Post AvDloeQTRi94dkftnE by maalmike@mastodon.social
(DIR) More posts by maalmike@mastodon.social
(DIR) Post #AvDlZaDOosijEjNGk4 by stux@mstdn.social
2025-06-17T12:16:44Z
1 likes, 0 repeats
How often do you use #AI tools like: ChatGPT, Claude, Gemini, DeepSeek or Grok?
(DIR) Post #AvDloeQTRi94dkftnE by maalmike@mastodon.social
2025-06-17T12:19:26Z
0 likes, 0 repeats
@stux But only my local modelsI find it useful with technical questions, or translations.
(DIR) Post #AvDluxTVXFthSO6LJ2 by cjd@pkteerium.xyz
2025-06-17T12:20:39.738253Z
0 likes, 0 repeats
They don't reason all that well, but they're fantastic for search, translation, and making a bullet-list into a presentable document.
(DIR) Post #AvDlwhKiWN2NyOoqbA by waldschnecke@chaos.social
2025-06-17T12:20:54Z
0 likes, 0 repeats
@stux I look at my printer in it’s isolated network, with a hammer at the ready, to smash it to bits if it does anything abnormal. As an IT guy working in the security sector - I don’t trust technology one bit.
(DIR) Post #AvDly834gAo77SFVxI by spanghero@ioc.exchange
2025-06-17T12:21:07Z
0 likes, 0 repeats
@stux Never knowingly.
(DIR) Post #AvDmtq85SanmL872WG by ranx@mastodon.social
2025-06-17T12:31:32Z
0 likes, 0 repeats
@stux I used ChatGPT 3 times since it was released, but I still prefer to use my brain and an old style search engine
(DIR) Post #AvDnaT88qWGnSfSu2q by HurricaneWeasel@mastodon.world
2025-06-17T12:39:19Z
0 likes, 0 repeats
@stux I check my texts for misspelled words or grammatical errors all the time, and it also helps with coding.
(DIR) Post #AvDndRTRSedbC0qVn6 by redstarfish@freesoftwareextremist.com
2025-06-17T12:39:56.198386Z
1 likes, 0 repeats
@stux Calling these LLMs, "AI" confuses people who don't know how these tools work. General public think they are actually thinking, consequently some people tell them their feelings and secrets. People also believe in what these models write: "use glue in pizza", cite non existent criminal cases, etc. I hope you'll not fall for buzzword and boost these companies.
(DIR) Post #AvDnukOENx1kmZf08e by stux@mstdn.social
2025-06-17T12:42:55Z
1 likes, 0 repeats
@redstarfish I picked them because of that unfort 😉 Since the mainstream media rlly fkced up the meaning of the word, we know better luckily but a lot of people don'tSo for this poll i went with that for a reason
(DIR) Post #AvDo7iaQOwVIxDY5lg by snarkweek@mstdn.social
2025-06-17T12:45:14Z
0 likes, 0 repeats
@stux Today I asked DDG a question and they auto seem to ask the LLM now if it is grammatically or conversationally like a question. They said they use a combination of the LLMs you named, but I'm not sure how that works. I asked it to identify the main difference between a bong a water pipe. You know, because people said you have to ask it something you know a lot about, if you want to properly test it. It didn't know the answer.
(DIR) Post #AvDoQn1a7oYCDjLTCS by trebach@functional.cafe
2025-06-17T12:48:39Z
0 likes, 0 repeats
@stux I only willingly use it to demonstrate how AI is stupid. Who the hell knows how often it's been used involuntarily because it's built into products without the option to turn it off
(DIR) Post #AvDp5NNCOuo0XMoGCe by crankylinuxuser@infosec.exchange
2025-06-17T12:56:06Z
0 likes, 0 repeats
@stux I'll use it to do what amounts to an interactive web search, and to make a thought-web of a subject I don't have a thorough understanding of.I treat it like Wikipedia with a possible contentious streak. Ask, and verify.Because of my hybrid solution, I've been able to determine narrow terminology that I wasn't aware of. And was able to properly research with exact terms, rather than converse with something that potentially hallucinates more and more.And... They're also good about getting behind paywalls 😂
(DIR) Post #AvDpGDz1eWM0bHfxSa by jmaris@eupolicy.social
2025-06-17T12:58:00Z
0 likes, 0 repeats
@stux I occasionally use them for extracting or transforming unstructured data, but not on a regular basis
(DIR) Post #AvDqlqIBN2JSLRHD7I by crashlogger@mastodon.eus
2025-06-17T13:14:59Z
0 likes, 0 repeats
@stux I tried a few of them a couple of times, they failed to produce helpful informationI asked for the documentation they used as a source, it never mentioned the use-case the LLM employed.Also -I know the image generation isn't directly tied to the LLM AI- I told one to draw a train once and it refused to draw anything other than a cartoon steam locomotive...? Not Thomas! Just a generic cartoon-style steam locomotive!
(DIR) Post #AvDr5QFftje1NDRLkm by Lazarou@mastodon.social
2025-06-17T13:18:28Z
0 likes, 0 repeats
@stux I'd rather drink paint than use this shit they call 'AI' I look at the people pushing it and I do not trust them or like them. They creep me out frankly.
(DIR) Post #AvDr9shY45w2labVWC by BmeBenji@mstdn.social
2025-06-17T13:19:21Z
0 likes, 0 repeats
@stux My manager tells me that it’s literally one of my performance metrics. Needless to say someone in the company desperately wants to justify that massive fucking check they wrote to Microsoft
(DIR) Post #AvDsD4uPEbuDRxhqjI by jaystephens@mastodon.social
2025-06-17T13:30:58Z
0 likes, 0 repeats
@stux "If I'm explicitly told to at work and at no other times."
(DIR) Post #AvDt4XWxY9rjZDEQvA by kate7@mstdn.social
2025-06-17T13:40:46Z
0 likes, 0 repeats
@stux never - too many concerns with stolen training datasets, environmental impact, bias, and the potential for devolving cognitive processes
(DIR) Post #AvDuDhxmEOqmU1i9GC by markdevries@mstdn.social
2025-06-17T13:53:39Z
0 likes, 0 repeats
@stux I think pretty biased crowd here. Maybe add 2 extra polls:- how often does your partner use …- How often do your kids use …And see if there are differences
(DIR) Post #AvDvP1bHf20mWi9NLM by davidhanzlik@mastodon.online
2025-06-17T14:06:53Z
0 likes, 0 repeats
@stux every time AI is mentioned I think of small models now because of this lady interviewed on PBS. She talks about AI being used to read CT scans for cancer, be used to place windmills, etc. Apparently the models are much smaller and require less data and energy than a ChatGPT. It would be so nice to know how to create a personal AI for a well defined task from a small dataset.https://www.pbs.org/video/the-world-of-ai-1749405092/
(DIR) Post #AvDvZ2BuJrhTFq3NVQ by WaitingOnABus@urbanists.social
2025-06-17T14:08:36Z
0 likes, 0 repeats
@stux I will add the caveat “willingly.” While quickly googling stuff on my work computer, I always forget that Gemini is going to pop up and I have to scroll past it.
(DIR) Post #AvDwKWAcftwhqO5vEW by Zergling_man@sacred.harpy.faith
2025-06-17T14:17:03.701472Z
0 likes, 0 repeats
@stux nigger faggot kike
(DIR) Post #AvDxmgFAup4F9HDUPo by arisummerland@beige.party
2025-06-17T14:33:30Z
0 likes, 0 repeats
@stux Never willingly and never on purpose. I recently had to turn AI suggestions off again in my browser. It turned them on again without my consent. 🤬
(DIR) Post #AvDxwzQcRUu7XoU9T6 by Cerrita@mastodon.social
2025-06-17T14:35:20Z
0 likes, 0 repeats
@stux I avoid them if I can as I value learning the why when working on issues, but admittedly I have used it once a couple of days ago to solve a specific problem I was having with terraform syntax related to foreach loops. I grumbled the whole time but I'll be damned if it didn't actually help. Fwiw though this is specifically AmazonQ. I haven't had nearly as much luck with other LLMs.
(DIR) Post #AvDy43YZUZW8NJZgfo by salva_pl@mas.to
2025-06-17T14:36:43Z
0 likes, 0 repeats
@stux never. I block all them.
(DIR) Post #AvDyPiIjCTpap5UUvw by CosmicCactus@mastodon.social
2025-06-17T14:40:23Z
0 likes, 0 repeats
@stux I use them often but not for serious stuff. They have a nice sense of humor sometimes: https://chatgpt.com/share/68517dca-189c-800a-a514-d81d70eeb8ed
(DIR) Post #AvDyRKOt28AxH98pPs by simone_z@mastodon.social
2025-06-17T14:40:48Z
0 likes, 0 repeats
@stux #MistralAI (#🇪🇺) with their #lechat is missing from your poll.
(DIR) Post #AvDyRg8KEO04g9Yuem by markku@fosstodon.org
2025-06-17T14:40:58Z
0 likes, 0 repeats
@stux and when I do, I am usually dossapointed by misinformation.
(DIR) Post #AvDyZGGOIZiYlGwzTs by lackthereof@beige.party
2025-06-17T14:42:20Z
0 likes, 0 repeats
@stux Has been made mandatory by my employer.
(DIR) Post #AvDyvOcptCtF4C0Hwm by attans@mastodon.nu
2025-06-17T14:46:20Z
0 likes, 0 repeats
@stux When this hype first started, I asked one such service (can't remember the name) to give me a list of ten gay celebrities. Then I went to Wikipedia. It got eight of them right. It started arguing with me when I tried to correct it about the other two.
(DIR) Post #AvDzHWhYXjssU8lQmm by Sk4tz@mastodon.social
2025-06-17T14:50:15Z
0 likes, 0 repeats
@stux as Therapy
(DIR) Post #AvDzwIxEYkWBipXeCG by TYB@mstdn.social
2025-06-17T14:57:43Z
0 likes, 0 repeats
@stux Rarely
(DIR) Post #AvDzxQnkryWqLhe48m by malte@radikal.social
2025-06-17T14:57:50Z
0 likes, 0 repeats
@stux I'm closest to never. Not literally obviously. I've tried it and realistically will probably test something or have a laugh once a year. So sometimes is most correct for me.
(DIR) Post #AvE1FzRotVvMy8nZvk by Sara_rey@eupolicy.social
2025-06-17T15:12:24Z
0 likes, 0 repeats
@stux Biased crowd here. According to Sam Altman, type of use differs per age group:- above 40: like a search engine - between 30-40: for all kinds of questions, including personal things like relationship advice - under 30: more like an operating system Doesn’t say % of population though. I use it for research, usually Perplexity because it mentions the source used, so I can click and go on from there.I used Notebook LLM too to upload a list of sources and ask questions.
(DIR) Post #AvE61NEEh8uu9RIKNU by grynnan@mastodon.social
2025-06-17T16:05:35Z
0 likes, 0 repeats
@stux [Never] use open source local AI.(My 4080 is running hot! =P )
(DIR) Post #AvE7Of9oIpPiN1tSdc by nazokiyoubinbou@urusai.social
2025-06-17T16:21:18Z
0 likes, 0 repeats
@stux I wasn't sure which one to vote on on this one. It's between "never" for literally what you're asking and I think maybe most of what you mean, but "sometimes" if you consider that I do run a local LLM on my own computer from time to time for various things. (These use less power and have absolute privacy, but the catch of course is that it's slower and I have to use smaller models that have fewer parameters in their training -- effectively "know less" for lack of a better way to put it, though obviously that is wrong too.)
(DIR) Post #AvE9Ddd3qOSwLzRid6 by em_and_future_cats@mastodon.social
2025-06-17T16:41:42Z
0 likes, 0 repeats
@stux I don’t use it on purpose, it is always the DAMN first thing on the search page! I am being very honest when I say that I use it “sometimes” because it is hard not to when it’s always there 😒And yes, sometimes it has an overview that is credible but I still go to Wikipedia or other credible sites (like a university or BBC etc etc) to double check the results. So I don’t WANT to use it, and I tried the “fuck” trick but it’s not easy for simple daily searching 😿
(DIR) Post #AvEARliafJQjRyKmY4 by harryomvlee@mastodon.nl
2025-06-17T16:55:26Z
0 likes, 0 repeats
@stux Deepl, bijna dagelijks.Maar geen van deze vijf.Nee, nooit.
(DIR) Post #AvEBKlpZwpBNj7pVrs by shammers@mastodon.coffee
2025-06-17T17:05:23Z
0 likes, 0 repeats
@stux Never. The handful of times I did use it when it was brand-spanking new, it got enough basic stuff wrong that I immediately stopped trusting it. Never mind all the other issues I have with it.
(DIR) Post #AvEJdsX43VmNe7nBTc by rebootl@techhub.social
2025-06-17T18:38:21Z
0 likes, 0 repeats
@stux some anti-bias here, i use them often/daily, i like duck.ai (from duckduckgo), it has privacy protection (anonymisation, chats stored locally, requests not used for training.sometimes results are disappointing, but often it helps getting unstuck. also grammar, spelling checking, translations etc..
(DIR) Post #AvEKeODEMyxvr2K4ga by TheRealKat@mstdn.social
2025-06-17T18:49:43Z
0 likes, 0 repeats
@stux #AIsucks #AIkillsHumanity
(DIR) Post #AvEezVD1Y8KsicZSDY by BrendanThompson@mastodon.social
2025-06-17T22:37:40Z
0 likes, 0 repeats
@stux my use of them is weird though I mostly just ask esoteric questions
(DIR) Post #AvEfP3q7vkndMb3356 by kevin@dice.camp
2025-06-17T22:42:15Z
0 likes, 0 repeats
@stux I voted "never" but might test them every few months. It always sucks, I don't see the appeal.
(DIR) Post #AvSLMIpl5Ol7V7XqIC by foolishowl@social.coop
2025-06-24T13:03:00Z
0 likes, 0 repeats
@stux I have to wonder how this would compare to people off the Fediverse. I keep hearing that we're extraordinary in our opposition to it.
(DIR) Post #AvSYT9jwas7ijDQJaS by bronakins@mastodon.world
2025-06-24T15:30:34Z
0 likes, 0 repeats
@stux Let me be more specific: I never KNOWINGLY use #AI #LLM etc.I don’t need to—I have vocabulary, knowledge & memory.BUT I’m not very technologically savvy, so who knows. 🤷🏼♀️For example I’d swear recent iOS autocomplete/correct is deliberately inserting typos/errors for me to correct so it can “learn” from me. (Or maybe I’m paranoid or just that bad at typing.)