Be advised: this thread is not for serious in-depth discussion of weighty topics (we have a link for that), this thread is not for anything Culture War related. This thread is for Fun. You got jokes? Share 'em. You got silly questions? Ask 'em.
- 171
- 1
Jump in the discussion.
No email address required.
Notes -
Item 1 of 7
Prominent publicly-traded mattress manufacturer Sleep Number has filed for bankruptcy, and has been taken private for 700 M$ by a Canadian mattress retailer. It claims that, of its 3000 employees, 38 bigwigs—including nine vice presidents and 13 directors—are important enough to the company's operations that they need to be enticed to stay during the transition to new ownership. It proposes to do so by giving to each of these employees a retention payment of approximately 48 k$.
The bankruptcy trustee objects to this idea, claiming that these payments would constitute illegal distribution of assets to insiders. But the bankruptcy judge overrules the objection. Despite being called vice presidents and directors, these employees are not members of the Executive Leadership Team (which consists of six senior and executive vice presidents in addition to the CEO), and are not members or appointees of the board of directors. Therefore, the employees in question are not insiders, and the retention payments are perfectly fine.
Item 2 of 7
Some borderline-shitpost writing
Previous discussion in this vein
Item 3 of 7
A truck is reported stolen by its owner. Two weeks later, a police officer pulls over a tow truck that is towing the allegedly-stolen truck. The tow-truck driver no habla inglés, but he makes a phone call to his English-speaking boss, who tells the officer that the owner of the property where the truck was stored thought the truck was abandoned and hired the boss to tow the truck away. However, the driver overhears the phone conversation and (through an interpreter) tells the officer that the boss told the driver a different story—that the boss purchased the truck from its owner and was in the process of selling it to a third party.
On these facts, the tow-truck driver receives a felony charge of receiving stolen property. However, the trial judge dismisses the charge, and the appeals panel affirms. Even assuming that the boss stole the truck (one judge on the panel cares enough to write a separate opinion complaining that the boss hasn't been charged with theft despite being unable to produce the allegedly-stolen truck's title), the prosecutor has produced no evidence whatsoever that the driver knew that the truck was stolen. Rather, the evidence shows that the driver had no idea that the boss was telling two different stories about how he obtained the truck (which would have justified him to suspect that the boss had stolen the truck) until after he was pulled over and overheard the boss's conversation with the police officer. And the crime of receiving stolen property requires proof that the alleged fence knew that the property was stolen.
(The driver remains on the hook for the less-serious offenses of driving without a license and without taillights.)
Item 4 of 7
Imagine that you object to a will as invalid, but your claim fails. Normally, you would pay your own attorney's fees. However, New Jersey apparently has a court rule allowing the probate judge to make the estate pay your attorney's fees if your objection "had reasonable cause".
Item 5 of 7
Hampton sends text messages to his ex-girlfriend, Stephanie. Stephanie's current romantic partner, Lawrence, sees the texts and calls Hampton on the telephone. An argument ensues. Hampton shows up outside Stephanie's apartment building, and Lawrence walks outside to have a fistfight. Less than a minute later, Stephanie walks outside and finds Lawrence beaten to a pulp while Hampton kicks him one last time and then is driven away by a friend. Lawrence is hospitalized for five months, and remains unable to walk without assistance.
Hampton is indicted for the second-degree felony of aggravated assault—attack resulting in "serious bodily injury" (inter alia, "protracted loss of function"), with a maximum sentence of ten years. The trial judge instructs the jury to additionally consider the "lesser included offense" of third-degree aggravated assault—attack resulting in "significant bodily injury" ("temporary loss of function"), with a maximum sentence of five years.
Hampton asks the judge to add instructions on two misdemeanor lesser included offenses, simple assault (attack resulting in mere "bodily injury"; maximum sentence six months) and simple assault with mutual combat (maximum sentence thirty days). But the judge denies the request. No reasonable jury could find that the impairment inflicted on Lawrence does not rise to the level of "significant bodily injury", so simple assault is not relevant to the jury's deliberations.
The jury convicts Hampton of second-degree aggravated assault, and the judge sentences him to seven years in prison. The appeals panel affirms the conviction, but remands for resentencing since the trial judge incorrectly counted Lawrence's "horrific" injuries in this case toward the "prior criminal record" aggravating factor.
Item 6 of 7
Mists of time past: My parents buy a corded lawnmower, and make me (and my brother) use it to mow the lawn.
2023: I stumble across an article claiming that battery-powered lawnmowers are far weaker than gasoline mowers. My own calculations on the basis of Home Depot product listings bear out the claims: 200–500 watts for battery mowers, 1400–1600 watts for corded mowers, and 2000–2500 watts for gasoline mowers.
Winter 2024–2025: I buy a vacant lot. Of course, it has no electric service.
Spring 2025: I buy a battery-powered lawnmower for use on the vacant lot. As indicated by my previous calculations, it is noticeably weaker than the corded mower that I have used for the past decade. In fact, despite being marketed as capable of handling a 1/2-acre property, its battery runs dry halfway through my 1/6-acre lot! I return the battery-powered mower in favor of a gasoline-powered mower, and use that for the 2025 lawnmowing season.
Winter 2025–2026: My contract with a homebuilding contractor does not explicitly mention lawnmowing, but does explicitly say that, if I am injured on the lot while work is ongoing, then I will not be covered by the contractor's insurance. I interpret this as implying that, during construction, lawnmowing will be performed by the contractor rather than by me. Accordingly, I sell the gasoline mower on Craigslist.
Late July 2026: The contractor previously razed all vegetation on the lot as part of excavation. But now the grass is getting pretty tall. I send an email complaint to the contractor. Its response: "We can take care of the grass. We were looking for someone to handle the mowing since we currently do not have anyone on staff available to maintain it."
Early August 2026 (one week later): The contractor still hasn't cut the grass. I attend the council meeting for final adoption of the municipality's new zoning code. Before the meeting, the code-enforcement official notices me, and informs me that he has sent to me an official letter telling me to cut the grass on my lot. I guess I'll have to rent either a gasoline-powered mower or a gasoline-powered generator from the local Home Depot.
Item 7 of 7
Raheem is standing on a street corner. Anwar drives up to him, and Anwar's passenger Kevon shoots him, but he survives. On the other side of the narrow, one-way street, Tiarra happens to be sitting in her parked car with her one-year-old child CC. She starts screaming, and Kevon turns around and shoots her dead. Both Anwar and Kevon are charged with, and convicted of, a laundry list of crimes—conspiracy to murder Raheem, attempted murder of Raheem, conspiracy to murder Tiarra, murder of Tiarra, and reckless endangerment of CC.
On appeal, Anwar admits that he conspired to kill Raheem, but complains that he had nothing to do with the crimes that Kevon committed against Tiarra and CC of his own accord. But the trial judge rejects this argument, and the appeals panel affirms. Anwar conspired to kill Raheem in a drive-by shooting, and it is reasonable to infer that Anwar contemplated the possibility that Kevon would additionally eliminate witnesses as part of that method of killing.
You held back there, as one does with this subject...
The whole point is not to shrink back!
Now I think about it I'd be worse at writing smut than Deepseek V4 Pro. I could do a better plot and more lively prose generally, more creativity and wouldn't get confused... but it'd be so awkward I'd be unable to do it.
Also I think you'd be better off with longer prompts. I die a little when I see these benchmarks for creative writing with one sentence prompts. A good creative writing prompt is at least 600 words I think, maybe more than that. I'm not all the way into lorebooks but you gotta give the bot something to work with. If you don't have the LLM executing your ideas it will go full Generic Genericsson. Then again, maybe local LLMs are more constrained for context length.
Common local LLMs range from 128k to 1m context windows, and while I'd recommend aiming to use less than half of the window per session, that still covers most short stories easily. Very long prompts extend prefill time, but this only really starts to matter if you're running on CPU or mixed GPU+CPU, or have most of a book written already.
For metrics, Hecatomb's 85k word Wild Pair series (cw: furry, mostly m/f, extreme kinks in ways that mean decensored local LLM review is the only option) is 120k tokens, and takes about 30 seconds to calculate prefill on an nVidia 3090. For a non-smut example, Doctorow's I, Robot (cw: very annoying writing), is 15k words, 20k tokens, and was less than a second on prefill.
Conversely, LLMs will invent details, and while they'll be average, they aren't going to be that generic. This can definitely go weird places -- the tendency for LLM-driven names to end up as variants of Kael, Lyra, so on -- but even mild pressure will get you to specifics, for better or worse.
Yeah Kael syndrome is what I was thinking about.
But also are you gonna get any good quality after about 50,000 words in? Even in bigger models I notice writing starting to decline, errors becoming more prominent...
More options
Context Copy link
I've recently been nooticing that something is definitely very weird with naming stuff in LLM-generated
smuttext. I'm usually a purveyor of uh, established can(n)ons so that usually helps corral the models into a more coherent mold, but recently I've been cautiously trying OC stuff based on old draftsand the naming "schemas" really stick out.I don't know how to put it in proper words, but it feels like without a defined frame of reference ("you are {{char}} from {{vidya}}, &c"), there are certain "clustered" defaults that models resort to, based on which rough genre/setting the LLM mentally sorts your scenario into. My prompts are detailed enough that I don't get super generic stuff like my good ol friend Elara on a trek to the Whispering Woods shudder anymore, I'm not talking about those (though the existence of such certified hood classics independently encountered by so many people is in itself quite interesting), but the "mental models" the LLMs seem to keep are still quite noticeable in generic/open-ended OC scenarios - to the point that when bringing up say an as-yet-unnamed nearby town (with only the general setting and maybe name of the current one to go off of), the model's response will declare exactly the same %town_name% across several regens, despite being vaguely prompted and ostensibly taking a fresh crack at the worldbuilding every time. I have yet to test if those differ across LLMs, likely they do. Hopefully someone smarter than me is looking into it, or at least has nooticed.
I had a look at what normal people put into LLMs as prompts, you should radically lower your estimates of consciousness, literacy and taste.
Don't think those can get lower at this point but sure, I'll count my blessings. Back when chub was still usable, I remember having seen a fork of one of my publicly posted cards, with only one difference - a very bobs-and-vagene-coded sentence added to its definitions and greeting message (from the user's point of view too, not the character's like the founding fathers intended!). Granted I have zero pretensions to poetic genius here, the card was like 600 tokens of simple descriptions, but the preferences and literacy of the average coomer have since become painfully clear. In that moment Rayon was enlightened.
I regret to remind you that 98% of people will never make a card or a fork or anything you can see on the user side... I'm talking about looking at it from the admin side, the shit people put in as prompts. AI is wasted on the masses.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link