I’m excited for the future, but not as excited for the transition period.

I have similar feelings.

I discovered LLMs before the hype ever began (used GPT-2 well before ChatGPT even existed) and the same with image generation models barely before the hype really took off. (I was an early closed beta tester of DALL-E)

And as my initial fascination grew, along with the interest of my peers, the hype began to take off, and suddenly, instead of being an interesting technology with some novel use cases, it became yet another technology for companies to show to investors (after slapping it in a product in a way no user would ever enjoy) to increase stock prices.

Just as you mentioned with the dotcom bubble, I think this will definitely do a lot of good. LLMs have been great for asking specialized questions about things where I need a better explanation, or rewording/reformatting my notes, but I’ve never once felt the need to have my email client generate every email for me, as Google seems to think I’d want.

If we can just get all the over-hyped corporate garbage out, and replace it with more common-sense development, maybe we’ll actually see it being used in a way that’s beneficial for us.

I understand why people seem to think we should tolerate these views, because “muh free speech,” but to them, I say:

IPFS seems similar to what you’re looking for.

(See: A copy of Wikipedia on IPFS being censorship-resistant, and globally distributed)

I like ArchiveBox, but in my experience, it kept on running into issues saving pages, and stopped functioning after it worked the first few times. I really wish there was a more streamlined application that did a similar thing somewhere out there.

I’ve been looking at Linkwarden’s page archiving solution, but it crashes whenever I try importing any large number of links, so that’s a bust too.

Reminds me of this great street art I saw posted at some point. Can’t find the source though.

That’s definitely true, I probably should have been a little more clear in my response, specifying that it can run at startup, but doesn’t always do so.

I’ll edit my comment so nobody gets the wrong idea. Thanks for pointing that out!

To put it very simply, the ‘kernel’ has significant control over your OS as it essentially runs above everything else in terms of system privileges.

It can (but not always) run at startup, so this means if you install a game with kernel-level anticheat, the moment your system turns on, the game’s publisher can have software running on your system that can restrict the installation of a particular driver, stop certain software from running, or, even insidiously spy on your system’s activity if they wished to. (and reverse-engineering the code to figure out if they are spying on you is a felony because of DRM-related laws)

It basically means trusting every single game publisher with kernel-level anticheat in their games to have a full view into your system, and the ability to effectively control it, without any legal recourse or transparency, all to try (and usually fail) to stop cheating in games.

Computers are a fundamental part of that process in modern times.

If you were taking a test to assess how much weight you could lift, and you got a robot to lift 2,000 lbs for you, saying you should pass for lifting 2000 lbs would be stupid. The argument wouldn’t make sense. Why? Because the same exact logic applies. The test is to assess you, not the machine.

Just because computers exist, can do things, and are available to you, doesn’t mean that anything to assess your capabilities can now just assess the best available technology instead of you.

Like spell check? Or grammar check?

Spell/Grammar check doesn’t generate large parts of a paper, it refines what you already wrote, by simply rephrasing or fixing typos. If I write a paragraph of text and run it through spell & grammar check, the most you’d get is a paper without spelling errors, and maybe a couple different phrases used to link some words together.

If I asked an LLM to write a paragraph of text about a particular topic, even if I gave it some references of what I knew, I’d likely get a paper written entirely differently from my original mental picture of it, that might include more or less information than I’d intended, with different turns of phrase than I’d use, and no cohesion with whatever I might generate later in a different session with the LLM.

These are not even remotely comparable.

Assuming the point is how well someone conveys information, then wouldn’t many people better be better at conveying info by using machines as much as reasonable? Why should they be punished for this? Or forced to pretend that they’re not using machines their whole lives?

This is an interesting question, but I think it mistakes a replacement for a tool on a fundamental level.

I use LLMs from time to time to better explain a concept to myself, or to get ideas for how to rephrase some text I’m writing. But if I used the LLM all the time, for all my work, then me being there is sort of pointless.

Because, the thing is, most LLMs aren’t used in a way that conveys info you already know. They primarily operate by simply regurgitating existing information (rather, associations between words) within their model weights. You don’t easily draw out any new insights, perspectives, or content, from something that doesn’t have the capability to do so.

On top of that, let’s use a simple analogy. Let’s say I’m in charge of calculating the math required for a rocket launch. I designate all the work to an automated calculator, which does all the work for me. I don’t know math, since I’ve used a calculator for all math all my life, but the calculator should know.

I am incapable of ever checking, proofreading, or even conceptualizing the output.

If asked about the calculations, I can provide no answer. If they don’t work out, I have no clue why. And if I ever want to compute something more complicated than the calculator can, I can’t, because I don’t even know what the calculator does. I have to then learn everything it knows, before I can exceed its capabilities.

We’ve always used technology to augment human capabilities, but replacing them often just means we can’t progress as easily in the long-term.

Short-term, sure, these papers could be written and replaced by an LLM. Long-term, nobody knows how to write papers. If nobody knows how to properly convey information, where does an LLM get its training data on modern information? How do you properly explain to it what you want? How do you proofread the output?

If you entirely replace human work with that of a machine, you also lose the ability to truly understand, check, and build upon the very thing that replaced you.

Schools are not about education but about privilege, filtering, indoctrination, control, etc.

Many people attending school, primarily higher education like college, are privileged because education costs money, and those with more money are often more privileged. That does not mean school itself is about privilege, it means people with privilege can afford to attend it more easily. Of course, grants, scholarships, and savings still exist, and help many people afford education.

“Filtering” doesn’t exactly provide enough context to make sense in this argument.

Indoctrination, if we go by the definition that defines it as teaching someone to accept a doctrine uncritically, is the opposite of what most educational institutions teach. If you understood how much effort goes into teaching critical thought as a skill to be used within and outside of education, you’d likely see how this doesn’t make much sense. Furthermore, the heavily diverse range of beliefs, people, and viewpoints on campuses often provides a more well-rounded, diverse understanding of the world, and of the people’s views within it, than a non-educational background can.

“Control” is just another fearmongering word. What control, exactly? How is it being applied?

Maybe if a “teacher” has to trick their students in order to enforce pointless manual labor, then it’s not worth doing.

They’re not tricking students, they’re tricking LLMs that students are using to get out of doing the work required of them to get a degree. The entire point of a degree is to signify that you understand the skills and topics required for a particular field. If you don’t want to actually get the knowledge signified by the degree, then you can put “I use ChatGPT and it does just as good” on your resume, and see if employers value that the same.

Maybe if homework can be done by statistics, then it’s not worth doing.

All math homework can be done by a calculator. All the writing courses I did throughout elementary and middle school would have likely graded me higher if I’d used a modern LLM. All the history assignment’s questions could have been answered with access to Wikipedia.

But if I’d done that, I wouldn’t know math, I would know no history, and I wouldn’t be able to properly write any long-form content.

Even when technology exists that can replace functions the human brain can do, we don’t just sacrifice all attempts to use the knowledge ourselves because this machine can do it better, because without that, we would be limiting our future potential.

This sounds fake. It seems like only the most careless students wouldn’t notice this “hidden” prompt or the quote from the dog.

The prompt is likely colored the same as the page to make it visually invisible to the human eye upon first inspection.

And I’m sorry to say, but often times, the students who are the most careless, unwilling to even check work, and simply incapable of doing work themselves, are usually the same ones who use ChatGPT, and don’t even proofread the output.

Just like how the moment their videotape rental history was exposed, that was when privacy became an absolute must in the case of video rental services.

Except we don’t have any widespread evidence for cocaine being taken outside of highly specific medicinal cases being helpful to the health and wellbeing of the individual.

When it comes to gender affirming care, we have substantial evidence that proves it is safe and effective, as even a cursory glance at medical research on the topic will show:

https://www.thelancet.com/journals/lanam/article/PIIS2667-193X(23)00118-7/fulltext

https://www.tandfonline.com/doi/full/10.3109/09540261.2015.1115753

https://jamanetwork.com/journals/jamanetworkopen/article-abstract/2789423

https://www.liebertpub.com/doi/full/10.1089/trgh.2015.0008

I mean, there are definitely people in the government working on it, but those often require much more substantial reforms and systemic changes before the changes could functionally work. (i.e. banning data brokers would kill off most free services, or banning targeted ads would kill most ad-funded news networks)

If you haven’t already, I recommend using the EFF’s Action Center to let your representatives know about specific changes you would and would not want made to our laws to protect privacy, free speech, and digital innovation, according to what they’ve found to be the most pressing issues at the moment.

so basically I’m being forced into trial.

Sovcit finds out what the law is for the first time (heartbreaking)

“1 in 10 people believe they are not at risk when using illicit sources to watch TV, film or sports.”

ONE in ten? Man, they’re even bad at cherry picking statistics 😂

They even cite a study with only 1,000 participants for their statistic that “32% OF PEOPLE HAVE BEEN VICTIMS OF FRAUD”

In the title, at least. The body of that claim’s card says that it’s the people, or someone they know that have been victims of fraud.

Gosh, I hate dishonest scare marketing campaigns.

I’m not a big expert on database technology, but I am aware of there being at least a few database systems (“In-Memory”) that use the RAM of the computer for transient storage, and since RAM doesn’t use files as a concept in the same way, the data stored there isn’t exactly inside a “file,” so to speak.

That said, they are absolutely dwarfed by the majority of databases, which use some kind of file as a means to store the database, or the contents within it.

Obviously, that’s not to say using files is bad in any way, but the possibilities for how database software could have developed, had we not used files as a core computing concept during their inception, are now closed off. We simply don’t know what databases could have looked like, because of “lock-in.”

That’s what some databases are. Most databases you’ll see today still inevitably store the whole contents of the DB within a file with its own format, metadata, file extension, etc, or store the contents of the database within a file tree.

The notion of “lock in” being used here doesn’t necessarily mean that alternatives don’t or can’t exist, but that comparatively, investment into development, and usage, of those systems, is drastically lower.

Think of how many modern computing systems involve filesystems as a core component of their operation, from databases, to video games, to the structure of URLs, which are essentially usually just ways to access a file tree. Now think of how many systems are in use that don’t utilize files as a concept.

The very notion of files as an idea is so locked-in, that we can rarely fathom, let alone construct a system that doesn’t utilize them as a part of its function.

Regardless, the files example specifically wasn’t exactly meant to be a direct commentary on the state of microblogging platforms, or of all technology, but more an example for analogy purposes than anything else.

What social media platforms don’t have some kind of character limit?

What platforms don’t use a feed?

What platforms don’t use a like button?

What platforms don’t have some kind of hashtags?

All of these things are locked-in, not necessarily technologically, but socially.

Would more people from Reddit have switched to Lemmy if it didn’t have upvotes and downvotes? Are there any benefits or tradeoffs to including or not including the Save button on Lemmy, and other social media sites? We don’t really know, because it’s substantially less explored as a concept.

The very notion of federated communities on Lemmy being instance-specific, instead of, say, instances all collectively downloading and redistributing any posts to a specific keyword acting as a sort of global community not specific to any one instance, is another instance of lock-in, adapted from the fediverse’s general design around instance-specific hosting and connection.

In the world of social media, alternative platforms, such as Minus exist, that explore unique design decisions not available on other platforms, like limited total post counts, vague timestamps, and a lack of likes, but compared to all the other sites in the social media landscape, it’s a drop in the bucket.

The broader point I was trying to make was just that the very way microblogging developed as a core part of social media’s design means that any shift away from it likely won’t actually gain traction with a mainstream audience, because of the social side of the lock-in.

short form content with just a few sentences per post sucks.

I 100% agree with this sentiment.

Jaron Lanier has a great book called You Are Not A Gadget, where he talks about the way we design and interact with systems, and he has some thoughts I think reflect this sentiment very well:

“When [people] design an internet service that is edited by a vast anonymous crowd, they are suggesting that a random crowd of humans is an organism with a legitimate point of view.” (This is in reference to Wikis like Wikipedia)

“Different media designs stimulate different potentials in human nature.”

He talks about how when a system becomes popular enough, it can “lock in” a design, when others build upon it as standard. Such as how the very concept of a “file” is one we created, and nearly every system now uses it. Non-file based computing is a highly unexplored design space.

And the key part, which I think is relevant to Mastodon, the fediverse, and social media more broadly, is this quote:

“A design that share’s Twitter’s feature of providing ambient continuous contact between people could perhaps drop Twitter’s adoration of fragments.”

Fragments, of course, meaning the limited, microblogging style of communication the platform allows for. I’ve seen some Mastodon instances that help with this, by not imposing character limits anywhere near where most instances would, opting for tens of thousands of characters long. But of course, there is still a limit. Another design feature by Twitter that is now locked in.

But of course, people are used to that style of social media. It’s what feels normal, inevitable even. Changing it would mean having to reconceptualize social media as a concept, and might be something people aren’t interested in, since they’re too used to the original design. We can’t exactly tell.

As Lanier puts it,

“We don’t really know, because it is an unexplored design space.”

Just because an LLM sounds smart and human-like doesn’t mean it will magically solve climate change after being directly implicated in resource consumption we know from actual scientists, today, will make the problem worse.

They don’t believe it.

They just think their investors will.

The problem is that welfare systems, such as those that provide housing, that distinguish who is eligible by how much they can afford it, to a certain degree, inevitably depress higher levels of economic activity, and good saving behavior, through the very nature by which they’re operated.

If we say that someone is no longer eligible for free housing if they earn, say, $2,000+ a month, and housing would otherwise cost $500 a month, then if they’re currently earning $1,500 a month (the same they’d effectively have if they had to pay $500 a month for housing on a $2,000 a month salary) they have a direct incentive to not make over $2,000, unless they can guarantee they’ll make at least that much plus $500 more to compensate for the difference. If they earned $1,800 a month, they’d be making $300 more than someone making $2,000, but paying $500 a month for housing after hitting the cutoff.

This isn’t just a hypothetical either. While this 2021 study does mention some benefits of means testing, such as more targeted expenditure, it ultimately shows that…

“An asset means-test incentivizes low-income households to hold few financial assets making them vulnerable to predictable and unpredictable income changes.”

…and sees that, in the end, while it can marginally increase the cost of these social programs to the taxpayer, it ultimately does more to benefit the individuals receiving the assistance.

Or how about this research done by the Cleveland branch of the Federal Reserve that states:

“the elimination of testing limits, such as in policies similar to a UBI, could present a welfare-improving alternative to the current system, though not without large economic trade-offs.” (They effectively mean worse targeting of funds, but better overall results)

Means-testing directly reduces the incentives that lead to higher overall household wealth, and quality of life.

Not having means-testing increases total income, which also means increased tax revenue. That same tax revenue can then go to funding the housing system as a whole, but it won’t directly, substantively punish people for an increase in income past an arbitrary threshold.

Not to mention the increased administrative cost of performing means testing, as opposed to doing unconditional support, which could reduce the amount of money actually going to funding housing, in favor of funding jobs for people that audit income levels of housing applicants.