Give Us This Day Our Daily Thread Feb 1st - Feb 28th 2026

[wibbleflea]

Google Gemini, 11th February. It was doing quite well till then, giving me exactly the pointers I was looking for. Oh well.

 

[Dotta]

AI

293K views · 22K reactions | When the suitcase comes out. 🧳🦴🧸 | Cocodoodle

When the suitcase comes out. 🧳🦴🧸

www.facebook.com

791K views · 18K reactions | #cat #catlovers #cute #cutecat #petlovers #kitten #facebookreel #facebookviral | Catnap Vibes

#cat #catlovers #cute #cutecat #petlovers #kitten #facebookreel #facebookviral

www.facebook.com

 

[iamshadow21]

Never use AI. It's a nonsense engine, and it's dangerous. It's telling people things that aren't just incorrect, but that are coercing them into acts of paranoia, self harm, suicide, and murder. It's generating CSAM and other sexual assault and harm material. It's accusing perfectly innocent people of crimes; a guy in Europe, it's telling everyone he murdered his children, something that never happened.

And people are relying on it like it's the word of God, rather than actually researching for themselves. It's giving frightening advice on mental health and medication and it's coaching children how to do dangerous things like light fires and use weapons.

It needs to go away until it's regulated and there are consequences for it giving out dangerous advice.

 

[MrsWatson]

My two experiences with AI.

1. I fancy myself a writer and belong to local writing groups. We generally receive a prompt to use for the exercise. I was having a difficult time with a prompt, so I stuck it into AI and asked AI to write a flash fiction piece, using the prompt. I thought it might kick start my imagination.

The result was absolute nonsense. Full of cliches, no plot, no character development. Just three paragraphs of sentences that didn't belong together and didn't tell any sort of story. It was laughably horrible.

2. My husband works for the largest cancer hospital trust here in the UK. He's a tech guy. He has a co-worker who is AI-crazy and wants to put it into all sort of uses at the hospital. One of those was the 24 hour care line. People with cancer or their loved ones can call at anytime for support, questions, do I need to see the doctor, are these side effects normal, etc. It was shot down, with my husband leading the pack. If you were suffering cancer, would you want AI to decide whether or not you talk to a real person? If you are feeling desperate because your partner of unmpteen years is suffering, do you want to talk to AI?

AI needs to do things like clean my house and order my groceries and do the back garden, not try to be a human.

 

[Chainspell]

A couple years ago when AI first started getting prevalent, I'd have students submit almost the exact same essay, with only minor composition/construction differences. The thesis would be the same, the evidence presented would be the same, the sources would be the same (even hallucinated ones!). Most of the time, they were just putting in the essay assignment instructions and taking whatever AI would spit out for them.

Students have gotten better about hiding AI use, but I did have one last year that submitted an essay about Revolutionary War hero Harry Truman. He admitted to using AI when I confronted him immediately, but I really wish I knew how that happened with the AI itself.

 

[Dotta]

14 February

Happy Valentine's Day!

"Dov'e L'Amore
There is no other,
No other love can take your place
or match the beauty of your face
I'll keep on singing til the day
I carry you away
With my love song"

Flamenco, Spanish guitar and passionate hearts
Not bad girls, eh?
My Slavic soul loves these rhythms

 

[tricia]

GIVE US THIS DAY OUR DAILY THREAD SATURDAY VALENTINE'S DAY FEB. 14TH 2026
​

Loving the AI stories. Keep them coming. I use chatgpt.

 

[HoveringWombats]

I spent most of my career in the tech industry, just recently retiring from that. I'm glad to be getting out now. AI is such an interesting phenomenon because IMO it is simultaneously the most overhyped innovation I have seen in my lifetime, and it also has some of the most potential. I think it's both true that AI is going to do incredible things and usher in big changes, and that the impact and capabilities are overblown.

I think one problem is that what we're currently calling "AI" (as an aside, "artificial intelligence" has been around for decades. I worked on "AI" stuff in the 90s that had no relation to what we're using today) is that it fundamentally breaks human brains in terms of how we expect technology to work. In many ways, social media broke human brains because we aren't designed or evolved to have that level of global connection and constant stimulation. And AI breaks brains because it can so convincing act like it is "thinking" or a living thing, while it's still just data and math under the hood.

So we see things like people expecting that having AI modify a photo will be based on some kind of true intelligence or cognition like a human would use, but really it's just probabilities and calculations. I don't think we're equipped to reason about the output of these computer programs that believably act like people but are still programs at the end of the day. So I really worry about the impacts to the world. But I also think that the tools can be very useful. A confounding set of contradictions....

 

[GatorsNorth]

I spent most of my career in the tech industry, just recently retiring from that. I'm glad to be getting out now. AI is such an interesting phenomenon because IMO it is simultaneously the most overhyped innovation I have seen in my lifetime, and it also has some of the most potential. I think it's both true that AI is going to do incredible things and usher in big changes, and that the impact and capabilities are overblown.

I think one problem is that what we're currently calling "AI" (as an aside, "artificial intelligence" has been around for decades. I worked on "AI" stuff in the 90s that had no relation to what we're using today) is that it fundamentally breaks human brains in terms of how we expect technology to work. In many ways, social media broke human brains because we aren't designed or evolved to have that level of global connection and constant stimulation. And AI breaks brains because it can so convincing act like it is "thinking" or a living thing, while it's still just data and math under the hood.

So we see things like people expecting that having AI modify a photo will be based on some kind of true intelligence or cognition like a human would use, but really it's just probabilities and calculations. I don't think we're equipped to reason about the output of these computer programs that believably act like people but are still programs at the end of the day. So I really worry about the impacts to the world. But I also think that the tools can be very useful. A confounding set of contradictions....

I agree. In the 90s I worked with a woman whose husband was working in AI, and it was nothing like we see now.

Sometimes it’s a good thing and sometimes it’s a bad thing. If I get an answer from AI that I think is sketchy, I check it out myself.

 

[HoveringWombats]

Oh, and in terms of what I use or have used AI for..

I've definitely used it for programming tasks. It is quite useful for that, but I'm not completely buying the idea that it is ready to replace programmers, because in its current state you still really need to be able to give it pretty specific direction and evaluate what it's doing, because it can and will go off the rails pretty quickly with complex problems. It's true that you can use AI to build entire apps without writing any code at all, but I think people will find that if you continue to use, maintain, and enhance those apps over a few years without ever touching the code, it's going to eventually be a mess. The primary debate is whether the continued rate of improvement in AI coding capabilities will reach a point where that's not a problem soon. I tend to think that we'll hit the point of diminishing returns before that, but I can't claim to truly know.

I definitely worry about what will happen if companies stop hiring and mentoring junior programmers, because you need a pipeline of junior developers in order to eventually have senior developers. Once all the senior people retire, if there are no more junior people to supervise the AI tools, we may be in a mess. I guess the bet is that by that point, the AI tools will be so good that you don't need any human to understand the code, but color me skeptical about that.

Beyond programming stuff, I don't let AI write for me, but sometimes I use it as a thought partner for writing. Occasionally if I'm stuck on how to start something, I'll tell an AI tool what I'm trying to do, and its ideas can give me a jump start. At times, when writing a complex document, I will give AI a stream of consciousness brain dump of what I want to get across, and have it suggest how to structure it. And occasionally I will feed in something I have written, and ask for feedback or a modified draft. But in none of these cases, do I ever use the AI output directly. I just use it to help improve my own writing.

Sometimes I will use an AI tool for data analysis, like I fed in a bunch of credit card transactions and asked it to look at trends by grouping together similar categories (i.e. the different credit cards use somewhat different categories, and I wanted to see spending summed up by category with them unified across cards). But you have to be really cautious with stuff like this because even with things that might seem as straightforward as math, it can make very stupid mistakes. I think it's ok for directional data, or things where you can easily verify the results, but I would not make important decisions based just on AI output for something like this.

AI can be really useful to help me figure out what to search for in order to do a more traditional search. I can enter a rambling description of what I am trying to find or figure out, and even if AI does not give me the right answer, it usually gives me enough direction or terminology to then go to Google and do a proper search for the right thing. (Sometimes even this fails in comical ways, but usually it works.)

Finally, I occasionally use AI for research in what I'd call a "bar trivia" context. Like I'll ask it to explain stuff that I'm just curious about, when it isn't really that important if the answer is perfect. The same as if I was chatting with friends at a bar about it over a few beers. If something seems off, or I get more curious based on what I read, I will go find actual sources and confirm facts. But for low-stakes stuff, AI answers are often good enough

I have two big concerns with these technologies.

First, as I mentioned above, I don't think that human brains are evolved to evaluate the veracity of AI output, and it's going to keep getting worse. We see that in threads here, where people post AI things like they are meaningful, and we see it on social media where AI-generated photos and videos go viral with people believing they're real. This is really dangerous and I don't think that simple education or regulation is going to solve it. I don't know what the answer is.

Second, we have to remember that AI works by being "trained" on content created by humans. All AI writing is based on writing that humans did. AI photos are made by models trained from photos humans took. AI code is written by models trained on code that humans wrote. If we get to a place where a large proportion of the new content in the world is AI-generated, what will the next AI models train on? It becomes a big self-perpetuating loop, and no genuinely new ideas or creativity can come out of that.

Anyway, it's both a brave and scary new world. It'll be interesting to watch.

 

[Dotta]

I agree. In the 90s I worked with a woman whose husband was working in AI, and it was nothing like we see now.

Sometimes it’s a good thing and sometimes it’s a bad thing. If I get an answer from AI that I think is sketchy, I check it out myself.

AI is still a baby.
What we see are short-term difficulties or glitches that occur during the early stages of a new project.
In other words "teething problems"

Besides,
let's be honest - AI is a mirror of ourselves.

As some say
it acts as a mirror by reflecting the data, biases, and values used to train it,
offering a composite image of humanity.

One thing is certain IMO.
There is no turning back.
AI is the future.

And as I wrote earlier in one of these threads "Give Us This Day" where we discussed this very subject
(but I don't remember the month it was)
those who turn back on the technological progress
will soon be left behind.

No one can stop technological progress.

We have intelligence to use AI wisely.
It can be our ally.

AI has never disappointed me.
It helps me with my job as a teacher.
And when giving info,
it always provides links which I can check.

JMO

 

[thefragile7393]

I used chatgpt to help tease out my perimenopausal symptoms. I thought I was going crazy. One doctor didn't take me seriously and simply switched some meds around. Turns out...I was having many many symptoms that have since simmered with the right type of hormonal help...and chatgpt helped me with that. I also use it to do layouts for my powerpoint presentations for school. I do the research and the legwork but it puts everything on slides for me, which seems to help my disorganized brain so much better. It has some good uses.

 

[warp3dwing]

I only use LLMs for programming work, and even then, only under close watch.

I’ve tried in the past to use it for work writing (editing mostly), but the context window is so small that it struggles to deal with larger, more complicated themes. It not only sucked whatever fun I could extract from the work, I don’t think it saved time at the end of the day.

And the more tired I got, the more likely I was to just accept its convincingly argued suggestions. Of course, I later realized it was rubbish.

The obsequiousness of LLMs become dangerous when ansking about subjects I know nothing about. When I asked it about aviation topics, I could immediately see where it was incorrect because of expertise in that field. But in other fields, I don’t have the knowledge to refute what it’s saying without research. And at that point, why not just write it myself?

I’m at the point where I only really use it now for coding, and since I’m not writing for work, I prefer to just write my own slop, thank you very much.

What confuses me is the number of posts (here and elsewhere) where the poster pastes an AI response or summary. We all have access to AI. If I wanted a summary, I’d get it myself. I want to know YOUR opinion, Mr or Mrs poster!

 

[justwannahelp]

I wish I was able to respond to your request but I've never felt the need to use it and I don't trust it.

Hopefully others will have interesting anecdotes on the topic.

I don’t use it. I just google. I’ve heard it’s bad for the ocean. No idea if true. I figure IF AI super computers really harm the ocean, though, google works just fine for me and “Hey Siri”

I’m kind of a can’t teach an old dog new tricks kind of person. Took me forever to trust my backup camera in my car

I do see where it could be very useful in certain work settings, so I’m not judging anyone who uses it. It’s just not useful for me personally. Seems really useful for general research and medical research