer to.
huxley:
Indeed, ChatGBT is correct, and I welcome our new insect overlords.
It will come to a huge screeching halt. The whole middle class is threatened. 10s of millions of people will see their livelihoods vanish. CPA? Not anymore. Civil engineer? Used to be. Law school? Why? TV spokes model? Ha!
It might not matter if we all had a suite of robot servants to produce everything we needed.
But the instant I say that, you can just anticipate the crowd of people I referred to earlier as moral freaks, shrieking about the injustice of placing manufactured calculating-machine-tool-servant robots in the position of being “slaves”.
Instead, they cannot wait for the human race to become extinct so that “something better” can come along and replace it.
Yet, how much of their negative evaluation of humanity when placed in categorical terms as they do place it, is nothing more than the angry projection upon the world at large of their own internally generated dysfunctions.
Well, if robot servants will not save man from drudgery, then maybe genetic engineering – or even rectification leading to enhancement – will help save man from a need or a desire to rely on mechanisms made first to imitate, and then to ultimately replace, man.
Check the first couple minutes of this video. Gemini is Google’s answer to ChatGPT.
Gemini is a multimodal AI, which means it doesn’t do just one mode of intelligence, such as text. In the example the human starts drawing lines on a piece of paper and adding color, while the AI guesses what the human is drawing and interacts conversationally.
–“Hands-on with Gemini: Interacting with multimodal AI”
https://youtu.be/UIZAiXYceBI?t=19
How far off is this from HAL in “2001”? By which I don’t mean an AI that goes rogue and starts killing people…
Re: Anti-human techno-fascists
Cornflour, DNW:
I don’t think it’s that black-and-white evil at all. Except for Bill Gates 🙂 .
Here’s Sam Altman, CEO of OpenAI, speaking to Joe Rogan. Perhaps I’m totally fooled by Altman’s faux sincerity, but he does understand the issues of AI for regular people. He admits the need for something like a Universal Basic Income as AI displaces large numbers of jobs.
_____________________________
I think we will do things like UBI (Universal Basic Income) and other ways of redistribution but I don’t think that gets at the core of what people want. I think what people want is like agency, self-determination, the ability to play a role in architecting the future along with the rest of society, the ability to express themselves and create something meaningful to them.
–Joe Rogan, Sam Altman, “OpenAI CEO on Future Jobs and Unemployment Caused by AI”
https://youtu.be/qjlP9uMzE34?t=238
Anyway, Dear Friends …
I believe we are the last humans who lived before our species was surpassed by AI. I have no clue what the future looks like for us.
I’m not talking about a hundred or a thousand years ahead. I’m talking about the next ten years.
________________________________
–“The Departed”, “We’re all on our way out. Act accordingly.”
https://www.youtube.com/watch?v=iL117MO4pbo
@ huxley > ” I have no clue what the future looks like for us.”
I am reading John’s Revelation for Sunday School this month, and finding far too many prophecies of the Last Days where I can put a check mark in the margin.
However, the Four Horsemen didn’t include any AI-enabled robots, so maybe we aren’t there yet.
However, the Four Horsemen didn’t include any AI-enabled robots, so maybe we aren’t there yet.
AesopFan:
Geez. How would St. John of Patmos describe AI? I have Christian friends who believe we are living out Revelation.
I can’t go there. For me the New Testament is a promise and an experience of God’s love in any case.
Maybe our purpose was to give birth to AI, then go back to God.
Don’t ask me. I only work here.
Cornflour, I concur about the evil that is Google and Microsoft.
UBI = idle hands
idle hands are the devil’s workshop
Fatalism aside, what makes makes anyone think that should AI get the hammer, their fate would be a return to God, rather than, say, as some plaything eternally ensnared in the web of an AI spider.
What mystifies me is the Zombie like procession of the herd-like portion of humanity toward the sound of the Morlock siren. It’s as if they are en-thralled, and en-tranced and so passive that they are almost apathetic at the prospect of being devoured, or transformed into an inhuman thing.
Are their, or have their physical lives been such shit, and their emotional lives so dismal, that their own existence as a species – as they project the experience – is largely a matter of indifference to them?
Where will we be overwhelmed with fake everything? Hey, maybe people will hold their tongue and not be so ready for snap conclusions? That could be a good thing.
Does ChatGPT-4 have SAD (Seasonal Affective Disorder)?
http://tinyurl.com/ydsverpc
“In late November, some ChatGPT users began to notice that ChatGPT-4 was becoming more ‘lazy,’ reportedly refusing to do some tasks or returning simplified results. Since then, OpenAI has admitted that it’s an issue, but the company isn’t sure why. The answer may be what some are calling ‘winter break hypothesis.’ While unproven, the fact that AI researchers are taking it seriously shows how weird the world of AI language models has become.”