Open Thread - Thurs 04 April 2024 - AI? I'm not Impressed
AI? I'm Not Impressed
Matt Taibbi has recently written a couple of articles about AI which made me pay attention. The first started with talking about how Google's AI tool, Gemini, made images which were wrong, but racially diverse! Like everyone, Taibbi thought something like, 'Yea this is bad, but it won't affect me...' Until it did. Right away. He stopped examining the issues with Gemini's images and asked Gemini a question about himself. And it gave him a list of a few things he supposedly did which he actually never did. It mixed up true statements and writings with things it had, apparently, made up. It credited him with articles he never wrote; articles which no-one ever wrote because they didn't exist. And that is both maddening and scary for everyone.
This make me laugh in a rueful way: from https://chappatte.com/en/images/ai-and-future-work
Taibbi's second article, about AI and the website Naked Capitalism hit me even harder. I realized that I had recently experienced some of the same AI mistakes, although I hadn't understood at the time that those mistakes were the result of AI.
So what is AI doing that is so aggravating and scary? All kinds of things! Here's two. As detailed in Taibbi's first article, AI makes facts, images, etc, up out of nothing. In his second article about Naked Capitalism AI picks out tiny bits of Naked Capitalism's website, without context or 'thought', and makes judgments about that website which affect income, web traffic and more.
I realized when reading the article about Naked Capitalism that I'd recently been subjected to the same AI stupidity, but in a slightly different direction. AI searched the web at an insurance company's direction. It found research papers I had written, and decided I must be selling the subjects of the papers. Yea, selling roman forts, or something
Like lots of people I'm not very confident, or trusting, about the new AI revolution. In fact, I don't like it at all. I don't want a car using it, I don't want commercials using it, I don't want phone calls using it, I don't want things suggesting stuff I might like using it. I don't want it judging the content that I create, or the things that I read or like. I don't think it's trustworthy in its current form. I don't trust google, or apple or whatever/whoever is running it.
What do you think?
Thanks for reading! Here's the open thread - and remember, everything is interesting if you dive deep enough, so tell us about where you're diving!
Comments
Happy Thursday!
A couple of days ago, we almost hit 70 degrees here in the PacNW. Now, we'll be lucky to hit 52 or so today. Gotta go pick up tax docs from the accountant today. FUN! Yea, not so much. But it'll be a nice drive maybe. Then, lots of transplanting in the greenhouse. The peppers are 6 inches tall and want to be in bigger pots! So does the basil! I bet the peppers will have flower buds starting in the next month or so. *drool*
Hope it is and has been a great day for you all. And the coming weekend is good!
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
Hey good morning
-
-
Not sure how accurate it is, but yesterday saw packaging
for an electric toothbrush that advertised AI on it.
Why anyone would want a computer connected toothbrush is
way beyond my ability to reason.
Thanks for the OT!
question everything
AI on a toothbrush...
good grief! I know everything seems to be connected now, but that's just insane. I remember teaching a friend's little kid how to use old fashioned, that is for me normal, light switches. You know, the switch that goes up and down. Apparently the newer houses/apartments don't have them, you either say something to the system to turn on the light, or it turns on if your phone/computer sez so, or they might have a newer slider type switch.
So weird. Is it really that much more work to flip a switch by hand, rather than telling something to do something?
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
I've started to see
presumably-AI-"fluffed" resumes. Lots more buzzwords, including some coinages. Claimed expertise in technologies that don't actually exist, or didn't exist in the wild at the time the person claims to have developed said expertise. That sort of thing.
They are almost as much fun to read as the one I got years ago when the resume for an old subordinate of mine came across my desk- and in it, he claimed that he was the one doing my job at the time he worked for me, instead of me. I was very amused. That was a very short phone screening. I hate kited resumes almost as much as I hate kited checks.
Caveat emptor to all of us poor saps who have to try to hire people, especially young engineers. I wish I could retire, because I really *am* getting too old for this.
Twice bitten, permanently shy.
Resumes...
'enhanced' (yea, not really) by AI? I never even thought of that aspect, but of course, it makes sense. Gads. Everything seems to be fake nowadays. I wonder if our parents felt the same way as the world changed around them. I can't imagine lying on a resume. I mean, it's going to be found out, isn't it? And isn't being a liar a death blow for employment changes? Err, except for politicians and some CEOs, I guess.
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
AI is real stupidity
HI all, Hey Sima!
Hope it's all good out there!
The problem with artificial intelligence is that it is REAL stupidity.
To give up thinking for anything or thought that comes along, and encourage this is not intelligent.
This is what they call intelligence now... a friend sends me screenshots from a newsfeed, to make make me fall out of my chair or bang head on desk...
So we need picture of talking pet birds? What do we have here? A begging juvenile fledgling WILD Barn Swallow (or near relative) and adult. Which are of course ILLEGAL to keep. It is idiotic.
What about that AI 'valhalla' video game image that showed up for a day or so at MS10 opening screen after an update two weeks ago, and quickly disappeared, back to the photos that had been being used. Do they think everything artificial is better than anything real? What a world to live in, but please leave me out of it.
I gotta fly and will be slammed all day into the noche...
Happy trails all!
edit - to change an B to an A.
We cannot solve our problems with the same thinking we used when we created them.
Look deep into nature, and then you will understand everything better.
both - Albert Einstein
That talking bird pic is nuts...
It's OBVIOUSLY not of pet birds, and... did anyone with a mind actually look over the post and approve it? Probably not, I guess. As you said, AI doesn't have a mind or real intelligence.
Hope things are going great for you! After getting the taxes today, I kinda took a break. Did a lot of online stuff and sorted potatoes for planting. Ohh boy.
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
Real Intellegence is more valuable
Pluto, I think uses AI to help with research and other tasks and finds it helpful.
The use of AI for narrative control is very much in place. The censorship industrial complex with real people (CIA et al) involved is also a threat.
We are the target.
Thanks for the OT!
“Until justice rolls down like water and righteousness like a mighty stream.”
Yes, we are the target.
I think that says just about everything. AI to control the narrative and set things up like whoever in power wants. Bleh.
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
Good morning Sima et. y'all. Thanx for the OT and toon, Sima.
Got back from Pismo yesterday afternoon, unloaded and unpacked and put away and all that and work up this morning to rain. We had, of course, used up all our perishables that we could before heading south, and I'm cooking this week, so I get to run down to the grocery store to restock. Pfah!
Somebody with the time to do so should train an AI on the archives of The Onion, Harvard Lampoon, Cal Pelican, Babylon Bee, and as many such sources as one could find and then loose it upon the world.
Meanwhile, Sputnik has a "water is wet" article up:
that is worth a quick read. https://sputnikglobe.com/20240404/75th-anniversary-nato-exists-to-respon...
So, we left in March and returned in April and thus have a bunch of eom financial shit to attend to as well, so I guess I'd best get into gear.
be well and have a good one
That, in its essence, is fascism--ownership of government by an individual, by a group, or by any other controlling private power. -- Franklin D. Roosevelt --
AI on the Onion archives...
and more. That would be interesting. Wonder if that's the source of some of the 'facts' Gemini and other AI machines expound?
I hope the shopping went well. And the financial stuff. My tax stuff went pretty well, nothing unexpected at least.
Have a good weekend!
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
I believe that AI is much more insidious and evil
than any moral, ethical person could imagine.
I absolutely believe that AI is currently being used to scrub any "undesireable" information from easily available research sites.
Perhaps an example is in order. Our state dept reps are in "negotiations" with the leaders of Niger to force the presence of a drone base in their country. They are having trouble getting support in Niger for their efforts.
A day or so ago a remotely piloted airpane did a suicide attack on a supposed Ru factory far from the Uk front. Hah! According to the snooze... they missed, but they were close. Bullshit. I think they hit what they were aiming at.
A dormitory housing African exchange students at a U.
Who could be so evil as to murder chidren to achieve an end? Perhaps someone who owes allegiance to a clan instead of humanity.
When I looked into the cv of our state reps I can't find much in the way of background anymore. Just pure whitewash.
I'm glad that I am reading the literature from before our coup d'etat. Finished Steinbeck, London, working on Hemmingway.
Be well, please.
Good example...
I think AI is being used unethically as well. And yea, it's probably scrubbing info from websites, info TPTB don't like. It's really infuriating how many normal people, companies, whatever, depend on AI and don't look at what it's actually telling them.
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
Thank you,
Based on some earlier info you provided I believe you are grappling with the same issues that I have been. In your case, tho, I think you are much more connected to the heritage. In my case I didn't have much connection to the land and in that locale legal protections weren't worth the paper they were written on. The developers and bankers owned the system and the gov't. It's not as easy for them in WA but when money is involved, the grifters will spare no expense that they can write off.
Please, please, be well.
When I started on my literary project it was to have a source of reading material that hadn't been crapified over the last few decades. Boy was I slapped with evidence of my naivete. The whitewashing has been rampant for a long, long time.
That whitewashing project has been infused with steroids by removing much of the manual efforts. Seeing that the players in our current societal upheavals have been suddenly cleared of any possible bias or conflict of interest by removing the evidence... wow.
What set this off for me was a linkdin bio someone posted to Moa about Evelyn Farkas and how different it was from the wiki.
Oh, to be cursed with memory. What hurts so much is the loss of intellectual value and potential progress.
Be well....
Please be as well as you can, yourself...
I do think we have a lot in common. Maybe most of us do. The fights over wills, land, finances, inheritance, it's nuts. And yes, TPTB want to take it as much as they can. Bleh.
Be well, it's the best we can do right now.
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
And just this moment, I had
a spam come-on pop up in my work email string: an invitation to a paid seminar entitled, and I quote,
Imagine my utter joy at the thought that AI-generated Excel Visual Basic code will be manipulating data contained in spreadsheets. Like my tax return, or my (nonexistent) retirement account, or the county voter registration data, or the targeting coordinates for our nuclear arsenal. What a great friggin' idea.
There are actually lazy idiots who will pay good money for this, and then wonder where the bugs come from. Not a fan.
Twice bitten, permanently shy.
AI sucks.
I get all kinds of offers for programs that supposedly helps lawyers get rid of secretary/receptionists. Let a bot screen clients. Sure. And they program them to have whatever friendly and warm voice you select. Right. They sound like prostitutes. Wonderful.
Got court this afternoon, but wanted to drop in and say high, and this is a fun OT, sima!
"We'll know our disinformation program is complete when everything the American public believes is false." ---- William Casey, CIA Director, 1981
I can't understand...
why get rid of the office crew? Why? People want to see and be with people, want to interact with people. I would freak if, for example, I went to my tax accountants and they'd gotten rid of the receptionist and had an automated document spewer (or whatever) to give me my tax returns. And if I had to ask a question, I got an automated non-answer.
Ahh well, hope court went well and have a good weekend!
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
I attended a course in the
It is not a defense to excuse yourself from liability because the AI was wrong.
AI is being used to write legal briefs on the theory it saves time and money. The problem is that an AI produced legal brief can cite cases that do not exist, cite a case in the wrong year, misstate which appeals court heard the case, and on and on.
The State Bar sees the shit show, but says to keep an open mind as to the potential. I guess they do not, as a body, want to project they are resistant to new tech. Still, they say if you use AI, understand that no malpractice insurance will cover it. The AI exclusion is in every single policy. You are on your own.
"We'll know our disinformation program is complete when everything the American public believes is false." ---- William Casey, CIA Director, 1981
So, AI writing a legal brief...
supposedly saves time and money, according to the experts, whoever they are . But it saves time and money only IF humans don't have to check over the brief, make sure the citations are correct (and actually exist) and all the rest. So, just let humans write the dang brief. heh.
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
I am not sure but I think that Genocide Joe has been a guinea
pig for AI.
Where would he be without his teleprompter?
Hahahahaha!
So true!
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
Sci-fi has ruled it part of dystopia.
I read an essay long time ago where the author claimed that sci-fi writers helped people understand that nuclear war was the end of humanity as we know it. The thus helped people back away from the use of nukes.
In the same vein, one of the characteristics of dystopian societies as imagined by current sci-fi is the use of AI in various formats even if the term AI was or is not used. Two films which show machine controlled societies and the eventual ruin of humans was the Forbin Project and yes Terminator. Probably add to that THX1138.
And on a lighter note-
Our friends in Israel are having a right old party over there. They've cancelled all military leave, put all their embassies on high alert, pulled some people out to undisclosed locations, and started jamming GPS signals- with the conjecture being that this is to make the incoming Iranian missiles or drones go off course.
Well, whod'a thunk it. I guess that there really are limits to what they can get away with. I hope that our "leadership" will be smart enough to let them take their casualties on their own damned dime, without shedding our blood. Presumably, the Israelis started that fight with Iran of their own accord- let them finish it.
And if they started it because we told them to, then fuck 'em: throw them under the bus like we do everyone else. The "christian" dominionists I have to work with are all agog that we're finally immanentizing the eschaton, and I'd really prefer that they be disappointed.
Gonna hit the liquor store tonight on the way home.
Twice bitten, permanently shy.
Israel...
goes to war with Iran. How stupid can they be? How stupid??
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
When I think of institutions currently
...both public and private, my subjective sense is the they are becoming further unable to deliver the services or material in kind or quality that they are purportedly designed to offer. This is not because of an actual scarcity of resources but an artificially imposed system of exploitation. So I don't think just about the potential of the digital world, or virtual world to create false narratives and libel those who don't obey the dictates of the blob but also the dangerous potential of AI for other institutions that lately appear to just dictate their terms to a relatively powerless public.
AI's potential could be used particularly in the banking, insurance, medical and administrative fields, but not limited to those alone, to create false historical accounts, records, and files to achieve goals that are unethical, obstructive, exploitative and destructive to the individual in particular and society in general. Not that this isn't already going on, but it will be accelerated. You agree that we may change the terms of this agreement at any time without notice to you.
語必忠信 行必正直
Good points
I think AI is already being used as you describe. Perhaps not with everything, but they are going to be doing that. Sci-fi world, here we come.
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
There are dozens, hundreds, of different AI's
They are not search engines. Looking things up is not an AI feature, per se. Any search engine with algorithms that measure popularity and rank credibility is a much better choice.
People insist on comparing AI's to search engines, because that's all they know and understand. But this comparison will only lead you into a dead-end way of thinking. Please don't let that happen.
Every AI you may have heard of has been trained and tested by a large team of people, for a specific purpose, for a long period of time. It learns what it is given to learn, and it learns how to complete tasks, based on what it learns. For example, you could feed an AI every single email you have received and responded to over the past 5 years. After testing and refining, the AI could write a response to every future email you receive that requires a response. And the response would sound exactly like you.
That's not a very exciting purpose, but it establishes that an AI is not a black box that looks up stuff up for you. Instead, you must first explain to the AI how to help you achieve your purpose. You tell it how to collaborate with you by using "prompts."
This article explains what prompts are, then offers a list of advanced prompts you can use with a particular AI (Chat GPT) to transform it into your personal factotum. Sure, Chat GPT can answer a question by looking things up — but that is not what it was designed to do. It was designed to have an ongoing conversation with you, so you can work together to find the solutions or get the job done. Chat GPT AI is free. Go get it. All you need to know is how to ask a question. If you start your AI conversation using one of the prompts from the linked article, above (scan them all) you could have a trained expert or brilliant advisor by your side in a matter of minutes.
This is how the world works, now.
We've turned a corner.
We have had this discussion before
-
-
writing is a creative art form. The 'prompts' you mention can be both
spontaneous and densely considered with different outcomes.
Thought creation is an internal blending of expression and almost
mystical connections. Programming a machine to duplicate this
process enables a false representation of imagination. The more people
lose critical thinking skills, the shallower the horizons become. Traveling
from point A to point B entails being open to subroutines which changes
the format. YMMV
question everything
Agreed.
"We'll know our disinformation program is complete when everything the American public believes is false." ---- William Casey, CIA Director, 1981
Yep.
Driving down a road, cleaning the floor, making dinner, staring out the window, all places where thoughts and inspirations just arrive.
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so
Here's my favorite AI artist Kelly Boesch doing just that:
.
Writing prompts and collaborating with an AI to create a scene from their combined imagination.
.
I hope viewers don't suddenly lose their critical thinking skills.
I enjoyed seeing the beautifully and uniquely portrayed birds
I don't know whether they exist only in an imaginary sense or in reality, even though I've seen documentary images of extraordinary looking birds. From its inception photography has always had the capacity to both reveal reality and alter it. I don't think though that the background music elevated the effect, as it sounded more like elevator-music to me. Generally, I think real nature holds the answers.
AI's can create highly detailed images from simple prompts
Soundtracks, if any, are added later and are not created by AI's. AI art is generally pictures and stills. This artist, Kelly Boesch is learning to animate the images, using other AI tools.
I watched some other of her videos on you tube
She is very prolific, and her enthusiasm for the genre is obvious. It reminded me a bit of psychedelia art of the past. I hope though, that this is not the only future of art and other things.
Count me out.
I write.
I write songs, and short-shorts, and essays. And rants, in comments. And sometimes, I design stupidly high-tech products that the market demands, as a corporate whore. And so I breathe, sleep, eat, and excrete, with a roof over the head of myself and the one person I really love in this life; and some critters as well. Sometimes, randomly, I even create Art. Nobody cares, but it feeds my soul.
Life is too short. And I can not, must not, and will never surrender my soul to a machine: one which will presumably vomit forth an acceptable version of *me*.
Sorry, you gotta deal with the guy with the foul mouth and an intrinsic hatred of *falsehood*. Warts and all. Fuck ‘em if they can’t take a joke.
Nor will any machine ever be able to do what I do, or you do, or we do together, as a community.
There was a point here, but it is lost to the noise. I do not expect to wake up alive tomorrow- but I haven’t expected that since my single-digit ages. That was so far back that I can no longer recall the feel of it. If sensible, gentle, non-coruscating, non-actinic, non-flesh-melting, non-world-destroying light reaches my eyes tomorrow morning: it will downright astonish me.
But all things are possible, amirite? I’ve been astonished, daily, for over 60 years. Could have used that energy for other purposes, I suppose, but it is what it is. So, perhaps, we will meet again on the morrow. Yay.
Peace (and, astonishingly, I mean it).
Twice bitten, permanently shy.
Don't apologize.
I like the way you think. You've earned the right to your skepticism. And you are very fair about it, and self aware. Plus, once you've drawn your line of disbelief in the sand, you are sometimes willing to jump back in and mix it up. Generally, you bring more to the table than you take away.
Hi Pluto
I have a friend who was just encouraging me to try out Chat GPT to answer some practical questions I have regarding an issue I am dealing with, but finding difficult to get concrete answers to. Using traditional means, actually talking with people informed in the area, usually turns out not to provide concrete answers, but more questions. I end up thinking maybe I just haven’t found the person who has the overall knowledge to provide the clarity I need. I also question whether I would find concrete answers from the amalgamation of questions regarding the subject in Chat, even if I could articulate it clearly. There always seems to be some loophole (blip) or another that can weave a web of confusion.
I'm sure Chat GPT is good for many purposes, but for others maybe more confusing. I guess it doesn't hurt to try. I might.
I was frustrated with Chat GTP for a long time.
I asked single questions and the responses were unsatisfying and conditional. It took a while to understand that Chat GTP is designed to inform through dialogue, which may incorporate establishing the rules of inquiry and understanding the the range and depth of the topic. (Chat GTP remembers the conversations and can pick it up again where you left off.)
Dialetics are fundamental in philosophical reasoning. Dialogue is a method of arriving at the truth. At east that's the idea. However, Chat GTP has limits programmed in that seem to avoid truth (or controversial conclusions). I've argued with the AI over this. Which is a somewhat crazy thing to do. You and I think alike on this issue:
The article I linked to in my prior comment is a list of prompts to enter into Chat GDP prior to discussion. In truth, some are actually hacks that turn off the AI's politically-correct behaviors that are programmed into the app. These prompts set up constructive AI behaviors and roles to facilitate inquiry into specific topics.
This is a powerful shortcut. I would not have mentioned this AI at all if I hadn't found these hacks to pass along. Suggest you look through the prompts. Prompts put you in the driver's seat. Just asking questions gives the advantage to the AI.
Thanks for the tips Pluto
I'll let you know what I experience if I give it a try.
Thanks for the OT Sima
The question is, “would we be in a better place if Presidents were real, or AI?"
Just joking, but who knows, with what we've got.
I wonder who or what...
the AI president would be supporting? And if all leaders were AI generated, then ... would they form a group and take over even more stuff? Gets very science-fictiony, doesn't it?
If you're poor now, my friend, then you'll stay poor.
These days, only the rich get given more. -- Martial book 5:81, c. AD 100 or so
Nothing ever changes -- Sima, c. AD 2020 or so