please vote anonymously and boost this poll for transparency and participation:

[falcennial@mastodon.social]

please vote anonymously and boost this poll for transparency and participation:

How frequently do you chat with LLM?

[falcennial@mastodon.social]

RE: https://social.tchncs.de/@jesuisatire/117181909837538051

this is a 3 monthly follow up, on request.

by participating you help us all to see how things have changed for fediverse users' LLM usage over time, which is very cool of you. thank you very much to the seven thousand, seven hundred and ninety-six community representatives who participated in the previous one, and special mention to the one hundred and seven people who share data exercises around: you shine a light on the world by including your friends.

https://mastodon.social/@jesuisatire@social.tchncs.de/117181909884842663

[daniele_hesse@mstdn.party]

@falcennial My wife is on ChatGPT a lot.

[zita@zitagravelbiker.eu]

@falcennial LMM ‍️ never heard of it ‍️

[codebeard@mastodonapp.uk]

@falcennial the frame of the question makes it hard to answer ‘chat with’ as in ‘conversation’ (not at all). ‘Use as a tool’ = few times a month...maybe more… but it’s not chat. So I answered that.

[bigpawedbear@masto.nu]

@falcennial what do you mean by chatting? do you mean all use cases? or just personal use cases? excluding formal use cases, research cases etc?

[adritheonly@mastodon.social]

@falcennial Question to the 25%. Is it worth burning the planet down for a few bytes of misinformation?

[flominator@genealysis.social]

@falcennial
Chat as in "use" or as in "having an actual conversation"?

[jbqueru@floss.social]

@falcennial Every few months, each time they claim "this time it's actually intelligent", to verify that it's not.

[goblinquester@dice.camp]

@falcennial What do you mean by chat? Just any propmt? Or prompt it to generate something that looks like a social conversation?
Kinda bit difference between those two promts versions.

[kittenkoder@mastodon.social]

@falcennial Delete all LLMs/ai. People who use them don't think.

[amy0223@mindly.social]

@falcennial @FreakyFwoof I’m interpreting “chat” to mean voice to voice conversation; something I never do. If this is not what is meant by “chat” the wording of this post may need changing. If you really mean chat, then awesome! My vote stands.

[rootsandcalluses@norden.social]

@falcennial Voluntarily? Or does every customer call and chat count?

[rwwh@mastodon.social]

@falcennial note: I only use locally run models.

[ann_in_a@livellosegreto.it]

@falcennial I would not do it at all, but my job has a new travel policy that requires us to book work-related trips via a chatbot. Only if the chatbot doesn't return suitable options (which is quite frequently tbh), I can book myself and claim reimbursement. But I have to chat with the bot first.

Additionally, there's the odd occasion in which I need to contact some customer service, and the first/only option is a chatbot.
This adds up to a few times per month

[adam_wysokinski@fediscience.org]

@falcennial I don't chat with LLMs, I chat with people. However, I frequently (each day) use LLMs for researching information, OCRing and translating documents, proofing my formal writing for language, style and grammar, checking code for hard to trace errors, explaining topics out of my fields, comparing products before purchasing and much more. I run some open, local models, but also use others.

[selvari21106@mastodon.social]

@falcennial not so far as I'm aware, though I can't fully rule out the possibility that I've spoken to one masquerading as a human and not realised

[shadsterling@mastodon.social]

@falcennial I voted “not at all”, but I am forced to use a customer service chatbot a few times a month

[shadowwwind@mastodon.social]

@falcennial when someone sends me a schedule in plain text, I use an LLM to turn it into an ics file to import into my calendar. I think that sums up my use in the past two months.

[rnichols1@mastodon.social]

@falcennial I suspect you'd get better results if you asked how often people use, rather than chat with, LLMs - chat is too vague