If you’ve watched any Olympics coverage this week, you’ve likely been confronted with an ad for Google’s Gemini AI called “Dear Sydney.” In it, a proud father seeks help writing a letter on behalf of his daughter, who is an aspiring runner and superfan of world-record-holding hurdler Sydney McLaughlin-Levrone.
“I’m pretty good with words, but this has to be just right,” the father intones before asking Gemini to “Help my daughter write a letter telling Sydney how inspiring she is…” Gemini dutifully responds with a draft letter in which the LLM tells the runner, on behalf of the daughter, that she wants to be “just like you.”
I think the most offensive thing about the ad is what it implies about the kinds of human tasks Google sees AI replacing. Rather than using LLMs to automate tedious busywork or difficult research questions, “Dear Sydney” presents a world where Gemini can help us offload a heartwarming shared moment of connection with our children.
Inserting Gemini into a child’s heartfelt request for parental help makes it seem like the parent in question is offloading their responsibilities to a computer in the coldest, most sterile way possible. More than that, it comes across as an attempt to avoid an opportunity to bond with a child over a shared interest in a creative way.
https://www.youtube.com/watch?v=wfEEAfjb8Ko
Now we need the machine to write a handwritten letter, and sign it. To complete the effect of genuine human connection
I saw a movie the other day, and all of the ads before the previews were about AI. It was awful, and I hated it. One of them was this one, and yes… Terrible.
It would’ve been cooler if they used it to write a cool PDF page of info and stats on Sydney McLaughlin-Levrone
Or finding/buying plane tickets at the best price by searching all the sites
But that would imply that it can be relied upon for accuracy.
Thank you! The ads from everywhere this Olympics have been so fucking weird. I even started a thread on mastodon and this ad was on it. https://hachyderm.io/@ch00f/112861965493613935
Ever since I moved to an ad-reduced life, everything has been nicer. I can’t completely escape them, they are everywhere. But minimizing with ublock and pihole helps, then only using video services that don’t have ads. Unfortunately, a lot have added ads, so I have quit those. I’ll pay extra for ad-free, just because ads make my life so miserable.
I can’t watch broadcast TV, it’s too irritating. I can’t browse the web on a device outside my network or phone. I don’t use free apps. Hell, I don’t listen to the radio.
I like to think it has made me a calmer person.
They were always weird but it is getting to the point where even normies are taking notice.
All that sex traffic that occurs for their event alone make it an abomination.
My best friend, the Uber driver, which I prefer to shut up all the way home. But hey, what are friends for, he keeps me hydrated!
Yeah, fully agree. This is one of the reasons big tech is dangerous with AI, their sense of humanity and their instincts on what’s right are way off.
Oozes superficiality. Say anything do anything for market share.
This is one of the weirdest of several weird things about the people who are marketing AI right now
I went to ChatGPT right now and one of the auto prompts it has is “Message to comfort a friend”
If I was in some sort of distress and someone sent me a comforting message and I later found out they had ChatGPT write the message for them I think I would abandon the friendship as a pointless endeavor
What world do these people live in where they’re like “I wish AI would write meaningful messages to my friends for me, so I didn’t have to”
The thing they’re trying to market is a lot of people genuinely don’t know what to say at certain times. Instead of replacing an emotional activity, its meant to be used when you literally can’t do it but need to.
Obviously that’s not the way it should go, but it is an actual problem they’re trying to talk to. I had a friend feel real down in high school because his parents didn’t attend an award ceremony, and I couldn’t help cause I just didn’t know what to say. AI could’ve hypothetically given me a rough draft or inspiration. Obviously I wouldn’t have just texted what the AI said, but it could’ve gotten me past the part I was stuck on.
In my experience, AI is shit at that anyway. 9 times out of 10 when I ask it anything even remotely deep it restates the problem like “I’m sorry to hear your parents couldn’t make it”. AI can’t really solve the problem google wants it to, and I’m honestly glad it can’t.
They’re trying to market emotion because emotion sells.
It’s also exactly what AI should be kept away from.
But ai also lies and hallucinates, so you can’t market it for writing work documents. That could get people fired.
Really though, I wonder if the marketing was already outsourced to the LLM?
Sadly, after working in Advertising for over 10 years, I know how dumb the art directors can be about messaging like this. It why I got out.
Yeah. If it had any empathy this would be a good task and a genuinely helpful thing. As it is, it’s going to produce nothing but pain and confusion and false hope if turned loose on this task.
A lot of the times when you don’t know what to say, it’s not because you can’t find the right words, but the right words simply don’t exist. There’s nothing that captures your sorrow for the person.
Funny enough, the right thing to say is that you don’t know what to say. And just offer yourself to be there for them.
it’s completely tone-deaf of that’s their understanding of humanity. Similar to the ipad app ad that theluddite blog refered to here in his post about capture platforms. Making art, music and human emotional connections are not the tedious part that need to be automated away ffs
These seem like people who treat relationships like a game or an obligation instead of really wanting to know the person.
If I was in some sort of distress and someone sent me a comforting message and I later found out they had ChatGPT write the message for them I think I would abandon the friendship as a pointless endeavor
My initial response is the same as yours, but I wonder… If the intent was to comfort you and the effect was to comfort you, wasn’t the message effective? How is it different from using a cell phone to get a reminder about a friend’s birthday rather than memorizing when the birthday is?
One problem that both the AI message and the birthday reminder have is that they don’t require much effort. People apparently appreciate having effort expended on their behalf even if it doesn’t create any useful result. This is why I’m currently making a two-hour round trip to bring a birthday cake to my friend instead of simply telling her to pick the one she wants, have it delivered, and bill me. (She has covid so we can’t celebrate together.) I did make the mistake of telling my friend that I had a reminder in my phone for this, so now she knows I didn’t expend the effort to memorize the date.
Another problem that only the AI message has is that it doesn’t contain information that the receiver wants to know, which is the specific mental state of the sender rather than just the presence of an intent to comfort. Presumably if the receiver wanted a message from an AI, she would have asked the AI for it herself.
Anyway, those are my Asperger’s musings. The next time a friend needs comforting, I will tell her “I wish you well. Ask an AI for inspirational messages appropriate for these circumstances.”
Another problem that only the AI message has is that it doesn’t contain information that the receiver wants to know, which is the specific mental state of the sender rather than just the presence of an intent to comfort.
I don’t think the recipient wants to know the specific mental state of the sender. Presumably, the person is already dealing with a lot, and it’s unlikely they’re spending much time wondering what friends not going through it are thinking about. Grief and stress tend to be kind of self-centering that way.
The intent to comfort is the important part. That’s why the suggestion of “I don’t know what to say, but I’m here for you” can actually be an effective thing to say in these situations.
Ask an AI for inspirational messages appropriate for these circumstances.
Don’t need to ask an AI when every website is AI-generated blogspam these days
I would abandon the friendship as a pointless endeavor
You’re in luck, you can subscribe to an AI friend instead. /s
You’ve seen porn addiction yes, but have you seen AI boyfriend emotional attachment addiction?
Guaranteed to ruin your life! Act now.
Those AI dating sites always have the creepiest uncanny valley profile photos. Its fun to scroll them sometimes.
Don’t date robots!
Shut up! You don’t understand what me and Marilyn Monrobot have together!
Brought to you by the Space Pope.
Uhh “subscribing to an AI friend” is technically possible in the form of character.ai sub. Not that I recommend it but in this day your statement is not sarcastic.
The article makes a mention of the early part of the movie Her, where he’s writing a heartfelt, personal card that turns out to be his job, writing from one stranger to another. That reference was exactly on target: I think most of us thought outsourcing such a thing was a completely bizarre idea, and it is. It’s maybe even worse if you’re not even outsourcing to someone with emotions but to an AI.
The obvious missing element is another AI on Sydney’s end to summarize all the fan mail into a one-number sentiment score. At that point we can eliminate both the AIs and the mental effort, and just send each other single numbers via an ad-sponsored Google service.
Hey, my buddy’s work is already doing that! Management no longer has any idea what the company does, but they know how often you click. It boils down to a decimal number, which is what they really need. Higher numbers are better.
Which they will unceremoniously murder after it fails to get enough traction in a month after launch.
This! I was appalled when this ad played, suggesting that ANYONE comes out of that fictional scenario pleased is ridiculous. No one wants to receive a crappy AI-written email, ESPECIALLY when the primary topic is emotional. Using an LLM to write a message for a loved one tells everyone that you don’t actually care enough to write it yourself. And Google is putting their big check of approval on the whole scenario saying, “This is what we want you to use Gemini for.” Absolutely abysmal.
The ONLY version of this ad that makes any sense is if the parent writing the email is illiterate or has a medical issue where they can’t type. But I’d rather see them use AI to make dictation better and more powerful instead.
We’re all switching to Kagi Search and moving our email to ProtonMail or the like right? I don’t need this kind of crap in my digital tool kit.
Proton recently introduced an AI “writing assistant” for emails called Scribe and a bitcoin wallet sadly.
Hate to say it, but Kagi is not great. Both in results and in stewardship.
Honestly they could have avoided this by asking the daughter to input some legitimate sentiments and had AI help her express them.
Instead they offload the task entirely, removing any thought or sense of legitimacy.
Well said.
So in the spring I got a letter from a student telling me how much they appreciate me as a teacher. At the time I was going through some s***. Still am frankly. So it meant a lot to me.That was such a nice letter.
I read it again the next day and realized it was too perfect. Some of the phrasing just didn’t make sense for a high school student. Some of the punctuation.
I have no doubt the student was sincere in their appreciation for me, But once I realized what they had done It cheapened those happy feelings. Blah.
You should’ve asked Gemini what to feel about it and how to response…
That’s the problem with how they are doing it, everyone seems to want AI to do everything, everywhere.
It is now getting on my own nerves, because more and more customers want to have somehow AI integrated in their websites, even when they don’t have a use for it.
We created a society of antisocial people who are maximized as efficient working machines to the point of drugging the ones that are struggling with it.
Of course they want AI to do it for them and end human interactions. It’s simpler that way.
… But why did it cheapen it when they’re the one that sent it to you? Because someone helped them write it, somehow the meaning is meaningless?
That seems positively callous in the worst possible way.
It’s needless fear mongering because it doesn’t count because of arbitrary reason since it’s not how we used to do things in the good old days.
No encyclopedia references… No using the internet… No using Wikipedia… No quoting since language and experience isn’t somehow shared and built on the shoulders of the previous generations with LLMs being the equivalent of a literal human reference dictionary that people want to say but can’t recall themselves or simply want to save time in a world where time is more precious than almost anything lol.
The only reason anyone shouldn’t like AI is due to the power draw. And nearly every AI company is investing more in renewables than anyone everyone else while pretending like data centers are the bane of existence while they write on Lemmy watching YouTube and playing an online game lol.
David Joyner in his article On Artificial Intelligence and Authenticity gives an excellent example on how AI can cheapen the meaning of the gift: the thought and effort that goes into it.
In the opening synchronous meeting for one such class this semester, I was asked about this policy: if the work itself is the same, what does it matter whether it came from AI or not? I explained my thoughts with an analogy: imagine you have an assistant, whether that is an executive assistant at work or a family assistant at home or anyone else whose professional role is helping you with your role. Then, imagine your child’s (or spouse’s, I actually can’t remember which example I used in class) birthday is coming up. You could go out and shop for a present yourself, but you’re busy, so you ask this assistant to go pick out something. If your child found out that your assistant picked out the gift instead of you, would we consider it reasonable for them to be disappointed, even if the gift itself is identical to the one you would have purchased?
My class (those that spoke up, at least) generally agreed yes, it would be reasonable to expect the child to be disappointed: the gift is intended to represent more than just its inherent usefulness and value, but also the thought and effort that went into obtaining it. I continued the analogy by asking: now imagine if the gift was instead a prize selected for an employee-of-the-month sort of program. Would it be as disappointing for the assistant to buy it in that case? Likely not: in that situation, the gift’s value is more direct.
The assistant parallel is an interesting one, and I think that comes out in how I use LLMs as well. I’d never ask an assistant to both choose and get a present for someone; but I could see myself asking them to buy a gift I’d chosen. Or maybe even do some research on a particular kind of gift (as an example, looking through my gift ideas list I have “lightweight step stool” for a family member. I’d love to outsource the research to come up with a few examples of what’s on the market, then choose from those.). The idea is mine, the ultimate decision would be mine, but some of the busy work to get there was outsourced.
Last year I also wrote thank you letters to everyone on my team for Associate Appreciation Day with the help of an LLM. I’m obsessive about my writing, and I know if I’d done that activity from scratch, it would have easily taken me 4 hours. I cut it down to about 1.5hrs by starting with a prompt like, “Write an appreciation note in first person to an associate who…” then provided a bulleted list of accomplishments of theirs. It provided a first draft and I modified greatly from there, bouncing things off the LLM for support.
One associate was underperforming, and I had the LLM help me be “less effusive” and to “praise her effort” more than her results so I wasn’t sending a message that conflicted with her recent review. I would have spent hours finding the right ways of doing that on my own, but it got me there in a couple exchanges. It also helped me find synonyms.
In the end, the note was so heavily edited by me that it was in my voice. And as I said, it still took me ~1.5 hours to do for just the three people who reported to me at the time. So, like in the gift-giving example, the idea was mine, the choice was mine, but I outsourced some of the drafting and editing busy work.
IMO, LLMs are best when used to simplify or support you doing a task, not to replace you doing them.
This is exactly how I view LLMs and have used them before.
These people in these scenarios aren’t going ‘Amazon buy my gf a gift she likes.’
They’re going, please write a letter to my professor thanking them for their help and all they’ve done for me in biology.
I don’t know of anyone who trusts AI enough to just carte blanche fire off emails immediately after getting prompts back either.
The fear and cheapening of AI is the same fear and cheapening as every other advancement in technology.
-
It’s not a a real conversation unless you talk face to face like a man
say it in a groupwrite it on parchment and inkpen and papertypewritertelegramphonecalltext messagefaxemail. E: rip strikethroughs? -
It’s not a real paper if it’s a meta analysis.
-
It’s not it’s not it’s not.
All for arbitrary reasons that people have used to offset mundane garden levels of tedium or just outright ableist in some circumstances.
People also seriously overestimate their ability to detect AI writing or even pictures. That dude may very well have gotten a sincere letter without AI but they’ve already set it in their mind that the student wrote it with AI as if they know this student so well from 10 written assignments they probably don’t care about to 1 potentially sincerely written statement to them.
If people like that think it cheapens the value, that’s on them. People go on and on about removing pointless platitudes and dumb culturally ingrained shit but then clutch their pearls the moment one person toes outside the in-group.
It just feels so silly to me.
IT’S NOT ART UNLESS IT’S OIL ON CANVAS levels of dumb.
It’s not altruistic/good-natured unless you don’t benefit from it in any way and feel no emotion by doing it! You can’t help the homeless unless you follow the rules! You can’t give them money if you record it.
In the end, they still got that money. But somehow it devalues it because instead of raising two people up higher, you only raised one? It’s foolishness.
People also seriously overestimate other’s abilities and cheapen what their time is worth all the damn time.
-
I’m curious, if they had gone to their parent, gave them the same info, and come to the same message… would it have been less cheap feeling?
And do you know that isn’t the case? “Hey mom, I’m trying to write something nice to my teacher, this is what I have but it feels weird can you make a suggestion?” Is a perfectly reasonable thing to have happened.
I think there’s a different amount of effort involved in the two scenarios and that does matter. In your example, the kid has already drafted the letter and adding in a parent will make it take longer and involve more effort. I think the assumption is they didn’t go to AI with a draft letter but had it spit one out with a much easier to create prompt.
When i saw this ad a few days ago, my immediate, audible, response was " I guess we don’t need humans anymore. "
I think AI has already taken over and it’s putting out the ads.
Being a non native English speaker this is actually one of the better uses of LLMs for me. When I need to write in “fancier” English I ask LLMs and use it as an initial point (sometimes end up doing heavy modifications sometimes light). I mean this is one of the more logical uses of LLM, it is good at languages (unlike trying it to get it to solve math problems).
And I dont agree with the pov that just because you use LLM output to find a good starting point it stops being personal.
The problem with this is that effectively you aren’t speaking anymore, the bot does for you. And if on the other side someone does not read anymore (the bot does it for them) then we are in very bizarre situation where all sorts of crazy shit starts to happen that never did.
You will ‚say’ something you didn’t mean at all, they will ‚read’ something that wasn’t there. The very language, communication collapses.If everyone relies on it this will lead to total paralysis of society because the tool is flawed but in such a way that is not immediately apparent until it is everywhere, processes its own output and chokes on the garbage it produces.
It wouldn’t be so bad if it was immediately apparent but it seems so helpful and nice what can go wrong
Well, if you get anywhere with that fake facade, then it will catch up to you.
Better start reading nicely written English books while doing this…
Learning to use the tools available to you is not “fake” it’s being smart. Anyone who would be like “oh you recognize your weak point and have found and used a tool effectively to minimize it…you’re fired/get out of my life” is an asshole and an idiot.
If you use binggpt as a translator tool, and put a disclaimer that these are not your own words - kudos, you removed the need for a translator and the latency associated.
However, if you claim that you speak English and use this tool to create a false impression of proficiency, that is just usual lying.
I see your point and can agree in the cases where the tool won’t be available to you, or if there is an intent to deceive.
But to flip the script, I’m pretty good at spelling but even then there are words I fuck up the spelling and it’s caught by a spell checker. Am I a liar for submitting things without pointing out my spelling errors that a computer caught? Or is there a recognition that this is a common tool available and I’ve effectively used it to improve my communication, so this is just standard practice?
I would accept spell checker, for a few reasons: one - it doesn’t really change the meanings, or the words, just polishes tiny fails; two - English is an abysmal language which has the largest percentage of dyslexic people of any language, and it’s associated with the fact that the dictionary is a mix of words from many languages, and neither they adhere to some single rule of spelling, or nor to 5 of them…
everyone assumes I am talking about taking something chatgpt spews out and using it as it is whereas only the thing I said was to use it as an initial starting point i.e over coming the blank slate block. When everyone is so horrible in understanding what other people try to convey I assume you wouldn’t lose much if you used chatgpt as it is anyways lol.
Furthermore, lacking proficiency in any language and using a tool to “beautify” a paragraph in said language will generally fail to improve communication, because chatgpt is trying to infer and add information which just isnt there (details, connotations, phraseologisms). Will just add more garbage to the conversation, and most likely words and meanings that just arent yours.
It’s fine. Eventually when people start using this crap en masse the people on the other end will just be using LLMs to distill the bullshit down to 3 key points anyway.
That would be bizarre, lol
Let’s say one person writes 3 pages with some key points, then another extracts modified points due to added llm garbage then sends them again in 2 page essay to someone else and they again extract modified points. Original message was long gone and failure to communicate occurred but bots talk to each other so to say further producing even more garbage
In the end we are drowning in humongous pile of generated garbage and no one can effectively communicate anymore
The funny thing is this is mostly true without LLMs or other bots. People and institutions cant communicate because of leviathan amounts of legalese, say-literally-nothing-but-hide-it-in-a-mountain-of-bullshitese, barely-a-correlation-but-inflate-it-to-be-groundbreaking-ese, literally-lie-but-its-too-complicatedly-phrased-nobody-can-call-false-advertising-ese.
What about using an LLM to extract actual EULA key points?
New movie just dropped Violet evergarden: Gemini’s dream
I’ve been watching quite a lot of Olympics coverage on TV, but never seen any ads. Is there an official Olympics TV channel with these ads?
Idk, I mean I think this is more honest and practical LLM advertising than what we’ve seen before
I like to say AI is good at what I’m bad at. I’m bad at writing emails, putting my emotions out there (unless I’m sleep deprived up to the point I’m past self consciousness), and advocating for my work. LLMs do what takes me hours in a few seconds, even running locally on my modest hardware.
AI will not replace workers without significant qualitative advancements… It can sure as hell smooth the edges in my own life
I’d view it as an opportunity for AI to provide guidance like “how can I express this effectively”, rather than just an AI doing it instead of you in an “AI write this” way.
putting my emotions out there
You think AI is better than you at putting your emotions out there???
I get what they mean. It can help you articulate what you’re feeling. It can be very hard to find the right words a lot of the time.
If you’re using it as a template and then making it your own then what’s the harm?
It’s the equivalent of buying a card, not bothering writing anything on it and just signing your name before mailing it out. The entire point of a fan letter (in this case) is the personal touch, if you are just going to take a template and send it, you are basically sending spam.
I am 100% for this if it’s yet another busywork communication in the office; but personal stuff should remain personal.
This is the same reason people think giving cash as a valentine’s gift is unacceptable LOL
Yeah I agree if you send it without doing any kind of personalisation. I think LLM shine as a template or starting point for various things. From there it’s up to the user to actually make it theirs.
exactly… AI used as a template factory would be good use.
The problem here (with the commercial in question) is that they present it as Gemini being able to write a proper fan letter that is not even prompted by the fan (it’s the dad for some reason)… THAT is what makes it incredibly cringey
And to state the obvious; of course it would be helpful for anyone with a learning or speech disability, nobody in their right mind would complain about a “wheelchair doing all the work” for a person who cannot walk.
It can be very hard to find the right words a lot of the time.
That can be, in many cases, because you don’t read enough to have learned the proper words to express yourself. Maybe you’re even convinced that reading isn’t worth it.
If this is the case, you don’t have anything worth saying. Better stay silent.
Talking to a rubber duck or writing to a person who isn’t there is an effective way to process your own thoughts and emotions
Talking to a rubber duck that can rephrase your words and occasionally offer suggestions is basically what therapy is. It absolutely can help me process my emotions and put them into words, or encourage me to put myself out there
That’s the problem with how people look at AI. It’s not a replacement for anything, it’s a tool that can do things that only a human could do before now. It doesn’t need to be right all the time, because it’s not thinking or feeling for me. It’s a tool that improves my ability to think and feel
Talking to a rubber duck that can rephrase your words and occasionally offer suggestions is basically what therapy is
well I am pretty sure Psychologists and Psychiatrists out there would be too polite to laugh at this nonsense.
That’s the problem with how people look at AI.
Precisely, you are giving it a TON more credit than it deserves
It’s a tool that improves my ability to think and feel
At this point, I am kind of concerned for you. You should try real therapy and see the difference
Psychiatrists don’t generally do therapy, and therapists don’t give diagnoses or medication
Therapy is a bunch of techniques to get people talking, repeating their words back to them, and occasionally offering compensation methods or suggesting possible motivations of others. Telling you what to think or feel is unethical - therapy is about gently leading you to the realizations yourself. They can also provide accountability and advice, but they don’t diagnose or hand you the answer - people circle around their issues and struggle to see it, but they need to make the connections themselves
I don’t give AI too much credit - I give myself credit. I don’t lie to myself, and I don’t have trouble talking about what’s bothering me. I use AI as a tool - these kinds of conversations are a mirror I can use to better understand myself. I’m the one in control, but through an external agent. I guide the AI to guide myself
An AI is not a replacement for a therapist, but it can be an effective tool for self reflection
I think it can be, if you know how to use it
It literally cannot since it has zero insight to your feelings. You are just choosing pretty words you think sound good.
The choices you make have to be based on some kind of logic and inputs with corresponding outputs though, especially on a computer.
The choices you make have to be based on some kind of logic
Sure, the ones I make… the ones the “AI” makes are literally based on statistical correlation to choices millions of other people have made
My prompt to AI (i.e. write a letter saying how much I love Justin Bieber) is actually less personal input, and value, than just writing “you rock” on a piece of paper… no matter what AI spews.
This would be OK for busywork in the office. The complaint here is not that AI is an OK provider of templates, the issue is that it pretends an AI generated fan mail, prompted by the father of the fan (not even the fan themselves) is actually of MORE value than anything the daughter could have put together herself.
Yes, but this is also its own special kind of logic. It’s a statistical distribution.
You can define whatever statistical distribution you want and do whatever calculations you want with it.
The computer can take your inputs, do a bunch of stats calculations internally, then return a bunch of related outputs.
Yes, I know how it works in general.
The point remains that, someone else prompting AI to say “write a fan letter for my daughter” has close to zero chance to represent the daughter who is not even in the conversation.
Even in general terms, if I ask AI to write a letter for me, it will do so based 99.999999999999999% on whatever it was trained on, NOT me. I can then push more and more prompts to “personalize” it, but at that point you are basically dictating the letter and just letting AI do grammar and spelling
The future will be bots sending letters to bots and telling the few remaining humans left how to feel about them.
The old people saying we have lost our humanity will be absolutely right for once.
