Are forums dying?

[Zero999]

Because that's not what happens in reality. Too many new members ask simple questions which are easily answered by a search engine, which annoys many regular posters. Offloading this work to an AI would save many peoples' time and patience.

Would it? Or would those same regular posters who read new member questions and then complain about them just read the autoreplies and complain about those instead? I could absolutely see them devolving into "nitpick the AI response" threads.

Probably yes, but it would be more fun than answering very common questions such as "I have a resistor with: brown, black, red and gold bands. What value is it?", "I have a potential divider, how do I calculate the output voltage?", "How do I work out the gain of an inverting op-amp circuit?" etc.

[PlainName]

You know, no-one is forced to answer any question on here. If it's a tedious newbie 100^th- time-this-week one, just push on to the next thread and forget about it.

[SiliconWizard]

You know, no-one is forced to answer any question on here. If it's a tedious newbie 100^th- time-this-week one, just push on to the next thread and forget about it.

Indeed. Either you feel like the topic is still interesting and you have something to say, or you can just move on to other threads. No big deal, really. Sure, some users could find it "rude" not to get any answer at all, but that's just how a forum works. No one is entitled to a reply.

It's not just "newbies" either - now a significant fraction of this kind of threads are from students (who often look like they haven't attended their courses) or people manually feeding AI for a few bucks (yes, that's a job).

[Randy222]

When could a bot actually reason that the child molester is an actual bad guy? In general responses the bot agrees, but due to the ruleset given to it there's no actions the bot would do, which for real humans would be reasonable actions.

All AI is very biased, simply because the rules are set by humans. It's the whole crux of the ML phase of the AI chain.

ML/AI needs to be purposely biased as a starting point. If I tell all the bots to prefer Socialism/Marxism/Communism over Capitalism, then bots will always have that preference in their "reasoning".

It brings new meaning to Battle Bots. Only the strongest or most capable bot will win.

[RoGeorge]

Is my cat not conscious?

Absolutely yes.  All animals are conscious.  Plants are conscious, too, in their own way, but less conscious than animals.  IMHO, consciousness is the ability to predict the world they live in.

[Andy Chee]

When could a bot actually reason that the child molester is an actual bad guy?

There are some humans that believe Pizzagate is real.

Failure to learn is not an exclusive problem with machine learning, it happens with humans too.

[Berni]

The human brain might operate in a pretty similar way to an modern AI, even if the way it does is is very different.

All living creatures that exhibit any semblance of intelligence learn their behavior over time. It is likely an evolutionary optimization that works well. Maybe the brain is too complicated to encode the exact blueprints for in the DNA, so it was easier to create a brain that figures out the exact wiring on its own as it develops (using environmental stimulus and simple chemical rewards to motivate it). Another evolutionary advantage is that an life form with a brain capable of learning can adapt to a changing environment much quicker. If it does something that hurts it stops doing that, rather than wait for thousands of generations of evolution to optimize that harmful behavior away.

Brains are essentially a biological implementation of statistics calculation. They find the good and bad patterns in all the input they are provided, generating output signals that are most likely to result in a good feedback (chemical reward). Human brains are just a particularly well optimized variant of all the biological brain iterations in the millions of years. We happened to get particularly good at knowledge transfer between individuals and this was a powerful multiplier to allow a single individual build on the experience of many many others, giving him the equivalent experience as if he lived and learned for thousands of years. Other animals have to mostly start over at birth.

All this is exactly what AI does. It is a pile of statistics calculation that is optimized to seek the highest possible reward by taking in input and producing output that has the best chance of getting rewarded.

You could build the "perfect LLM" by simply taking all human written text and calculating the probability of the next word for every possible combination of words that come before it. Problem is that there are so many combinations of words that by the time you get to a context length of 4 or 5 you run out of memory to store the statistics results. Modern LLMs can do context lengths of 1000000 or more, they do this by using machine learning to approximate these statistic results on the fly (much smaller and less computation). Where the machine learning model gets the approximation wrong is where hallucinations happen. But the bigger the model is, the more able it is to get the approximation close to spot on.

That being said even if you have such a fully accurate "perfect LLM", the text you used to train it is made by humans and every human has a bias. So the LLM will pick up a bias that is roughly the average of all the text. If you were you train a LLM on mostly text in Arabic languages you will get a very different bias as compared to training it on English text. What we consider as being morally correct is specific to our culture. There is no pure untainted bias free training set in existence.

[SparkyFX]

Too many new members ask simple questions which are easily answered by a search engine, which annoys many regular posters. Offloading this work to an AI would save many peoples' time and patience.

Everyone was a newb at some point (be it netiquette, the topic or the internet in general). It is a development and some people forgot that they went through that themselves :-). Newbs that obviously did not use search properly also did not come across why that is being seen as bad form and why it is better to properly use search first or somehow utilizing the resources that are already there before asking the question.

AI also does not teach good prompting until being specifically asked for being that guide.

There is a bootstrapping problem in both approaches. So coming back to the question are forums dying out, it depends on what experience a new user makes.

[paulca]

Is my cat not conscious?

Absolutely yes.  All animals are conscious.  Plants are conscious, too, in their own way, but less conscious than animals.  IMHO, consciousness is the ability to predict the world they live in.

To be aware of the environment enough to be present in it.  It requires the understanding of "self" and "environment" to distinguish between them.  Probably, IMHO the minimal form of consciousness.

Cats are far more advanced than that.  They do have significant reasoning.  Their perception is also quite distinct from our own.  Which is why they surprise us with the examples like given above.  The cat knows more about you than you do.  They don't have the complex memories and frontal cortext and "language" forming our thoughts.  They just work on the moment, mostly.  They can see and smell your intent before you have it.

[paulca]

I think the main damage to forums like this one is done by "Reddit".  It unfortunately took a lot of forum traffic.  However it's a toxic cesspool. Mostly because it is not "mono-topic" or "reduced audience" but open and with millions of active users.  "Topics" are very thin walled "rooms" with windows and open doors.

The deadlock on modernising forums like this one is not helping attract people when they have a choice between a "classical forum" or Reddit.

I expect this forum, sadly will age into non-existance with it's "defacto" members.  Maybe that's fair.  All good things...

[jonovid]

There have been resurgence in young people seeking physical media.
this may also include a return to a back to basics with an internet with less spam , AI or a time of more civilized forums.
hopefully

[Randy222]

with an internet with less spam ,

Abilene
Internet2
other

At some point "internet" must become a choice from Internets.

But the digital world is ultimately shaped in some way, usually based on some society standards, which is not uniform around the world.

So maybe one day we'll get the "Modern Western", "Asia", "EU" versions of internet, and you'll get to choose which one your ISP connects you to.

The US has it's 1st Amendment for free speech, which is slowly being shut down, but try and find a free speech place online, you cant, we cant just go off the rails here on eevb, there's no free speech anywhere.

I myself dont care if some fool spouts bad words and BS ideas, because it has zero traction. It may be some noise up front, like what's going on in the US with the commies trying to make a come back, but it's going nowhere fast.

Globalists is not a fringe thing, people need to be aware, otherwise the govt will run your life (spain, uk, n.korea, china, etc etc).

[PlainName]

I myself dont care if some fool spouts bad words and BS ideas, because it has zero traction.

Problem is some of it does have traction.

Classic explanation for calling it out is: if you shout "fire" in a cinema you'll be causing a lot of pain, possibly deaths. So free speech has limits. But... perhaps the way to counter it is to just allow any old shit and then eventually the norm will be to treat everything as probably bullshit. However, back to the fire: what if there really is a fire, someone shouts "FIRE!" and no-one believes them because the norm is now not to?

Seems to me the proper solution is not to be an absolutist and recognise there are shades of grey. But no-one in the argument is going to accept that - you gotta either 100% win or you've 100% lost, whatever it is.

[Andy Chee]

Seems to me the proper solution is not to be an absolutist and recognise there are shades of grey. But no-one in the argument is going to accept that - you gotta either 100% win or you've 100% lost, whatever it is.

There's a real and practical difference between arguing for the purpose of winning, and arguing for the purpose of seeking the truth.

The internet is full of the former, because the internet lacks accountability/oversight.

[EEVblog]

Seems to me the proper solution is not to be an absolutist and recognise there are shades of grey. But no-one in the argument is going to accept that - you gotta either 100% win or you've 100% lost, whatever it is.

There's a real and practical difference between arguing for the purpose of winning, and arguing for the purpose of seeking the truth.

Some people just like to argue for fun.



This forum works because it's social, not a questions and answers database.
The social aspect can also destroy forums if not kept in check, hence the delicate moderation dance we have to play.

[Randy222]

I myself dont care if some fool spouts bad words and BS ideas, because it has zero traction.

Problem is some of it does have traction.

Classic explanation for calling it out is: if you shout "fire" in a cinema you'll be causing a lot of pain, possibly deaths. So free speech has limits. But... perhaps the way to counter it is to just allow any old shit and then eventually the norm will be to treat everything as probably bullshit. However, back to the fire: what if there really is a fire, someone shouts "FIRE!" and no-one believes them because the norm is now not to?

Seems to me the proper solution is not to be an absolutist and recognise there are shades of grey. But no-one in the argument is going to accept that - you gotta either 100% win or you've 100% lost, whatever it is.

Why should a group of people suddenly believe the one yelling "fire"? Doesnt anyone in the group care to verify, which is validating the one as savior or fool?

It's not about the "norm", it's about people thinking for themselves w/o some pre-indoctrinated mindset. We dont ignore the one yelling "fire", we use our brain to validate it, then we hero or shun the person who yelled "fire".

[PlainName]

OK, for 'fire' try 'bleach injection would kill covid' or similar.

The fact is that most of us can't verify stuff because it is simply beyond our ken. And when it comes to fires, sticking around until you can feel the heat to prove it is real is generally something you won't get to try a second time.

[paulca]

OK, for 'fire' try 'bleach injection would kill covid' or similar.

The fact is that most of us can't verify stuff because it is simply beyond our ken. And when it comes to fires, sticking around until you can feel the heat to prove it is real is generally something you won't get to try a second time.

And absolutely perfect fuel for Meta.  "Roll up roll up, place your bets, fight about it in the comments"

[Zero999]

OK, for 'fire' try 'bleach injection would kill covid' or similar.

The fact is that most of us can't verify stuff because it is simply beyond our ken. And when it comes to fires, sticking around until you can feel the heat to prove it is real is generally something you won't get to try a second time.

Trump didn't suggest injecting bleach at all. He inarticulately rambled about the idea of investigating disinfectants in a press conference with his advisers and journalists misquoted him, but it's irrelevant. If someone shouts fire, then most people will simply evacuate the building, not only because they're not going to hand around and find out, but because no one would think someone would lie about something like that.

[abeyer]

Trump didn't suggest injecting bleach at all. He inarticulately rambled about the idea of investigating disinfectants in a press conference with his advisers and journalists misquoted him, but it's irrelevant.

That's not even an accurate representation of what your own source says... it quotes/paraphrases his statement as "and I see the disinfectant where it knocks it out in one minute. Is there a way we could do something like that by injection inside or almost a cleaning because you can see, it gets in the lungs." I don't see that as anything but questioning if you can inject disinfectant. He corrects himself later, and they say that too, but it's pretty clear what he said despite people trying to spin it one way or the other.

But to the original point, and also to why this is a problem... it's true that most people don't have the expertise to evaluate claims they get presented with, so even more reason why a public figure shouldn't be spitballing dumb ideas in front of a media audience: People end up relying on "experts" whose expertise they can't/don't evaluate to evaluate the claims for them. Media publishes whatever gets views and ad dollars, which often means whatever is just controversial enough to stir up arguments while not being beyond a line that turns away advertisers. This is like the perfect storm of bad situations for a prominent public figure to make random health conjectures.

[Zero999]

Trump didn't suggest injecting bleach at all. He inarticulately rambled about the idea of investigating disinfectants in a press conference with his advisers and journalists misquoted him, but it's irrelevant.

That's not even an accurate representation of what your own source says... it quotes/paraphrases his statement as "and I see the disinfectant where it knocks it out in one minute. Is there a way we could do something like that by injection inside or almost a cleaning because you can see, it gets in the lungs." I don't see that as anything but questioning if you can inject disinfectant. He corrects himself later, and they say that too, but it's pretty clear what he said despite people trying to spin it one way or the other.

If you read the entire conversation he had with them, you'll find he didn't suggest anyone inject themselves with disinfectant.

But to the original point, and also to why this is a problem... it's true that most people don't have the expertise to evaluate claims they get presented with, so even more reason why a public figure shouldn't be spitballing dumb ideas in front of a media audience: People end up relying on "experts" whose expertise they can't/don't evaluate to evaluate the claims for them. Media publishes whatever gets views and ad dollars, which often means whatever is just controversial enough to stir up arguments while not being beyond a line that turns away advertisers. This is like the perfect storm of bad situations for a prominent public figure to make random health conjectures.

It was the wrong platform for Trump to have had that conversation, which should have taken place in private, but the journalists are equally to blame for their handling of it.

[PlainName]

If you read the entire conversation he had with them, you'll find he didn't suggest anyone inject themselves with disinfectant.

If you read my comment properly you'll note I didn't attribute that phrase to anyone, never mind Trump.

I am sure some very public and divisive figure may have said 'fire' once, and yet that's not picked up to argue about 

[EEVblog]

The Trump politics stops now.

[u666sa]

And that's the point: where is everybody?
On social media?
On anti-social media?
YT? TikTok?

Telegram and Discord. Because you can say pretty much anything without violating TOS.

Ai though, can not replace expert's advice. Because Ai uses forums for its answers. Guess who writes in those forums? There are actually communities, old communities, whose model of operation requires earning credit from participation and sharing, and often these communities have tests for newbies, which must be taken right after registration to be able to post and view hidden content. Such communities work just fine. Because their members are always motivated to participate in discussions, not only by writing and sharing with accumulated experience, but also by looking at schematics and solving problems -- they earn credit every time they provide useful advice.

Ai isn't smart, because advice it takes doesn't even come from these forums, which are hard to come by via a simple google search. Ai uses mainstream "amateurish" forums, which are free to index and are easily accessible. Let's take this real life question:

PROBLEM: Compal LA-M401P laptop turns on and works only with battery, no 19 volts. Problem TPS65994 doesn't bring up 19V though type-c, battery does not charge. Type-c is only 5V. Charge battery with PSU and laptop works from battery. In the beginning DT22 and CC2 had low resistance. Replaced TPS65994 and low resistance is gone. The way I understand it TPS65994 is controlled by UT2 and UT2 Intel JHL9040R Thunderbolt which takes firmware from W25Q80EWZPIG.
AI: 1. Thunderbolt power rail missing. 2. Thunderbolt reset not released. 3. SPI flash not powered. 4. Corrupter Thunderbolt firmware.
EXPERT_1: TPS65994 what letters you have after numbers? They are all different.
EXPERT_2: Don't look for differences in datasheet, buy and install exactly the same.

Ai obviously did not solve the problem, nor was it even close to solving it. Instead experts with their years of experience solved the problem right away.

Such forums have great value in them, and participants often professionals, motivation for participation are credits which are used to get material, and also community which can provide solid solutions to everyday problems. I am a member of several such forums and I haven't even noticed a slowdown. On the other hand, other forums did slow down, and looking back at it in retrospect through number of replies gotten it had happened sometime around 2023.

What so special about 2023? Ai became mainstream, not experimental. So your assessment about Ai being the culprit in driving this change is spot on. 

[PlainName]

they earn credit every time they provide useful advice

Not that simple. It prompts users who don't actually know to butt in anyway to get the credit. On a small scale we sometimes see that here, where some new person with blast a bunch of useless posts around just so they're not zero-posters and can send PMs and drop stuff to sell.