chaorace

Have you considered overriding the Emacs process EDITOR variable via setenv? i.e.:

    (setenv "EDITOR" "/usr/bin/emacsclient -c")

If you do that, it should cause any child processes (e.g.: vterm) to inherit the replaced EDITOR variable and properly open new editor frames instead of trying to do a terminal takeover.

It lets you hold hands with strangers

What you describe sounds almost exactly like the definition of playful teasing laid out in the article:

> For playful teasing to be successfully interpreted as affiliative rather than aggressive, the teaser, to some extent, has to understand the recipient's expectations and predict their likely reaction. Likewise, the recipient needs to draw accurate inferences and correctly identify the teaser's intent as affiliative, looking beyond any mildly abrasive behavioural elements.

The stark difference in social boundaries you observe is actually an affirmation of the hypothesis that playful teasing acts as a social tool for discovering how to navigate different individual social dynamics.

pixel displacement

Well that was a bit of a doomy article, though I find myself agreeing with the general throughline. Dating's gotten pretty messed up and realistically speaking that's unlikely to change going forward.

The reason I characterize the article as "doomy" is because it implies some very terrible things will happen as a result:

> "A society with too many isolated men and no war to send them to is a scary situation. [...] they will decay your country from the inside."

Far be it from me to suggest that endemic loneliness and population decline are positive developments, but... will society really decay from the inside-out as a direct result? I'll cop to being one of those men which the article characterizes as having "given up" -- a sub-yearly sex haver, a porn watcher, and an enjoyer of the video games -- but I'm no closer to blighting civilized society because of it.

Cards on the table, that's probably thanks to various tailwinds in my own personal life: a fulfilling job, a good paycheck, and a modest yet supportive network of friends and family whom I love dearly. There're plenty of guys (and girls) out there with none of those things and no sex life to top it off. Life sucks for them in general and that is probably the stronger (and more actionable!) source of this malaise creeping over general society which many of us are feeling.

I admire your pragmatism. I think it's quite neat how your chosen name reflects that aspect of your self, too.

In their defense: the OP's deployment of the term "payday" is sloppy when discussing a world which predates salaries, bank transfers, and indeed... weeks[1]

[1]: https://en.wikipedia.org/wiki/Ancient_Mesopotamian_units_of_...

> but is there any reason other than promoting the GPL why a non-free module should not use linux-only APIs?

I think this question illustrates something of an invisible generation gap in software. GPL is not a utilitarian license. There is no materialistic value in choosing to use the GPL whatsoever. The text of the GPL is itself a legal edifice specifically designed to dismember the very framework in which it exists, rendering it anti-utilitarian.

This is all to say that the GPL is politics. One chooses the GPL specifically for political reasons. In politics, you do not go out of your way to aid your ideological adversaries; you either convert them or destroy them.

> if getting rid of non-free modules is the goal, why even have a syscall exception? why not force all modules to be GPL?

Despite what I just said about GPL as ideology, the GPL is simultaneously a legal construct which necessarily must bend to the rules of the framework it exists within. The logic of GPL extending through syscalls is based in a legal argument similar to that used in the Google vs. Oracle debacle -- we created the "shape" of this API, so we have certain rights regarding it. The "shape" of the Unix syscalls predate Linux and are therefore beyond the project's reach, legally speaking.

To a certain extent, I think it's just IBM disease. A company the size of Meta is expected to have an AI research department like Microsoft or Google, even if their core business (social media) derives relatively less benefit from the technology.

Pretend you're an uncreative PM on an AI team; what part of Facebook or VR could you feasibly improve by iterating on LLMs? Perhaps the content moderation system... but that would require wrangling with the company ethics comittee and someone else at the company probably already took ownership that idea. You've gotta do something compelling or else your ML engineers are going to run off somewhere else.

If I were to ask my ML engineers about what they wanted to work on, they're going to avoid areas where their model is outgunned (i.e.: chat) and instead prefer lower hanging fruit which generalizes well on a resume (i.e.: "Pioneered and published key innovations in LLM code-generation").

Of course, the alternative answer is that Meta wants to replace all of their jr. developers with GPUs, but I think their leadership is a little too preoccupied with VR to be actively pushing for such a transformative initiative in anything more than a very uninvested capacity (e.g.: "Sure I'll greenlight this. Even if it doesn't pay off I don't have any better ideas")

A little yes, a little no. Your phone knows when it's receiving a call from an unauthenticated number, but it would be very unreliable to use this metric alone to decide when a call is untrustworthy. My understanding is that carrier telemetry is what drives the final yes/no verdict.

Sounds like an energy weapon to me!

The neat thing is that -- whether or not LK-99 is a hoax -- the public will have engaged with real scientists doing real science in a rather personal capacity. It's novel and interesting to be able to tune into the materials science equivalent of live-coding.

At the end of the day, materials science is still science. The institutional framework is optimized for a very specific process, so it's generally faster to let the process play out as usual rather than go and cut corners. Rest assured; there are a lot of scientists out there! We can afford to let a few of them chase clouds once in a while.

In any case... the creation process described in the original paper is relatively cheap and low-tech enough that labs will likely generate their own samples in less time than any procurement process would take.

Yes, it's generally legal to apply an additional surcharge on card transactions beyond the cash price, thus making it appear as if the sales tax is being removed. It's all above board as long as the full value of the sale is reported and taxed (exact details vary by tax code, of course).

It's generally illegal, however, for a business to evade taxes by failing to report the full value of cash sales. It's also legally dubious as a customer to knowingly help facilitate such transactions (e.g.: anti-laundering laws)... but I don't think it's common to catch heat for this if you're an otherwise law-abiding person.

A case like this would be a landmark battle in the making, I think. Advocacy groups like the IEEE would most likely be all over this due to the weightly ramifications that would arise from Nintendo setting a new unfavorable legal precedent.

Yeah and if I lit a match and dropped it in my house it'd be a housefire. Calling skins NFTs is patently false, because they are both not tokens and not non-fungible (nametags notwithstanding).

Surprisingly, some of the most convincing studies indicating the dangers of PFAS come from 3M themselves[1][2]. These studies were of course never intended for public release:

> In 1961, company lab tests linked C8 exposure to enlarged livers in rats and rabbits. DuPont scientists then conducted tests on humans, asking a group of volunteers to smoke cigarettes laced with C8. “Nine out of 10 people in the highest-dosed group were noticeably ill for an average of nine hours with flu-like symptoms that included chills, backache, fever and coughing,” the researchers noted. “Concerns about the potential toxicity of C8 had been raised internally within DuPont by at least 1954, leading DuPont’s own researchers to conclude by at least 1961 that C8 was toxic and, according to DuPont’s own Toxicology Section Chief, should be ‘handled with extreme care,'” Bartlett’s February 2013 suit against DuPont alleged.

> In one study, for instance, Olsen looked at blood tests of 3M employees at the company’s plants in Antwerp, Belgium, and Decatur, Alabama, both of which made PFOA and PFOS, among other products. By the late 1990s when Olsen was embarking on this research, these chemicals were known within the company to accumulate in humans and alter cholesterol levels in lab animals. Because the workers had undergone three separate rounds of blood tests, Olsen was able to trace the levels of the chemicals in workers’ blood over time. And by combining his results with various clinical measures the company had been tracking in its workers, he was able to see whether there was a relationship between the chemical and these health outcomes. Olsen’s findings, written up in an draft report in October 2001, were clear. There was a positive association between the amount of PFOA in workers’ blood and their levels of cholesterol and triglycerides, states the report, on which Olsen is listed as the principal investigator.

They've also taken some... drastic measures regarding which factory workers are allowed to be exposed to certain PFAS despite publicly denying any connection with birth defects[3]:

> In a November 1982 memo, DuPont's chief medical director, Dr. Bruce Karrh, expressed concern over employee exposure to C8. The company then barred women of child-bearing age from working around C8.

Far be it from me to condemn an abundance of caution, but the timing seems odd when you consider what happened a year prior[3]:

> When Bailey returned to work in 1981, she discovered a memo on a bench in the women's locker room. It detailed a study by 3M, which sold C8 to DuPont, that detailed eye deformities in lab animals whose mothers were fed C8 during pregnancy. Bailey went to the plant's medical doctor and asked if this is what happened to [her child born with an eye deformity]. The doctor denied a connection, according to Bailey, but arranged an appointment at Johns Hopkins University Hospital in Baltimore.

[1]: https://theintercept.com/2018/07/31/3m-pfas-minnesota-pfoa-p... [2]: https://www.ecowatch.com/teflons-toxic-legacy-dupont-knew-fo... [3]: https://www.delawareonline.com/story/news/2016/04/02/c8-susp...

Instances = Servers. Communities = Subreddits.

What you're looking for are home automation communities, such as:

- https://lemmy.ml/c/homeautomation

- https://lemmy.ml/c/homeassistant

- https://lemmy.world/c/homeautomation

Regardless of which instance you specifically join, you can subscribe to all three of these communities to follow & participate in their discussions. Here's an instance browser to help you decide where you want to make your account: https://lemmyverse.net/

It generally doesn't matter which instance you sign up on, but here's a few guidelines if you just want maximum access with minimum hassle. This information isn't really particularly necessary, but it may help alleviate the choice paralysis if you happen to be anything like me:

- Don't sign up for Hexbear -- it's a non-federated instance, so accounts there can't browse outside content.

- Try to avoid joining the largest instances (5k+ users) -- large instances are more likely to be blocked by others for various reasons (e.g.: lemmy.world & sh.itjust.works are blocked by beehaw.org)

- Try to avoid joining instances which specifically advertise themselves as catering to contentious or risque content (e.g.: lemmynsfw.com, burggit.moe, lemmygrad.ml) -- pretty self-explanatory why admins tend to like blocking these, I think!

- Try to avoid joining small (<1k users) and/or joke instances -- if your home-instance goes down, your account & comments will disappear with it, so choose something that looks like it has some staying power. This will also help streamline your onboarding experience because bigger instances are generally more strongly connected to everything else in the threadiverse.

- Still stumped? I can happily recommend my home-instance: https://lemmy.sdf.org -- not too big, not too small, highly innocuous content, and a long history of hosting robust fediverse servers.

Yup. You can join any of the dozens of medium-sized, unassuming instances and not have to worry about defederation at all. It's only the tall poppies that get reaped (i.e.: very large instances & instances with controversial stances).

I'm glad you asked! First, some resources:

- Community Browser: https://lemmyverse.net/communities

- Anime Community Compilation: https://sopuli.xyz/post/779648

In my experience, these have been the most reliable for seasonal anime episode discussions:

- https://lemmy.ml/c/anime

- https://lemmy.world/c/anime

I haven't browsed the manga communities as much, but so far these have been good for chapter discussions:

- https://lemmy.ml/c/manga

- https://feddit.de/c/manga

Yes. Instances are also publicly indexible by search engines.

So, I looked at the table appendix you're referencing and I think you're overstating your case a bit.

Among books within copyright, GPT-4 can reproduce Harry Potter and the Sorcerer's Stone with 76% accuracy. This is, apparently, the highest accuracy GPT-4 achieved among all tested copyrighted books with 1984 taking a distant 2nd place at 57%.

With this in mind, we can verifiably say that GPT-4 is unusually good at specifically reproducing the first Harry Potter book. An unscrupulous book thief may very well be able to steal the first entry in the series... assuming that they're able to get past one quarter of the book being an AI hallucination.

Welcome to HN: where NixOS is evangelized with the thunderous battlecry of "People who hate configuring love configuring NixOS! If you hate configuring too, you should stop using that system you've already finished configuring and then learn to configure this new one!"

There are lots of "big deals" in the environment that we silently tolerate -- things which studies conclusively and repeatedly show are substantially detrimental:

* Roadside air pollution

* Lead pipes

* Indoor CO2 buildup

Studies relating to flouride are nowhere near as conclusive and show nowhere near the same magnitude of effect. If flouride is contributing negatively in a similar way, it's doing so to such a small extent that it very nearly falls within the margin of error. Anyone genuinely worried about flouride should consider double-checking that they've already taken more meaningful action regarding their health, such as going on a walk, getting a full night's rest, or opening a window.

Well, this is the state we're talking about... the threat of 30-60 days in jail might be sufficient to scare most small-time tax dodgers into line. Anyone bigger than that's probably already legally avoiding taxes in the first place anyways.

First: these studies are unrelated to AI and mostly predate modern LLMs, so no, they draw no direct comparisons. At best, this is pointing out two surprisingly similar behaviors we observe in LLMs and the actual functioning of the human mind.

Second: Being unable to reconstruct a prior mental state =/= remembering a thought process. I'll use an accursed programing analogy because we are on an accursed programing website: it's like the difference between debug logs and time-travel debugging. In plain language: humans retroactively reason about their thought process based on the artifacts produced by that thought process. The quality and reliability of these artifacts will vary heavily, but the reasoning itself is always backward-looking and speculative in nature.

This probably varies quite a lot, but whenever I fail such tests[1], the first person they contact has always been the employer/school/team, not me. Usually, the business is discrete enough to privately reach out and allow me to explain myself. More than once, however, I would only find out about my failure and imminent expulsion after-the-fact via letter/email -- I've always been able to fight my way out of these situations, but often without my personal reputation intact.

[1]: Which I always do, as a person who has been prescribed Adderall since childhood[2]

[2]: Yes, the diagnosis is legitimate. It was diagnosed by a trusted family pediatrician during a regular check-up and was not solicited by my parents[3]

[3]: Yes, it is exhausting to explain myself, my parents, and my medical history like this whenever the topic comes up. Woe is me, woe is me, etc.

Same boat. A few years back I spent a summer living off LTE internet (before it was cool!) using a Sierra Wireless USB modem bodged into barely working with the help of pirated reference docs. Man am I glad that I gave up on that idea -- the OG AT protocol is, indeed, a totally obtuse crock of shit to the point where, honestly, it barely even qualifies for the moniker of "protocol" at all.

This reminds me of the recent so-called "Glitch Token" phenomenon[1]. When GPT-3 was presented with reserved tokens it never encountered during training, it reacted in extremely unpredictable ways -- often with a simple "fuck you".

For those unfamiliar with LLM architecture: "tokens" are the smallest unit of lexical information available to the model. Common words often have their own token (e.g.: Every word in the phrase "The quick brown fox jumped over the lazy dog" has a dedicated token), but this is a coincidence of compression and not how the model understands language (e.g.: GPT-3 understands "defenestration" even though it's composed of 4 apparently unrelated tokens: "def", "en", "est", "ration").

The actual mechanism of understanding is in learned associations between tokens. In other words: the model understands the meaning of "def","en","est","ration" because it learns through training that this cluster of tokens has something to do with the literary concept of violently removing a human via window. When a model encounters unexpected arrangements of tokens ("en","ration","est","def"), it behaves much like a human might: it infers the meaning through context or otherwise voices confusion (e.g.: "I'm sorry, what's 'enrationestdef'?"). This is distinctly different from what the model does when it encounters a completely alien form of stimulation like the aforementioned "Glitch Tokens".

At the risk of anthropomorphizing, try imagining if you were having a conversation with a fellow human and they uttered the following sentence "Hey, did you catch the [MODEM NOISES]?". You've probably never before heard a human vocalize a 2400Hz tone during casual conversation -- much like GPT-3 has never before encountered the token "SolidGoldMagicarp". Not only is the stimulus unintelligble, it exists completely beyond the perceived realm of possible stimulus.

This is pretty analagous to what we'd call "undefined behavior" in more traditional programming. The model still has a strong preference for producing a convincingly human response, yet it doesn't have any pathways set up for categorizing the stimulus, so the model kind of just regurgitates a learned lowest-common-denominator response (insults are common).

This oddly aggressive stock response is interesting, because it's actually the exact same type of behavior that was coded into one of the first chatbots to (tenuously) pass a Turing test. I'm of course referring to the "MGonz" chatbot created in 1989[2]. The MGonz chatbot never truly engaged in conversation -- rather, it continuously piled on invective after invective whilst criticizing the human's intelligence and sex life. People seem predisposed to interpreting aggression as human, even when the underlying language is, at best, barely coherent.

[1]: https://www.youtube.com/watch?v=WO2X3oZEJOA [2]: https://timharford.com/2022/04/what-an-abusive-chatbot-teach...

There are no officially adopted text-representations (for IPv4), only binary ones. In other words: there are simultaneously infinitely more possibilities and exactly zero.

For the most part, applications tend to punt the job of interpreting such text address representations to the IP stack (usually embedded in the OS kernel). These vary in what they'll accept by implementation and version, but they tend to be extremely good at interpreting whatever arbitrary nonsense people have historically been likely to try. As a result, there are surprisingly few application-level libraries which even attempt to deal with that mess.