• 0 Posts
  • 56 Comments
Joined 3 months ago
cake
Cake day: October 1st, 2024

help-circle


  • Sometimes I wonder if it’s a complete waste of time to think through a post that I’m writing, if only a couple people are going to read it. But then I figure: a) doing so is its own reward: practice putting sentences together, keeping the mind sharp; b) some texts/ideas can be seminal, just as a music band may have very few fans but each of those fans goes on to create their own band; c) contemporary scholars study texts and articles (including ephemera such as handbills) from past decades, so it’s likely that future scholars will trawl and study social media posts from our era, using techniques we can barely imagine. Plus, it’s fun!



  • If Lemmy had weekly awards, you would win one for this post. The bland, LLM-inspired structure creates a feeling of rising dread until the very end when one is left with the horror of realizing this human (if they can still be called human?) has spent way too long talking only to AIs.

    What’s more, the text is not a story or essay submitted as a post; the text only really works AS a post, with its references to Lemmy, Aspect, and SocialAI and contextualized among a stream of posts. The fact that it’s in [email protected] provides ironic distance, but not so much to prevent it from being read unironically for at least the first couple paragraphs. I don’t know what Aspect and SocialAI are like, but the differences between them and Lemmy that are pointed out in the text creates a picture of a platform that problematizes modern identity and the individual’s role in a society mediated by social media (ha) and AI bots. I bet someone could write a half-decent critical theory research paper expounding on your post. Well done.



  • It looks like your “edit” is right (version of lemmy and instance configuration are an issue), with the additional complication that sometimes where it is hosted makes a difference (sometimes the youtube summary is in a different language and/or won’t grab it because of regional restrictions) and sometimes the reader app/web interface grabs their own thumbnails and/or may not display thumbnails and/or may even try to just embed the video. There was a post on techsupport a while back about this: https://lemmy.world/post/20180043

    I post to [email protected] from slrpnk.net and to get a thumbnail, I do the following:

    • get a thumbnail image file. either grab it from wikipedia, imdb, or by a screencap that I trim. put that file somewhere on my computer.
    • use the web interface with default client. It has a “thumbnail URL” field (which I didn’t see back when I had a lemmy.world account - probably version differences) but I can’t just upload the image to that field, so I have to…
    • go to the post “body” input field and click on the “upload image” button to upload the thumbnail image file from my computer. This uploads the image to the server and adds a line of code to the body input field that looks something like this: ![](https://slrpnk.net/pictrs/image/a-long-UUID.png)
    • copy the part from “https” to “png” and put it in the “thumbnail URL” field
    • usually I then delete that code from the post “body” input field
    • edit: of course, I still have to add the youtube URL in the “URL” field…

    I think this adds thumbnails, but of course I can never be sure because someone can always use a client or configuration I’m unaware of. That’s the price of freedom from a single corporate server and a limited choice of viewer clients.
    Ideally this whole process would be easier or automatic. I imagine the situation/process will change as time gones on.