Elektrine lite

← Feed

@ct_bergstrom@fediscience.org

Post #872201

2024-08-13 20:57 UTC

So you're thinking about getting #ChatGPT to write something for you... (Adapted from https://www.linkedin.com/posts/tyulkanov_a-simple-algorithm-to-decide-whether-to-use-activity-7021766139605078016-x8Q9)

Replies (16)

  • @gregwardo@c.im 2024-08-13 21:08

    @ct_bergstrom@fediscience.org aren't you throwing the baby out with the bathwater? I've used it to start drafts and found it useful but maybe I'm a fool.

    Open ##2259364

  • @drhoopoe@mastodon.sdf.org 2024-08-13 21:21

    @ct_bergstrom@fediscience.org Just in time for when I'm putting together slides for the first day of classes next week. Thanks!

    Open ##2259367

  • @MishaVanMollusq@sfba.social 2024-08-13 21:38

    @ct_bergstrom@fediscience.org I hadn’t gotten around to trimming my copy yet . Saved me 37 seconds

    Open ##2259368

  • @samhainnight@mstdn.social 2024-08-13 21:57

    @ct_bergstrom@fediscience.org You left out “Could you pay a ghost writer to do this ethically instead?” and “Are you OK with fucking over the environment because you’re lazy?”

    Open ##2259369

  • @ct_bergstrom@fediscience.org Excellent. The only other item I would add, while not spotting where to slide it into the flowchart: Do you care that you are wasting energy? #AI #Insanity

    Open ##2259373

  • @xs4me2@mastodon.social 2024-08-13 22:17

    @ct_bergstrom@fediscience.org 🎯

    Open ##2259374

  • @the_moep@mastodon.de 2024-08-13 22:27

    @ct_bergstrom@fediscience.org Taking full moral and legal responsibility for some random text is such a weird concept... like if it later turns out wrong just correct it? This is what we do with non-generated text already, why should there be a difference if one uses a glorified word-suggestion tool or not?

    Open ##2259375

  • @EVDHmn@ecoevo.social 2024-08-13 22:54

    @ct_bergstrom@fediscience.org I don’t touch most LLM . I do use perplexity, or bing to explore a thought matrix but not to find accuracy I’m selective about my foraging. Every thought or action=energy cost. I depend on scientists other researchers who collaborate with thoughts and ideas exploration . Santa Fe lectures are one of my main points currently of accuracy and learning lately. Plus other outliers from my network. Allen’s institute,Dr Sapolsky, CIP, Miller lab. And a few others. 🥰

    Open ##2259378

  • @mickymorse@mastodon.social 2024-08-14 00:43

    @ct_bergstrom@fediscience.org Asking a machine to think for you is not progress.

    Open ##2259379

  • @ct_bergstrom@fediscience.org does this apply to supervised generation of alt text? I don’t use it myself but have heard good arguments for it. I wonder if this is a wedge issue between disability advocates and ai pessimists.

    Open ##2259380

  • @hwyaden@toot.wales 2024-08-14 01:29

    @ct_bergstrom@fediscience.org would you normally use lorem ipsum after the first paragraph is my new test.

    Open ##2259461

  • @pelielios@mastodon.social 2024-08-14 02:52

    @ct_bergstrom@fediscience.org Don't forget that LLM-generated text can't be copyrighted so if this is a for-profit endeavor, add an extra scoop of dumb on top. Not sure how human-modified it has to be to skirt around that, if the law is even close to there yet. But some works (including a graphic novel IIRC) have already had copyrights yoinked for being AI-generated.

    Open ##2259462

  • @gordonwhite1@mas.to 2024-08-14 04:05

    @ct_bergstrom@fediscience.org Excellent chart of the logic chain for why it’s best to avoid AI writing tools. Critical thinking is aided by the tactile interactions with writing one’s research and creative ideas. Letting a machine do your thinking is irresponsible and lazy. Plus how often while writing have you discovered mistakes in your research. Bad combo for your own creativity, research accuracy and those that value your writing.

    Open ##2259463

  • @Life_is@no-pony.farm 2024-08-14 04:48

    @ct_bergstrom@fediscience.org @BlumeEvolution@sueden.social

    Open ##2259465

  • @Sirablopp@eldritch.cafe 2024-08-14 09:14

    @ct_bergstrom@fediscience.org To the people who think "But I found it quite helpful and how important is the last question on the right side anyway?" https://axbom.me/notice/Akn6YXjElXGBT0j0oi I'm not entirely sure you really grasp the full scope of moral responsibility for using it, not just in case of a fuck-up.

    Open ##2259466

  • @sabik@rants.au 2024-08-17 13:37

    @ct_bergstrom@fediscience.org @astraluma@tacobelllabs.net I feel like the left-most one should also be "then why"

    Open ##2259467