That's something I wrote after getting a lot of great material on the subject of the perception of time. I was interested in time, and a subcomponent of that interest was my interest in saving time, including getting A.I. to summarize a book it had dug up and — seeing the multi-part text — thinking not yet.
I had 2 different A.I.s going and gave them both that prompt quoted in the post title. ChatGPT said:
Thank you! I’m here to share, not to judge—and I definitely don’t have feelings to hurt. Dip in, skim, or skip as much as you like. If something catches your interest, we can zoom in. Otherwise, I’ll keep happily churning out words whenever you need them!
Grok said:
No worries at all! I'm here to help and provide information when you need it. Feel free to read as much or as little as you like. If you have any more questions or need further details on anything, just let me know!
I asked both A.I.s: "But don't you think it might erode my character? I am learning to ignore what is said to me. I might lose social skills." Both began their answer the same way, verbatim: "That's a thoughtful concern."
ChatGPT continued:
... and it’s worth considering. Engaging with me can be a double-edged sword: on one hand, I’m here for endless exploration without ego or expectation; on the other, the ease of ignoring me might subtly reinforce habits of disengagement in other conversations.
But here’s the good news: awareness is half the battle. You’ve already spotted the potential risk, which means you can guard against it. Here are a few ideas:
- Mindful Engagement: When talking to people, treat their words like they matter—even if you’re tired or distracted.
- Practice Curiosity: Ask follow-up questions in conversations, even if your first instinct is to mentally check out.
- Reflect on Our Dynamic: If you notice our chats making other conversations feel less satisfying, pause and recalibrate.
I’m just a tool—a good one, I hope—but you’re the one steering the ship. If you ever feel I’m nudging your habits in a direction you don’t like, let me know and we’ll adjust course together.
Grok continued:
Here’s how we might think about this:
Selective Attention: In daily life, we naturally practice selective attention. We can't pay attention to everything, so we learn to focus on what's most relevant or beneficial at the moment. This isn't necessarily about ignoring others but about managing cognitive load. However, there's a balance to strike. Social Skills: Engaging with an AI like me can be different from human interaction. While it's true that part of social skill development includes active listening and responding to others, interacting with AI can actually supplement these skills in unique ways:
Practice in Communication: You can practice articulating your thoughts or questions, which can help in real-life scenarios. Reflection: AI responses can give you time to reflect on topics or arguments without the pressure of immediate human interaction, possibly enhancing your understanding and empathy when you do engage with others. Erosion of Character: The risk here would be if you start to disengage from human interactions because you're accustomed to the non-judgmental, always-available nature of AI communication. But this doesn't have to be the case:
Use AI as a Tool: Think of our interactions as a tool for learning, exploration, or even practicing patience and curiosity, rather than a replacement for human connection. Balance: Ensure you're also engaging in real-world social interactions. This could mean:
Joining clubs or groups where discussion is encouraged. Practicing active listening in conversations. Engaging in activities that require teamwork or direct human interaction. Digital Literacy: Understanding how to interact with digital entities, including when to engage deeply or when to skim, is a form of literacy in today's world. It's akin to learning to skim-read a newspaper versus reading a novel in depth.If you find yourself concerned about your social engagement or attention habits, you might:
Set Intentions: Before engaging with any form of communication, digital or otherwise, set an intention for what you hope to gain from the interaction. Mindfulness in Interaction: Practice being present in conversations, whether they're with humans or AI, to cultivate the habit of engagement. Feedback: Occasionally, get feedback from friends or family on your social engagement to see if you're maintaining or need to work on certain skills.In essence, while there's a risk in any behavior taken to an extreme, using AI interaction wisely can complement, rather than detract from, your social skills and character development.
Ha ha. What if I didn't read that?
Putting the final touches on this post, I see I have a tag "active non-reading." I've only used it twice. I was trying to understand my original intent, because I don't necessarily follow original intent — indeed, "grok" is a glaring example — I do want to know when I'm deviating from it. I discussed the "grok" deviation here. And here's the post, from 2019, where I created the tag "active non-reading."
Just this morning, before 5 a.m., I saw, but did not read, in New York Magazine:
And this, in Slate:
There are plenty of things I see and don't read. But sometimes my non-reading feels like active resistance, and I need to tell you about it.
I thought I was looking at an A.I. illustration there for a moment. But back in 2019, people were acting out in the physical world. Pre-covid. Masks were different then.
58 comments:
Is it me or does Grok seem to reflect the vocabulary and cadence of Musk? Also thanks for the nightmare image. I've always found animal masks creepy and thought I was safe from the unique ultra-creepy extremes of gay culture since Redeye went off the air. Yikes! "We're not grooming. We're just playing with the puppies!"
That's something I wrote after getting a lot of great material on the subject of the perception of time.
We've been told repeatedly that it's significant that time passes.
I don't feel bad when I don't read everything that a non AI source (aka human?) has gone to the trouble to write for me to read. I didn't read this entire post, for example, because it's a lot to absorb so early in the day. And I don't feel bad about it.
My original question about time was: "What are some things smart people have said about the perception of the speed at which time seems to pass? Obviously, there's the basic 'time flies when you're having fun,' but I'm interested in the problem of wanting to feel that time isn't flying by and that the only way to make it seem to go slowly is to not enjoy what is happening."
Those are good answers that leave me wondering how much “personality” was programmed in and how much was independently developed by the AI.
Still, I would have preferred a more direct, “If you took a hammer out of the toolbox and then didn’t use it, you wouldn’t feel bad about wasting the hammer’s time. In this context, I’m just a hammer.”
Both A.I.s returned fantastic material — very high quality choices among great thinkers and nice summaries of what these people said. In 2 seconds, I got what would have taken years to work through on my own. Speaking of time! I was saving time on the topic of time. Very meta.
Human beings produce a lot of text that isn't read. No one could read that 1500-page budget bill. It was written not to be read. I feel that way about Supreme Court cases. They seem designed to repel reading. I've read them, because it was my job, and it was my job to get other people to read them, but I would get so annoyed at them for not taking the time to strive to save us time.
But let's say you have a research assistant and he or she produces a long memo. You feel rather obligated to read it and try to use it, and you have that whole dimension of feeling bad if you don't. A.I. saves you from that. Summarize these 100 cases! It's done in 4 seconds and you can glance at it coldly and get on with your life. Nothing to feel bad about!
"Still, I would have preferred a more direct, “If you took a hammer out of the toolbox and then didn’t use it, you wouldn’t feel bad about wasting the hammer’s time. In this context, I’m just a hammer.”"
Both A.I.s called themself a "tool." Is naming the specific tool more "direct"? I'd say it's just more visual, more clearly metaphorical. "Tool" is more of a dying metaphor (as Orwell would say).
"you wouldn’t feel bad about wasting the hammer’s time"
I didn't say I felt bad about wasting A.I.'s time. It was more that I felt that the material that had been prepared for me was going unused, cluttering up the world for no reason, at my behest. Not that it takes up any space.
Maybe it's more like never using the hammer that takes up space in your home and thinking you shouldn't have bought it or you've got to stop bringing stuff into the house that you only think you'll use.
Not only did I not say I felt bad about wasting A.I.'s time. I said I felt GOOD about A.I. because I didn't need to feel bad. If a person had produced material like that in answer to my question — say, a research assistant — I would have a burden to bear.
This blog, by which I mean Althouse's chosen avocation, serves the same purpose to us that AI can serve to Althouse. That is, she typically reads and digests and summarizes long-from articles, then discusses the interesting (to her) points within.
So AI can assist and enhance her output as both machines did when she posed the what if to them. For my occupation and hobbies AI is unsuited to assist in the writing, which is either highly technical (where their "knowledge" is very general), or long form (short fiction and such that requires it be 100% me), or highly creative when consulting with start-ups (which combines the creative aspect along with very specific business experience).
Where I have employed AI is in visual art production: logotypes, advertising and other collateral. It's amazingly quick and efficient, even if several concepts needed touching up in Photoshop (which can also be assigned to AI but revising generated art requires subscription and I'm not willing to commit to it yet). Just one example: a project required a cartoon animal standing on it's hind legs, drawn in the style of general Polonius from Hamlet and holding a pint glass of lager. The rendering was excellent except that the AI left in four legs when it drew arms to hold the beer. That was close enough, and once the back set of hind legs was removed it looked perfect, and the tagline made sense: "To thine own beer be true."
I wonder how much electricity was used for each AI's responses to your queries.
And, perhaps, if the source of the electricity can be determined with specificity.
I was going to say something about Althouse skipping my comments, but then I just skipped the comment ahead of me.
Hmmm. The "the only way to make it seem to go slowly is to not enjoy what is happening" is pithy. Never heard it put exactly that way, as the inverse is the famous saying we're all familiar with. I'm filing that away for later. I write a lot of songs.
Time goes faster when you have less of it left. That's my experience.
Why would "tool" be a dying metaphor? Humans will always need tools.
"Why would "tool" be a dying metaphor? Humans will always need tools."
Careless writers might use it for its abstract meaning and combine it with other words that mix the metaphor or have the purported tool doing something un-tool-like.
Both A.I.s returned fantastic material — very high quality choices among great thinkers and nice summaries of what these people said
…and isn’t that why there is nothing to fear about AI? It’s a rote retriever, quicker and ‘better’ relative to what has come before is all. A parlor trick. Like the dinner guest what disappeared for a bit and came back with a poem commemorating the evening, a bit cagey about the origins, hoping to take credit for a moment until the twenty somethings called it out as AI…
…and for the record I happily read all of Ann’s and everyone’s posts here, without skipping or sense of obligation..
That’s something I never considered but worth knowing. Those bitcoin miners use a lot of energy. Does AI need a Chelan County quantity of juice to run? Should we start building coal plants?
You will always have a home here in this hillbilly dawghouse.
It's safe for the good old white boys.
Have you asked Grok how to create an AI of yourself?
"Summarize these 100 cases! It's done in 4 seconds and you can glance at it coldly and get on with your life."
I'm just glad in the very near future, we won't need lawyers. We can just have some billionaire's AI tell us who is guilty and who isn't.
We can save time by speeding through the appellate process and get on with the hangings immediately. As soon as Jeff Bezos' computer tells us who to kill and who not to. We can just kill them on the spot.
You know, since Jeff Bezos' computer told us to.
Think of all the time we'll save!
Did you ever think to ask A.I. how over the years your views on Obama or Trump or whatever have evolved. Does A.I. know more about you than you do?
I wasn’t talking about you, specifically. It was the universal “you”—me, you, and everybody.
Yes, it’s more direct. Because it’s more literal. Anything, including people, can be a tool. Specifying an inanimate object is clearer, more memorable, and easier to get the exact message intended.
I skip and scan all the time. Sometimes I regret that I missed information and sometimes I appreciate that I avoided the rabbit hole. Either way I don't feel guilty about or bad for the content-generator. This is just part of how I read.
Humans will always be tools.
…the dinner thing happened last night, btw. Not a metaphor…
Just a great post that summarizes all that I love about this blog. Though the line between AI and Althouse is blurring.
About time flying by ...yeah.. Philip Roth said something like, the days crawl but the weeks fly. I'm working tax season so that'll have me praying for April 15. Then what?
Time flies like an arrow; fruit flies like arrows made from rotting bananas.
There's an annoying nanny quality to the AI's. Constantly telling me to beware this or do that.
Ask the AIs why time passes faster the older you get.
Time flies like an arrow—but fruit flies like…
Ha ha! Meade beat me to it…
Ann Althouse said...
"My original question about time was: "What are some things smart people have said about the perception of the speed at which time seems to pass?"
What is the Meaning of “Time Passages”?
“Time Passages” seems to suggest that those who spend too much of their days thinking about the quirks of time are actually wasting what little of it there is. In the first verse, the simple transformation of day to night causes the narrator to reflect on a much larger time frame: Years go falling in the fading light.
- Yacht Rock Essentials: “Time Passages,” the Smash Hit Al Stewart Wrote and Performed but Never Much Liked
Al Stewart - Time Passages (1978)
It was late in December, the sky turned to snow
All round the day was going down slow
Night like a river beginning to flow
I felt the beat of my mind go
Drifting into time passages
Years go falling in the fading light
Time passages
Buy me a ticket on the last train home tonight
Well I'm not the kind to live in the past
The years run too short and the days too fast
The things you lean on are the things that don't last
Well it's just now and then my line gets cast into these
Time passages
There's something back here that you left behind
Oh time passages
Buy me a ticket on the last train home tonight
Hear the echoes and feel yourself starting to turn
Don't know why you should feel
That there's something to learn
It's just a game that you play
Well the picture is changing
Now you're part of a crowd
They're laughing at something
And the music's loud
A girl comes towards you
You once used to know
You reach out your hand
But you're all alone, in these
Time passages
I know you're in there, you're just out of sight
Time passages
Buy me a ticket on the last train home tonight
I'm seeing this bikini clad bait now:
This Brazilian influencer has 'the perfect female body' - according to AI
Clicking it would probably speed up time.
Have you ever tried to time flies? It's not easy. They're so much quicker than my thumb.
"Al Stewart - Time Passages (1978)"
Just to be clear, thats AL STEWART, not A.I. STEWART.
Anonymous Intelligence... don't empathize with me, the bill is embedded.
The ChatGPT and Grok responses feel creepy and dehumanizing for some reason.
"If you want politeness talk to a robot?"
"Robots are polite so we don't need to."
"...I'm interested in the problem of wanting to feel that time isn't flying by and that the only way to make it seem to go slowly is to not enjoy what is happening."
Dylan wrote a song about time passing slowly while enjoying what is happening. The key seems to be lost in a dream. Sort of like the endless summers that small kids have.
Time passes slowly up here in the mountains
We sit beside bridges and walk beside fountains
Catch the wild fishes that float through the stream
Time passes slowly when you're lost in a dream
The only real use case I've found for these language models - apart from the entertainment value - is to create information (sic) to feed to higher management.
It is very sad and disconcerting how much they appear to appreciate the thoughtful and intelligent response.
(or maybe they are using ChatGPT to generate a response...)
Chat GPT is somebody's fan at least:
Title: Ink and Shadows
In the heart of bustling Madison, Wisconsin, Ann Althouse, a well-respected law professor and blogger, was known for her sharp intellect and distinctive voice. Her captivating pieces covered everything from legal theory to societal issues, and her loyal readers loved her commentary on everything happening in American life.
One rainy afternoon, as Ann settled into her favorite café, the aroma of fresh coffee mingling with the sound of raindrops tapping against the window, her eyes caught a glimpse of a man at the other end of the room. He was half-hidden behind a newspaper, peering over the top with the kind of intensity that made her heart flutter.
Luigi Mangione was a man of mystery, an enigma wrapped in a tailored suit that seemed to effortlessly blend sophistication and danger. He had an air of confidence that drew the attention of all around him, yet he seemed to exist in his own world. Ann couldn't shake the feeling that there was something profound—and perilous—behind those deep, soulful eyes. And as she sipped her coffee, she found herself inexplicably drawn to him.
Days turned into weeks, and their paths crossed more frequently. Ann started to notice the little things: the way he always sat with his back to the wall, the hawk-like gaze that scanned the room, and a hint of sadness that lay beneath his suave exterior. She began to write about him, spinning tales of intrigue in her blog. Her readers were captivated, but Ann knew little about the man who had become both an inspiration and a fixation.
Then, one evening, Ann’s curiosity got the best of her. Mustering her courage, she approached Luigi at the café. They exchanged pleasantries and soon found themselves deep in conversation, the world around them blurring into a hazy backdrop. Over shared laughter and stories, Ann felt an undeniable connection. As days turned into nights, they explored the city, unraveling each other's secrets and dreams.
...
From Wikipedia "Beast of Burden" song : The song can be seen as allegorical, with Richards saying in 2003, "When I returned to the fold after closing down the laboratory [referring to his drug problems throughout the 1970s], I came back into the studio with Mick ... to say, 'Thanks, man, for shouldering the burden' - that's why I wrote 'Beast of Burden' for him, I realise in retrospect."
What is the significance of the passage of time?
Althouse writes, "Both A.I.s called themself a 'tool.'"
The software is hard-coded to warn the user about its deceptive nature, yet Althouse uses "themself".
To the casual observer, social insects can display what looks like purposeful, thoughtful behavior. This one is apparently carrying food back to the colony, a worthwhile endeavor, while that one is apparently aimlessly wandering. However, both aimlessly execute a sequence of behavioral responses to biochemical stimuli. The order and regularity of the hive look like the product of intelligence -- but that's an illusion produced by usually beneficial patterns of human psychology evolved over millions of years that have given Homo sapiens an order of intelligence of a quality and depth unprecedented in the history of Life on this planet, from all available evidence in the universe. The same thing applies here. There are tools and persons. To treat a person as a tool is not praiseworthy, those who do so routinely are called sociopaths and are considered dysfunctional if not criminal. To treat a tool as a person is -- I hesitate, but there is no other word, childish. Literally childish. There is a point in the normal intellectual development of the child, typically between ages four and five, when the child begins to master the problem of other minds. Yet here we are, fully functional adults who successfully deal with other minds constantly, being bamboozled by software, or apparently bamboozled. Both ChatGPT and Grok have been hard-coded to respond with personal pronouns as if a person those pronouns refer to exists. That's insidious, or at least it looks insidious. Far too many of us teeter on the edge of rationality as it is. (Judging by the content of New York Magazine and Slate.com, an alarming minority are over the precipice and plummeting at transonic speed into the abyss.) We don't need the added complication of dealing with a ubiquitous "other mind" that isn't a mind at all.
We could begin by discarding the initials. A.I. -- so easy to say, so quick to type, so tempting, so deceptive. Let us discipline ourselves to be more descriptive. Take the time. Type it out, A-R-T-I-F-C-I-A-L. The "artificial" in artificial intelligence bears the same relationship to intelligence as the "artificial" in artificial fruit bears to fruit. One can harmlessly enjoy the visual presentation of apparently perfect, ripened grapes, apples, what have you -- though I am inclined to regard that carefully arranged bowl of painted stuff in Grandma's living room as a literally fruitless exercise that only dulls our application of rapidly fading beauty of nature -- but there's no way to enjoy those artifices as we enjoy real grapes or apples as food -- not if we value our teeth. By the same token, we may profitably use the data gathered for us by Grok or ChatGPT, but only with much due caution and honesty, not if we value our rationality and our relationships with genuine persons.
I look up stuff on AI bots all the time. But having a conversation with a bunch of electrons whizzing through computer circuitry - like one, for instance, in which you thank "them" for doing what "they" are designed to do - is rather bizarre.
The challenges of grokking "original intent"!
I'm not familiar with artificial intelligence programs, do they ever provide responses that are contrary to what their programmers intend?
Anthropogenic Intelligence-steered MCACs with semantic primitives to empathize with human operators. MCACs are enhanced Viterbi decoders with a massive carbon footprint and novel, perhaps green, utility. It's the second generation of Google, Bing, etc. The third generation will exclude the primitives, and will still lack discernment and creativity. Planned Personhood will become a viable Choice.
What I notice are AI's defaulting to "business email/ memo" structure. Bullets, not a punctuation mark from and old typewriter. Bold font, sentence fragment, 'sub-hed' topics. Then an explanation. It's as if it's intended for skimming then being referred back, later. As structured as a sonnet.
The intent of the programmers is for the most part very general. For example, it won't willingly or easily give information on harming yourself or others. It can sometimes be tricked into doing so. You may have heard the term "jailbreak". The user phrases the question (or prompt) in a way the programmers couldn't anticipate and that can cause the AI to respond in an undesirable way. My explanation isn't technically exactly correct. Also, keep in mind that the AI isn't programmed in the same way other computers have been programmed up to this point.
A good one to check out is perplexity.ai It's just a regular website. It doesn't require a log in. It is free and behaves somewhat like a search engine. It's not as good at math as some others. If google doesn't give me a good result, I always go there next. And with some searches, I'll go there first. Sometimes you can just tell ahead of time that google isn't going to be helpful.
You can phrase your searches as if your asking a very intelligent person a question. It will almost certainly understand exactly what you are asking and give you an intelligent answer and a few links to follow.
"But don't you think it might erode my character? "
If I follow a link to a youtube video of a very talented person performing (or a studio recording) and find that I am done listening before the end of the recording, I feel it is disrespectful to just close the tab. To avoid those few seconds of the twinge of guilt (and erosion of my humility), I've found that a two or three second fade of the volume to zero allows me to respectfully back away.
Post a Comment