Kelson Reviews Stuff - Page 18

It’s hard to pin down Picard Season 2.

The opening episode should have been an opening 5 minutes. From there it turns into something that looks like a Mirror Universe story but isn’t. By episode 3 it’s a fix-the-timeline story set in the present day (more or less – the Europa mission is way too soon for real-world 2024, but Artemis II is targeted for roughly the same window!) interwoven with an intriguing story about mental illness and trauma.

The next few episodes are actually pretty good! Interesting concepts, social commentary, some mystery, character exploration, some nice references to Star Trek IV for the nostalgia, Brent Spiner as yet another ancestor of Dr. Soong.

But once they figure out what’s changed and what they need to fix, the story just turns into a messy kitchen-sink stew. Things make sense individually but the connections don’t, or vice versa. Overkill is the rule, as character developments are sidelined or discarded, thematic resolutions are undermined, and the timey-wimey ball unravels into a timey-wimey mess.

A fun, accessible introduction to how machine learning works…and how it sometimes doesn’t!

Still relevant despite recent advances in AI-generated imagery and text, because the new systems still work on the same principles as the ones that were around three years ago. They just have a lot more data and processing power. This also means they have the same limitations and blind spots. What was it trained on? How was it trained? (This is the most obvious way human bias can leak into an AI model.) How well is the goal specified? And of course, did the AI actually latch onto relevant details, or did it notice that all the training pictures labeled sheep had green fields and blue skies, and completely ignore the actual sheep?

These are things to keep in mind as we enter the landscape of generative AI tools like ChatGPT: You can train an LLM to write a book review, and it’ll give you a great piece of text that reads like a book review – but it’s not going to have actually evaluated the book. For that, you’d have to train another AI to categorize books as good, bad, interesting, dull, and so on. But even that can only be as good as its training data. (I don’t remember whether the classic phrase “garbage in, garbage out” is used anywhere in the book, but it still applies today!)

The author has a blog/newsletter, AI Weirdness, where she pushes AI over the edge to sometimes hilarious results.

I really wanted to like The Book of Boba Fett. I like Fett well enough and I like Fennec Shand…but I just couldn’t get into the story of him trying – clumsily – to become a crimelord.

The flashbacks showing his escape from the Sarlaac, his captivity and later acceptance into a Tusken tribe were fascinating, and I would cheerfully have watched more of that!

I think the problem with the show is that they put everything that audiences wanted out of a Boba Fett show into The Mandalorian. Cool armor and weapons, has his own code, always seen with the mask on, flies around the galaxy and gets into fights… They had to do something different for a Boba Fett show…but the approach they took just didn’t work for me.

(Spoiler: Also, I was so mad at Luke and Ahsoka in their guest spot. The two people in the galaxy best suited to understand that denying Anakin familial attachment is what enabled the dark side to get its hooks in him, and there they are repeating the old order’s mistake with their first student.)

Hilarious self-parody of TNG-era Star Trek. Funny on its own, but even better if you know the shows it’s riffing on. Great as a diversion or as a palette cleanser after watching something more serious (like, say, Picard).

Toward the end of the first season they start mixing in episodes with Big Changes™, Important Events, Mysteries and continuity. It’s not bad when it does those episodes. But I think it’s better when it plays to its strengths: the absurdity of the situations and the way the characters play off each other.

A very lightweight social networking server, with a very clean web interface for viewing public posts. Compatible with Mastodon apps and interacts well with Mastodon and other major ActivityPub platforms on the Fediverse.

It’s not meant for handling zillions of users. GoToSocial is more for setting up a server for your group or family or friends (or even just yourself) on a spare Raspberry Pi or cheap VPS host. I’ve got it running quite well on a 1GB Linode.

You can run it as a semi-private server, so your group can interact locally without the posts leaking, and still share public posts with followers on Mastodon etc.

(It is server-only, though. You can manage settings on its website, and view public posts, but all the interaction has to be done through an app like Elk, Ivory or Tusky.)

Right now it’s still in alpha.

• Not all features are implemented yet.
• It has trouble with some other Fediverse platforms (but that’s improving as bugs get sorted out!)

I wouldn’t recommend it as a complete solution yet, because, well, it’s not complete.

But it already handles the basics: Sharing pictures and words, and interacting with people on the same and other servers.