/reader

I am usually pretty good about remembering birthdays since I live in my calendar. And calendar aside I cannot believe he did not let it slip for the entire day. I don't think I'd be able to handle the absurdity of it but maybe I'm just a diva.

Anyway, I felt awful and fretted about it. J and B and them told me that it was nothing to worry about. Would you be mad if someone forgot your birthday? they asked. Probably not, but I imagine I would find it a little strange if my close friend who lived with me forgot for almost the entire day. Perhaps I felt bad because last year M completely forgot about mine and I was a little more sad about it than I'd like to admit.

In my defense I had been thinking about his birthday gift for the past few weeks and I did get him something. So I guess my self-flagellation results stems from the illusion that I was not the type of person to forget close friends' birthdays being shattered.

If J was bothered by my duncehood he did not let it show. We went to Cap't Loui for dinner today, which puts a Korean spin on the traditional seafood boil. One year later, I am again dreaming of a low country boil.

The house sauce they use is sweetened by gochujang and one of their specialty sides is garlic noodles. What do you like in your seafood boil?

*   *   *

I got breakfast with S this morning at Paloma in Greenpoint. Her friend called mid-hangout and S said to her, "The next time you come here we have to go get these croissants." They were fine. I don't really think my brain has wrapped itself around the fact that a single croissant can cost eight dollars.

And here I am writing home about them. Clockwise from top left: guava and key lime danish, herbs de Provence croissant, shakshuka croissant

After breakfast I suggested going to the library to work. Libraries here tend to have a lot of homeless people in them and in this location a few slept in the sunlit chairs facing the windows. I wasn't bothered, but when one started snoring loudly A said it was our cue to get out of here. "Doesn't that bother you?" she asked.

I was confused. "Why? Just tune it out."

"You're so tame," she replied as she packed up her things.

tame: INFORMAL
(of a person) willing to cooperate.

Later I asked her how she was feeling about her breakup. Her face lit up. "Oh! I don't know if it's because I'm getting older, but I feel like I just forgot and moved on after a while. Don't you feel that way too?"

I think the only part of breakups that has gotten easier for me is that I have learned the hard way what I should be doing and (more importantly) not doing after they happen. But when it comes to how much I suffer, if anything they get harder the older I get. My body takes longer to heal. Forgetting is so long. I think about things more now. I want everything to make sense.

*   *   *

I brought J to work with me on Friday. She'd brought me to her office a few times in the past, but her office isn't quite as nice and she doesn't know anyone that works there. I introduced her to my friends at work, who later asked me if we were dating. I explained that we were not. I always find it a little mind-boggling (and sad) how many people struggle to comprehend the existence of close platonic friendships between genders, not least because they make up the majority of my friendships.

*   *   *

I loved Sylvia's moments that stick in [her] mind, so much so that I have half a mind to steal the idea and write my own list. But what a depressing list it would be...

Hacks, Season 4
This season was a return to form with a hilarious yet heartfelt storyline, non-stop one-liners, new and exciting characters, and fun celebrity cameos. The relationship between Deborah and Ava is much more mature as the season develops, and the major turn of the season is unexpected. Jimmy and Kayla remain the MVPs of the series, and the addition of Randi to their twosome is inspired. I laughed, a lot, and can’t wait for next season.

Great British Baking Show, Season 15
There’s not much I can say about this show that I haven’t said before: its reality tv at its sweetest (pun intended) and perhaps best. You grow to love every contestant, are enamored by the judges and hosts, and the baking is always mouth-watering. This season, however, was probably the first where, after the first few episodes, I had no idea the eventual winner was going to win. Was a lovely result.

Another Simple Favor
There was a moment, about a third of the way through the movie, when we contemplated giving up on the film, but decided to give it a little bit more to see if it got better. It didn’t. And despite that we watched the whole thing. The writing is pure schlock and the outfits weren’t even as good this time as they were in the first film. Anna Kendrick and Blake Lively are usually fun to watch, but this material is abysmal rendering the two leads barely watchable. Did make me want to go to Capri, though.

Severance, Season 2
Undoubtedly, a second season of a series will be compared to its first, but in this case, the comparison would be inapt. This season of Severance is wildly different than the first, making it better and worse at the same time. The show keeps its tone and timbre, and the look stays the same, but the story goes to places you couldn’t imagine. Season one left us with some cliffhangers, and those are aptly dealt with in season two, which introduces new plot twists and characters and setting all with intrigue and mystery. I really hope they don’t keep us waiting three years for the next season.

New laws for new technologies

The tl;dr is: the court ruled that (a) piracy for the purpose of training an LLM is still piracy, so there’ll be a separate case about the fact that Anthropic did not pay for copies of all the books their model ingested, but (b) training a model on books and then selling access to that model, which can then produce output based on what it has “learned” from those books, is considered transformative work and therefore fair use.

Compelling arguments have been made both ways on this topic already, e.g.:

• Some folks are very keen to point out that it’s totally permitted for humans to read, and even memorise, entire volumes, and then use what they’ve learned when they produce new work. They argue that what an LLM “does” is not materially different from an impossibly well-read human.
• By way of counterpoint, it’s been observed that such a human would still be personally liable if the “inspired” output they subsequently created was derivative to the point of  violating copyright, but we don’t yet have a strong legal model for assessing AI output in the same way. (BBC News article about Disney & Universal vs. Midjourney is going to be very interesting!)
• Furthermore, it might be impossible to conclusively determine that the way GenAI works is fundamentally comparable to human thought. And that’s the thing that got me thinking about this particular thought experiment.

A moment of philosophy

Here’s a thought experiment:

Support I trained an LLM on all of the books of just one author (plus enough additional language that it was able to meaningfully communicate). Let’s take Stephen King’s 65 novels and 200+ short stories, for example. We’ll sell access to the API we produce.

The output of this system would be heavily-biased by the limited input it’s been given: anybody familiar with King’s work would quickly spot that the AI’s mannerisms echoed his writing style. Appropriately prompted – or just by chance – such a system would likely produce whole chapters of output that would certainly be considered to be a substantial infringement of the original work, right?

If I make KingLLM, I’m going to get sued, rightly enough.

But if we accept that (and assume that the U.S. District Court for the Northern District of California would agree)… then this ruling on Anthropic would carry a curious implication. That if enough content is ingested, the operation of the LLM in itself is no longer copyright infringement.

Which raises the question: where is the line? What size of corpus must a system be trained upon before its processing must necessarily be considered transformative of its inputs?

Clearly, trying to answer that question leads to a variant of the sorites paradox. Nobody can ever say that, for example, an input of twenty million words is enough to make a model transformative but just one fewer and it must be considered to be perpetually ripping off what little knowledge it has!

But as more of these copyright holder vs. AI company cases come to fruition, it’ll be interesting to see where courts fall. What is fair use and what is infringing?

And wherever the answers land, I’m sure there’ll be folks like me coming up with thought experiments that sit uncomfortably in the grey areas that remain.

🤘 You're subscribed to DanQ.me using the RSS feed. You rock! 🎸

I am writing to express my extreme disappointment that you voted to table impeachment charges against the president over the bombing of Iran. I understand that the issue would never pass because of the numbers, but YOU did not have to join Republicans in their complicity with a traitor to the Constitution. Symbolic votes *are* symbolic… so right now I am skeptical of your commitment to our laws. Are you up to the task of defending us from fascist oversteps, or will you sit by and do nothing to stop the actually pretty fast erosion of separation of powers and allow the creation of an American monarch? I urge you and other Democrats to act together as a true opposition party, not collaborators. If you had held the line and all voted together, that would still send a message of unity. Symbolic votes feel like the least I should be able to expect of my Representative.