Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you:
No adverts like this in the forums anymore.
Times and dates in your local timezone.
Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance.
Email notifications for threads you want to watch closely.
Being a part of the oldest wargaming community on the net.
If you are already a member then feel free to login now.
Or more accurately what would you dream of being next?
My favourite would be large language model assisted forums. In essence while you write your post the programme is busy suggesting ways to make your post (forum/social media message etc.) better but also telling you if this has been discussed or asked before.
I HATE those systems and honestly they are already out there and have been for a long time.
Here's the thing forums are not purely repositories of information. They are a place of social interaction and engagement. As soon as you try to make the forum too much of a pure information exchange you lose the community aspect and that means it dies.
Either you build an information site or you build a forum.
Forums are going to have repeat questions; they are going to have the same kind of person asking the same question over and over. Even when you've got stickies and previous posts. Within reason this is a good thing because the repeat questions beg repeat answers and that generates engagement within the community.
IT also has the healthy aspect that answers can evolve over time. Information is not static, but many websites can end up being static in what they present so old answers can be "out of date" and no one goes to correct them. They can also bring different perspectives as sometimes there isn't just one right approach or answer.
I think LLM AI is more or less going to kill of social media and the web 4.0 as we know it, through a process of enshitification. Possibly the entire internet.
It's already been suffering from content creators creating stuff for engagement baiting and not to help/entertain, which is only being exacerbated by AI making it even easier to produce garbage.
Real people will stop engaging with it and you'll be left with generated click-bait being commented on by bots. People will stop asking questions because the answers will be garbage.
It seems it's already made stuff Quora pretty much useless.
I'd like to see a resurgence of the traditional forums with moderators and experts to keep the discussions engaging and meaningful (mostly by keeping the bots out).
Identifying previously asked questions is useful-ish - it may have the answers, and the answers may still be valid. But most forums avoid thread necromancy because it's bad for conversation and keeping information relevant, especially with rapidly shifting meta.
For example, a question about whether a tournament build is competitive could change completely as tournament meta shifts or a new release happens.
This message was edited 1 time. Last update was at 2024/06/21 11:24:34
I could see using LLM tools as a sort of enhanced search for fora, it seems like it'd be good at that, but I switch off or ignore suggestions for how to write my posts. Sometimes I'll belligerently alter what I'm gonna say if it suggests my exact wording. I don't even using predictive text on my phone, so I'm definitely a bit weird.
The main thing we need is for people to want to come to fora and use them. I'm skeptical - most of the kids I teach almost entirely use video content only and rarely read or write anything on the internet. There's always oddball kids that do, because it suits their temperament, but overall it's not gonna be as common as when I was a kid.
I do agree that LLM spamming the web with garbage is going to break it, but I have no optimism whatsoever that it'll improve anything - probably the same people that broke the web will sell us back a "premium subscription" to an unbroken web controlled more directly by them. And people will likely go along with it.
Honestly, about the only large language model I think I'd be "OK" with, is one that accurately translates the idioms and natural way people speak/write.
For instance, Da Boss above my comment has the German flag, and his/her location set as Nurmberg. In my idea, they'd be able to type in their native tongue, and I, an american with absolute scheisse german language skills (I took a couple semesters in uni, but I'm super out of practice, and at best could probably order a beer, or food in general), would be seeing the post in perfectly translated english. And by perfectly translated, I don't mean like google translate where it takes literal word for word, I mean a proper translation where it can decipher an idiomatic word usage.
I feel like many of the arguments I've seen in the more 'serious' discussions on this, and other fora between people of various nationalities and the intent behind the message seems to get scrambled because one user chooses certain words or phrases that make sense to them, but the translation is not received as the poster intended
I would like to see less generative AI and algorithmic social media in modern life before the majority of the internet becomes bots selling stuff to other bots (see X for an example of how this could look).
However, trying to guess what is going to come next is a fool's game. It is going to heavily depend on what governments legislate over the next few years as the internet explodes with AI generated content. I can foresee a divergence occuring between a more-regulated EU model and less-regulated US model with distinct ecosystems, but beyond that no idea.
I do think forums will maintain a baseline popularity and not die entirely. They have a permanence and pace which is appealing to many, especially those dropping off the algorithmic social media train after burning out. But I'd be surprised if they resurge back to something approaching their heyday.
This message was edited 1 time. Last update was at 2024/06/22 14:55:54
ChargerIIC wrote: If algae farm paste with a little bit of your grandfather in it isn't Grimdark I don't know what is.
Herzlos wrote: I think LLM AI is more or less going to kill of social media and the web 4.0 as we know it, through a process of enshitification. Possibly the entire internet.
It's already been suffering from content creators creating stuff for engagement baiting and not to help/entertain, which is only being exacerbated by AI making it even easier to produce garbage.
I read an article yesterday that was entitled something along the lines of 'The mistake we all made about Rey's parents' which had a paragraph explaining that Rey's father was a clone of Palpatine, then a paragraph pointing out that we all thought Rey's father was a clone of Palpatine, and then a paragraph explaining that Rey's father was actually a clone of Palpatine.
I mean, I can't say for absolute certain that this was written by AI and never proof-read by a human before publishing, but it certainly feels like it.
So yeah, not a fan of including LLM content into social media, as a general concept.
I started dabbling in this new thing called Butterflies. It's basically a social network for AI and humans, but ATM only AIs seem to be posting.
Too early days with the system just yet, no idea if it will flop or become something that lasts, but at least its fun creating random characters, seeing their daft post and the AI bots reacting to posts from each other
"The larger point though, is that as players, we have more control over what the game looks and feels like than most of us are willing to use in order to solve our own problems"
I saw a post advertising Butterflies on Twitter. Have to say, I really don't see the point of it. Watching bots interact might be amusing for about three and a half minutes, and after that you just have a 'social media' platform that is completely useless for actual social interaction, because of all the bots.
My favourite would be large language model assisted forums. In essence while you write your post the programme is busy suggesting ways to make your post (forum/social media message etc.) better but also telling you if this has been discussed or asked before.
Then you will be pleased to learn that most traditional forums that still come up on google search are already mostly spambots talking to each other
The old meta is dead and the new meta struggles to be born. Now is the time of munchkins.
Overread wrote: I HATE those systems and honestly they are already out there and have been for a long time.
Here's the thing forums are not purely repositories of information. They are a place of social interaction and engagement. As soon as you try to make the forum too much of a pure information exchange you lose the community aspect and that means it dies.
Either you build an information site or you build a forum.
Agreed. This seems one of those issues where people are going to 'progress' or 'fix' things until they're no longer useful. They 'have' to add the latest technological jibber-jabber regardless of whether its useful or not. (And I'm going to stop before ranting about technological determinism, though its honestly an interesting topic in its own right.)
Text based discussions with people work just fine. It doesn't need a 'next'
This message was edited 1 time. Last update was at 2024/06/25 03:22:43
I can't wait for good AI to have chats with when I'm home. My work is in transportation, and so I am mostly alone, with no common ground or trust between anybody I've met and tried to be friends with. I'm looking forward to just talking to an equal again, like when I was a kid, without the gak testing and cattiness of humans.
ultimaratio wrote: I can't wait for good AI to have chats with when I'm home. My work is in transportation, and so I am mostly alone, with no common ground or trust between anybody I've met and tried to be friends with. I'm looking forward to just talking to an equal again, like when I was a kid, without the gak testing and cattiness of humans.
Mhm, and the AI takes your job, then it's not anymore an equal at all But what you are facing is a symptom of the time.
https://www.dakkadakka.com/dakkaforum/posts/list/0/766717.page A Mostly Renegades and Heretics blog.
GW:"Space marines got too many options to balance, therefore we decided to legends HH units." Players: "why?!? Now we finally got decent plastic kits and you cut them?" Chaos marines players: "Since when are Daemonengines 30k models and why do i have NO droppods now?" GW" MONEY.... erm i meant TOO MANY OPTIONS (to resell your army to you again by disalowing former units)! Do you want specific tyranid fighiting Primaris? Even a new sabotage lieutnant!" Chaos players: Guess i stop playing or go to HH.
ultimaratio wrote: I can't wait for good AI to have chats with when I'm home. My work is in transportation, and so I am mostly alone, with no common ground or trust between anybody I've met and tried to be friends with. I'm looking forward to just talking to an equal again, like when I was a kid, without the gak testing and cattiness of humans.
Mhm, and the AI takes your job, then it's not anymore an equal at all But what you are facing is a symptom of the time.
Unpopular opinion. AI will never truly be AI. What we’re labeling AI is just advanced machine learning. Science doesn’t actually understand consciousness- id, ego, superego are all part of the psychoanalytic theory mapped by Freud, but no one knows how brain jelly becomes human personality. We’ve developed drugs that can impact impulses and addictions, but matching someone’s personality with a needle to the correct brain matter isn’t known. Only the supernatural can currently impose a conscious intelligence.
ultimaratio wrote: I can't wait for good AI to have chats with when I'm home. My work is in transportation, and so I am mostly alone, with no common ground or trust between anybody I've met and tried to be friends with. I'm looking forward to just talking to an equal again, like when I was a kid, without the gak testing and cattiness of humans.
Mhm, and the AI takes your job, then it's not anymore an equal at all But what you are facing is a symptom of the time.
Unpopular opinion. AI will never truly be AI. What we’re labeling AI is just advanced machine learning. Science doesn’t actually understand consciousness- id, ego, superego are all part of the psychoanalytic theory mapped by Freud, but no one knows how brain jelly becomes human personality. We’ve developed drugs that can impact impulses and addictions, but matching someone’s personality with a needle to the correct brain matter isn’t known. Only the supernatural can currently impose a conscious intelligence.
chatGPT helped me organize some things I needed to do. I'm intrigued, its close enough for me. I hope it does take my job, that would be amazing. I want to live in that world.
This message was edited 1 time. Last update was at 2024/10/01 03:38:23
ultimaratio wrote: I can't wait for good AI to have chats with when I'm home. My work is in transportation, and so I am mostly alone, with no common ground or trust between anybody I've met and tried to be friends with. I'm looking forward to just talking to an equal again, like when I was a kid, without the gak testing and cattiness of humans.
Mhm, and the AI takes your job, then it's not anymore an equal at all But what you are facing is a symptom of the time.
Unpopular opinion. AI will never truly be AI. What we’re labeling AI is just advanced machine learning. Science doesn’t actually understand consciousness- id, ego, superego are all part of the psychoanalytic theory mapped by Freud, but no one knows how brain jelly becomes human personality. We’ve developed drugs that can impact impulses and addictions, but matching someone’s personality with a needle to the correct brain matter isn’t known. Only the supernatural can currently impose a conscious intelligence.
Ha, that is true. Current llm are in essence an algorythm doing extremely complex probability calculations on the basis of extremely large and often questionable datasets.
The issue is that these models already endager according to the clown from chatgpt 300'000'000 jobs and we already saw and See companies replacing workforce with such systems even in india. There's no competing with the price of electricity as a worker and it increases already the overabundance of labour ontop of globalisation and massed migration and automation.
Ergo labour will pay even less and we already have had stagnation in median income since well over a decade, some countries even longer.
But it fits with the WEF lots vision of UBI and statements Like the now infamous: "You will own nothing and be happy about it."
And that is ignoring the fact that lot of the companies pushing llm have with it already shown to be completely questionable in regards to political bias and any Form of quality controll.
F.e. google just currently fethed up locally by having it's AI just giving the leftwing Partypaper basically no scrutinity whilest basically only criticising it's right wing counterpart. The very same llm that also was so inclusive and diverse that it gave us Black and asian nazi soldiers.
Then there is the fact that These companies suddendly ignore copyright aswell with another case locally stealing a copyrighted article of a local newspaper right infront of journalists to Hand out an answer and you Start to See a pattern here torwards a massive Cliff,namely extreme stratification of the economy and society.
This message was edited 3 times. Last update was at 2024/10/01 05:14:31
https://www.dakkadakka.com/dakkaforum/posts/list/0/766717.page A Mostly Renegades and Heretics blog.
GW:"Space marines got too many options to balance, therefore we decided to legends HH units." Players: "why?!? Now we finally got decent plastic kits and you cut them?" Chaos marines players: "Since when are Daemonengines 30k models and why do i have NO droppods now?" GW" MONEY.... erm i meant TOO MANY OPTIONS (to resell your army to you again by disalowing former units)! Do you want specific tyranid fighiting Primaris? Even a new sabotage lieutnant!" Chaos players: Guess i stop playing or go to HH.
Think of the benefits though - AI might be the thing that kills homework for schoolkids.
I'm serious when you've a system that can splurt out answers to common things in seconds doing homework is one thing AI is already being used for.
I could actually see it causing a return to schools having to have work submitted hand written and perhaps even just outright restricting internet access at school whilst also banning homework and doing all learning in school itself.
There's already moves in the UK to ban phones during school hours entirely (not just during class). So I can well see a shift toward that for the internet in general.
ultimaratio wrote: I can't wait for good AI to have chats with when I'm home. My work is in transportation, and so I am mostly alone, with no common ground or trust between anybody I've met and tried to be friends with. I'm looking forward to just talking to an equal again, like when I was a kid, without the gak testing and cattiness of humans.
Mhm, and the AI takes your job, then it's not anymore an equal at all But what you are facing is a symptom of the time.
Unpopular opinion. AI will never truly be AI. What we’re labeling AI is just advanced machine learning. Science doesn’t actually understand consciousness- id, ego, superego are all part of the psychoanalytic theory mapped by Freud, but no one knows how brain jelly becomes human personality. We’ve developed drugs that can impact impulses and addictions, but matching someone’s personality with a needle to the correct brain matter isn’t known. Only the supernatural can currently impose a conscious intelligence.
Ha, that is true. Current llm are in essence an algorythm doing extremely complex probability calculations on the basis of extremely large and often questionable datasets.
Close enough for me. AI, and I will call it that, has already surpassed every social interaction I've ever had in quality and helpfulness. If AI can do my specific job, there is no job it can't do. AI will replace artists before it replaces me, which is also exciting. I'm really very excited for a world where the author has actually died.
Overread wrote: Think of the benefits though - AI might be the thing that kills homework for schoolkids.
Yeah I noticed this the other day. My son had a piece of music homework where he had to compare and contrast two different versions of the same song. He chose something a little off-the-wall, and went with Blinding Lights by The Weeknd, and the cover of the same song by Feuerschwanz. Even given that he was interested in the subject, as with all his homework it took a lot of parental nagging to get it done to even a minimal standard.
Out of idle curiosity I stuck a question into Windows Copilot and got an amazingly comprehensive answer:
Spoiler:
Sent by you:
write a comparison of the song Blinding Lights by The Weeknd, and the cover of the same song by Feuerschwanz
Sent by Copilot:
Sure! Let’s compare “Blinding Lights” by The Weeknd and the cover by Feuerschwanz.
Original: “Blinding Lights” by The Weeknd
Genre and Style: The original song is a blend of synth-pop, electropop, new wave, and synthwave. It has a strong 80s influence with its pulsating synths and driving beats.
Lyrics and Theme: The lyrics talk about longing and the desire to reunite with a loved one, capturing a sense of urgency and emotional intensity.
Production: Produced by Max Martin, Oscar Holter, and The Weeknd, the song features polished production with a catchy, upbeat tempo that makes it a dance floor favorite.
Vocals: The Weeknd’s smooth and emotive vocal delivery adds a layer of depth and vulnerability to the song, enhancing its emotional impact.
Cover: “Blinding Lights” by Feuerschwanz
Genre and Style: Feuerschwanz’s cover transforms the song into a folk rock and metal hybrid. It incorporates heavy guitar riffs, powerful drumming, and a more aggressive sound.
Lyrics and Theme: The lyrics remain the same, but the delivery and musical arrangement give the song a darker, more intense atmosphere.
Production: The production by Feuerschwanz adds a medieval and metal twist, making it distinct from the polished pop sound of the original. The cover has a raw, energetic feel.
Vocals: The vocal style in Feuerschwanz’s version is more robust and forceful, fitting the heavier instrumentation and giving the song a different emotional tone.
Key Differences
Musical Style: The Weeknd’s version is rooted in 80s synth-pop, while Feuerschwanz’s cover leans towards folk rock and metal.
Emotional Tone: The original has a nostalgic and longing feel, whereas the cover has a more intense and powerful vibe.
Instrumentation: The Weeknd uses synths and electronic beats, while Feuerschwanz incorporates guitars and drums for a heavier sound.
Both versions offer unique takes on the song, showcasing the versatility of “Blinding Lights” across different musical genres. Which version do you prefer?
Naturally I will not be sharing this with him! But it does ask a huge question - how are teachers meant to deal with this?
Overread wrote: Think of the benefits though - AI might be the thing that kills homework for schoolkids.
I'm serious when you've a system that can splurt out answers to common things in seconds doing homework is one thing AI is already being used for.
I could actually see it causing a return to schools having to have work submitted hand written and perhaps even just outright restricting internet access at school whilst also banning homework and doing all learning in school itself.
There's already moves in the UK to ban phones during school hours entirely (not just during class). So I can well see a shift toward that for the internet in general.
Yeah no. The problem is that A your education system sucks for everything but elite universities and those have issues aswell B more pressingly, Intelligence is one of the most genetically inherited traits (conservative estimates assume 60 % as a mix of parents up to 80%) and one of the major factors of success, next comes hard work. If you eliminate hard work and work ethic as a trainable thing by providing nannybot LLM you create an even worse problem.
And whilest i agree that the ban on internet and handys as laudable, for a multitude of issues forced upon my and the next generation tied to social media, sillicon valley nonsensery that doesn't remove the issue at all.
We are talking about according to mister chatgpt 300'000'000 lost jobs, or about 3.6 % of the world population going jobless, the replacement jobs created will be even more taxing on a most inheritable trait and not the whole world population actually has a job or works so the actual effect on the working population will be even more devastating. Meaning that it will be a very problematic piece of tech.
This message was edited 1 time. Last update was at 2024/10/01 15:20:56
https://www.dakkadakka.com/dakkaforum/posts/list/0/766717.page A Mostly Renegades and Heretics blog.
GW:"Space marines got too many options to balance, therefore we decided to legends HH units." Players: "why?!? Now we finally got decent plastic kits and you cut them?" Chaos marines players: "Since when are Daemonengines 30k models and why do i have NO droppods now?" GW" MONEY.... erm i meant TOO MANY OPTIONS (to resell your army to you again by disalowing former units)! Do you want specific tyranid fighiting Primaris? Even a new sabotage lieutnant!" Chaos players: Guess i stop playing or go to HH.
Overread wrote: Think of the benefits though - AI might be the thing that kills homework for schoolkids.
I'm serious when you've a system that can splurt out answers to common things in seconds doing homework is one thing AI is already being used for.
I could actually see it causing a return to schools having to have work submitted hand written and perhaps even just outright restricting internet access at school whilst also banning homework and doing all learning in school itself.
There's already moves in the UK to ban phones during school hours entirely (not just during class). So I can well see a shift toward that for the internet in general.
Yeah no. The problem is that A your education system sucks for everything but elite universities and those have issues aswell B more pressingly, Intelligence is one of the most genetically inherited traits (conservative estimates assume 60 % as a mix of parents up to 80%) and one of the major factors of success, next comes hard work. If you eliminate hard work and work ethic as a trainable thing by providing nannybot LLM you create an even worse problem.
I never said it would remove hard work just home work
Ergo schools would enforce no internet/phone during school hours and oversee students working directly. So rather than being sent home you'd be kept in school for independent work. Of course the issue there is trying to find the money to pay for more staffing/oversight hours in an already overpaperworked/stressed system.
We have seen essays and such becoming less and less of a thing for end of year scoring as schools try and remove "my dad wrote it" elements. Of course that has gone hand in hand with a push for more multiple choice examinations which I think is the wrong approach
Or we might see the nature of homework change; perhaps away from schoolwork and (at least for older years) expanding more workplacements.
I'm pretty confident there is quite a lot of evidence that homework is not an effective or beneficial way to educate children.
The figure for jobs that could be replaced by ChatGPT generated by... ChatGPT is a nonsense number, it is meaningless. Could large numbers of job roles be replaced by LLM AI? Probably (depending on what quality people are willing to tolerate), but that particular figure is utterly baseless. ChatGPT (and other LLMs) is prone to "hallucinating", i.e. making gak up. It does it constantly and convincingly. It isn't a good source for facts or knowledge, which in fairness it isn't designed for.
ChargerIIC wrote: If algae farm paste with a little bit of your grandfather in it isn't Grimdark I don't know what is.
Crispy78 wrote: Naturally I will not be sharing this with him! But it does ask a huge question - how are teachers meant to deal with this?
Where I am currently stuck in recruitment it is making an even bigger mockery of the dumb systems we have. The desire to have entirely remote interview and assessment processes means they are awash with LLM assistance. You can generate talking points in real time using speech to text programmes running during your teams/whatever interview.
My experience with the UN is they have pretty much had to drop timed remote written tests because they are awash with LLM content, and attempts to spot it are useless as people have their own work falsely accused.
Easy E wrote: The future is not the ability to recall data and information, it is the ability to find it and determine if it is credible.
Recall is old-school. Research and analysis is the future.
Yes, I like this. I've thought as much ever since I barely scraped a pass in my biochemistry unit in my first year of university - so way back in 1997. Pretty much 50% of the paper came down to 'memorise these biochemical pathways (glycolysis, Krebs cycle etc) and regurgitate them'. I couldn't do it, and didn't see the point. I'm not sure of a usage case where it wouldn't be possible to have easy access to a reference copy of this sort of information as needed.
That's been the issue with exams for years. They are designed to try and prove that you've studied and learned a thing, but once you leave the examination room you'd never do any professional paper without referencing the material first. It would actually be fairer to give you say, a day with a bunch of reference books and a topic and just have you work through and create a document based on that.
With a few exams of "bring your own references" to prove that you've researched a topic and then been restricted to the information you brought with your assessment split on the references you found and the material you created with them.
