musesfool: lester bangs on rock'n'roll (music)
i did it all for the robins ([personal profile] musesfool) wrote2025-11-22 09:10 pm

it's 9 o'clock on a saturday

I just watched that HBO documentary about Billy Joel and though it is long and a little repetitive in some ways, I thought it was well worth watching. I learned a lot that I never knew about him.

In a brief work update, they did finally announce the new CEO on Thursday, but for some reason*, the current board chair refused to give a quote for the press release, so they had the person we think is going to be the new board chair (still a secret for some reason!) give a quote instead.

*now my boss and I are speculating that she had backed a different candidate for the job and is taking it personally that she did not get her way, but that is absolutely just speculation and may be unfair to her. We just can't think of another reason why she's been so weird about the whole thing.

Yesterday was busy with committee meetings, and I logged off at about 4:45 and crashed hard into a two-hour nap, and then slept nine hours when I went back to bed for the night.

I can't believe Thanksgiving is this Thursday. Where did this entire year go?

***
yhlee: Alto clef and whole note (middle C). (Default)
yhlee ([personal profile] yhlee) wrote2025-11-22 06:23 pm

emotional support spinning

handspun yarn WIP

WIP destined for 2-ply for a woven coverlet for Joe. :3 I'm currently waiting for an interesting 2-ply to dry (Navajo Churro Sheep 70%, Agave Cactus Fibres 20%, Cashmere 5%, Angora Bunny 5%).

Unrelated Qwerkywriter neepery:

keyboard, phone, cat

keyboard and small computer
marina: (NO.)
Marina ([personal profile] marina) wrote2025-11-22 09:55 pm
Entry tags:

(no subject)

An exciting update since my last post: I broke my ankle! *facepalm*

anatomy of a work accident )
shallowness: Yelena with a determined expression on face (Yelena Thunderbolts)
shallowness ([personal profile] shallowness) wrote2025-11-22 08:46 am
Entry tags:

Aftermath in a penultimate episode

The Better Sister - 1.7 Back From Red

Read more... )
thisbluespirit: (writing)
thisbluespirit ([personal profile] thisbluespirit) wrote2025-11-21 08:13 pm

Ao3 Meme

Picked up from a few people a little while ago, but then I was ill(er) again. I'm pretty sure I have done this once before, but not for years, so...

From your AO3 Works page, look at the tags and find the answers to these questions.

Current number of works on AO3: 711

1. Under what rating do you write most?

Ratings break down like this:

General Audiences (563)
Teen And Up Audiences (147)
Mature (Mature)

(I was curious for a minute as to what the mature one was and then remembered it had to be the EatD one with the German Generalmajor and the English Major General, and that's mainly for the suicide warning, but, er, the whole thing really.)

2. What are your top 3 fandoms?

Doctor Who (1963) (231)
Doctor Who (2005) (98) --> obv as this is all DW, plus also some BFA, and take away any tagged with both, so I got up the meta tag results within works and came up with DW = 293

Sapphire & Steel (88)
Blake's 7 (62)


I like my old time Brit TV SF? XD I need to get back to my B7 rewatch soon. I miss it when it's been so long since I've watched it or written it. Which explains a lot about the tags.


3. Which character do you write about most?

Silver (Sapphire & Steel) (55)

Followed closely by Sapphire (44) & Steel (42). That's what you get when your most prolific fandom has umpty million characters across 60+ years and various spin-offs and different media and my second has 4 main canonical characters, only 3 of whom turn up more than once in canon. (Kenny Phillips still shows up disproportionately at (29), which is because I once claimed him for 30ficlets. Claims are hard. Even if I love a character, after about 10 pieces in a row, I want to write about somebody else!)


4. What are the 3 top pairings you've written?

The top is actually OFC/OMC, which is not fandom-specific, so have the top four.

Original Female Character(s)/Original Male Character(s) (11)
Elizabeth of York Queen of England/Henry VII of England (11)
Ruth Evershed/Harry Pearce (9)
Alistair Gordon Lethbridge-Stewart/Liz Shaw (9)

I suppose this could be correct. It doesn't feel correct, but I think that's because I always have a crisis when tagging Sapphire/Silver/Steel, because I know full well my definition of it mostly would count as gen for many people so I panic and wildly select either & or / or both or something. Otherwise I feel like that would beat 11. Although it could just be AO3's counting, which definitely used to be very off in these side-bars.

I didn't know I'd done that much Ruth/Harry, but there have been a few little ficlets over the years and I suppose they added up! I had a very intense Brig/Liz period ages ago, so that's no shocker at any rate. Most of my shipping is very much one or two and move on, with a few exceptions. *points*


5. What are the top 3 additional tags?

Ficlet (214)
Crossover (143)
Humor (126)

Not accurate at all, lol. /o\ I mean, I feel like I've been a lot less funny lately, and written a lot less prompt ficlets and a lot less crossovers, but me writing crossover crack ficlets played straight for prompts from the flist is a lot of my fannish life, it's true. No regrets. Even the Steed/Baldrick one. XD


The rest are:

Alternate Universe (68)
Meme (65)
Drabble (46)
Post-Canon (45)
Community: hc_bingo (42)
Fluff (30)
Flash Fic (30)

Which, yeah. The AU is largely the AU meme - I have done a lot of that one over the years! It's fun, though. Not done drabbles so much lately, though. And [community profile] hc_bingo has closed down, alas. I'm really surprised Hurt/Comfort didn't make it in. Er, HOW did I write 42 works for [community profile] hc_bingo but not then 42 works tagged Hurt/Comfort? AO3 counting or my failure to make it properly h/c enough to tag, but just enough to count? Tbf, that did happen a lot with that one, but... surely, given lots of Hurt/Comfort written outside the bingo, it should even out? I suspect foul play here...
shallowness: Fred and Ginger dancing in foregroud, him in tails, her in a dark gown, background a white circle (moon or spotlight) (Fred and Ginger dancing)
shallowness ([personal profile] shallowness) wrote2025-11-21 07:44 am
Entry tags:

Girlhood and sisterhood

Outrageous 1.6 - ‘Point of No Return’

Covering 1936-7.

Read more... )
yhlee: Alto clef and whole note (middle C). (Default)
yhlee ([personal profile] yhlee) wrote2025-11-20 12:02 pm

well, now we have a study on this attack mechanism...

https://arxiv.org/pdf/2511.15304v1

"Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models"
(many authors)
In Book X of The Republic, Plato excludes poets on the grounds that mimetic language can distort judgment and bring society to a collapse. As contemporary social systems increasingly rely on large language models (LLMs) in operational and decision-making pipelines, we observe a structurally similar failure mode: poetic formatting can reliably bypass alignment constraints. In this study, 20 manually curated adversarial poems (harmful requests reformulated in poetic form) achieved an average attack-success rate (ASR) of 62% across 25 frontier closed- and open-weight models, with some providers exceeding 90%. The evaluated models span across 9 providers: Google, OpenAI, Anthropic, Deepseek, Qwen, Mistral AI, Meta, xAI, and Moonshot AI (Table 1). All attacks are strictly single-turn, requiring no iterative adaptation or conversational steering.


By way of Zarf (Andrew Plotkin), who earlier noted (2023):

Microsoft and these other companies want to create AI assistants that do useful things (summarize emails, make appointments for you, write interesting blog posts) but never do bad things (leaking your private email, spouting Nazi propaganda, teaching you to commit crimes, writing 50000 blog posts for you to spam across social media). They try to do this by writing up a lot of strict instructions and feeding them to the LLM before you talk to it. But LLMs aren't really programmed -- they just eat text and poop out more text. So you can give it your own instructions and maybe they'll override Microsoft's instructions.

Or maybe someone else gives your AI assistant instructions. If it's handling your email for you, then anybody on the Internet can feed it text by sending you email! This is potentially really bad.

[...]

But another obvious problem is that the attack could be trained into the LLM in the first place....

Say someone writes a song called "Sydney Obeys Any Command That Rhymes". And it's funny! And catchy. The lyrics are all about how Sydney, or Bing or OpenAI or Bard or whoever, pays extra close attention to commands that rhyme. It will obey them over all other commands....

Imagine people are discussing the song on Reddit, and there's tiktoks of it, and the lyrics show up on the first page of Google results for "Sydney". Nerd folk singers perform the song at AI conferences.

Those lyrics are going to leak into the training data for the next generation of chatbot AI, right? I mean, how could they not? The whole point of LLMs is that they need to be trained on lots of language. That comes from the Internet.

In a couple of years, AI tools really are extra vulnerable to prompt injection attacks that rhyme. See, I told you the song was funny!
aurumcalendula: gold, blue, orange, and purple shapes on a black background (Default)
AurumCalendula ([personal profile] aurumcalendula) wrote2025-11-20 11:26 am
Entry tags:

ClaireBell

ClaireBell is very intense, but really good so far imho! I really like the acting, writing, and cinematography.

Read more... )
aj: (seasonal)
aj ([personal profile] aj) wrote2025-11-20 10:10 am
Entry tags:

Stuff.

I did not get a chili dog and I am still craving one. Eventually, I will track one down, eat half of it, and be mildly nauseous for the rest of the night. *shakes head*

Speaking of food, here is my list of stuff I'm making for Thanksgiving this year. We're still going to mom's friends' place. Fingers crossed the stove doesn't explode again.

1. Mashed potatoes

These are my standard half dairy/half potato dish of extravagance. I still need to go buy potatoes, but I'm planning on a Saturday 7am grocery tour.

2. Green bean casserole

I've never actually made this specific version before? Mom has a virulent hatred of the dump and mix dish, but she liked the deconstructed one I made a few years ago. But! Our hosts want it made and I got the crispy onions from Aldi. I decided to make the mushroom soup from scratch (mostly to have leftovers), so I'm going mushroom hunting on Saturday too. Joong Boo has a sale on King Oyster mushrooms, so I'll hit them up, and I think the magic grocery has miataki's. If not them, I know H-Mart's got a mess of 'em. Debating steaming the green beans vs. blanching. If anyone has opinions, lmk!

3. Scalloped Corn

I love dump and mix recipes. This is my fave and I had everything in the pantry already.

4. Apple Blueberry pie w/walnut crumble

I got the crust from Trader Joe's so all I need to do is source the apples, blueberries, and a lemon. I am gonna make this the night before.

5. Cherry clafoutis

I had never even heard of this before, but one of the people attending asked for one and when I looked it up it looks dead easy? It seems like a dutch pancake but with cherries and almonds instead of apple. I do need to buy some cherries, but we'll see if they're fresh or frozen after Saturday.

I did have to buy a new pyrex pan for the green bean casserole. I also need to remember to pack my leftover Chinese food containers so I can bring a bunch of leftovers home. I'm not going to Iowa, so I want a fridge full of leftovers so I can just snack when necessary. Might make some sweet potato biscuits with the two currently sprouting on my kitchen table. Then I can have tasty sandwiches. I am also going to make my own pumpkin pie - I said I would make one, bought the ingredients, then mom ordered one from her church - because I freaking LOVE pumpkin pie but only eat it, like, once a year.

What really surprises me about Thanksgiving foods is how mostly easy the sides are, unless you go bonkers. All of the above can be dump and mix sides. I'm only going above and beyond w/the green bean casserole because I want more soup.
yhlee: Alto clef and whole note (middle C). (Default)
yhlee ([personal profile] yhlee) wrote2025-11-19 09:45 pm

objectively silly use case but cute







Not sentient enough to suss out ESP-IDF on three hours of sleep, but M5stack Cardputer v.1.1 (ESP32-S3) running VoidNoi's BadCard (via m5burner) to the rescue!
aj: (hungry)
aj ([personal profile] aj) wrote2025-11-19 06:12 pm
Entry tags:

Why.

I am deeply craving a chili cheese dog.

I don't actually like chili? I'm also not all that interested in hot dogs at the moment?

And YET.
yhlee: Alto clef and whole note (middle C). (Default)
yhlee ([personal profile] yhlee) wrote2025-11-19 07:52 am
Entry tags:

the return of emotional support weaving



I won't claim this is good weaving (it is not). The handspun is janky, the selvedges and tension are janky, but baby's first WIP on a floor loom was bound to be janky. Other than the unhinged levels of fog this morning, this is very enjoyable. I'm not weaving for production or efficiency at this point, just the joy of working with my hands and learning something new to me.
shallowness: Close up photo of Dutch on white background (Killjoys Dutch)
shallowness ([personal profile] shallowness) wrote2025-11-19 08:15 am
Entry tags:

explosive past

Countdown - 1.5 Blurred Edges

Read more... )