Opinions on AI

[Neo_Morpheus]

I will admit that I am conflicted with AI.

As a tool, I have to admit, it does makes certain things easier and in others, it is straight up allowing me to do tasks that I simply dont know how, like programming.

But on the other hand, the power consumption, high prices and other things are very negative.

And of course, we cannot ignore the fact that humans are behind this and when humans and money are in the same sentence, we know very well that it will be abused and used in very nefarious ways.

 

[thesmokingman]

Yup. It's copyright infringement to resell another artist's work, but the colossal data trawl (undertaken by third parties and bought by the AI giants) has evaded every fucking trademark office. These AI companies paid nobody to appropriate their IP. Now they're complaining someone else is stealing their work.

Their techbro entitlement is astonishingly offensive.

You guys are ironic and kind of lol worthy that a mod would take this stance. A lot of AI used trademark and copyrighted stuff early on and it was sort of a wild west until the law and regulations caught up. The question of what, where, and how much is a fight in the courts yet to be finished. By posting the fact that DeepSeek is actually a DeepFake does in no way mean I am taking a side on trademark or copyright. You guys are the comical ones, turning a cheek one way to applaud a bigger thief in your own words, LMAO. Irnoically Open AI which I'm not a fan of, is one of the few that has started to license their use of data.

DeepSeek pretended to be a new model made on the cheap. That has proven to be all a lie. Even the starting amount of Nvidia gpus for training is a lie. Not disclosing the tens of thousands of gpus they got circumventing embargos was another hint. We know now they got to where they are by distillation, basically running the logic question thru Open AI's model then copying the answer. No wonder huh?

Instead of talking about the facts yall go into this delusional, but they stole first. Lmao, I'm not a Open AI supporter so I'm well aware of what transpired before, but as I wrote that's a separate issue to what DeepSeek has done. And yall are ignoring that it really is a politcal attack on the US and some kind of warning or play by Xi to counter the chip sanctions which is a whole other can of worms.

 

[TumbleGeorge]

DeepSeek pretended to be a new model made on the cheap. That has proven to be all a lie

What, when? Because nOpen "AI" cry this statement public in websites isn't proof, isn't evidence.. To this moment for me is just crying of loser in competition.

 

[thesmokingman]

What, when? Because nOpen "AI" cry this statement public isn't prove.

Dude I can never understand your posts, wtf did that mean?

 

[Aleksandar_038]

You guys are ironic and kind of lol worthy that a mod would take this stance. A lot of AI used trademark and copyrighted stuff early on and it was sort of a wild west until the law and regulations caught up. The question of what, where, and how much is a fight in the courts yet to be finished. By posting the fact that DeepSeek is actually a DeepFake does in no way mean I am taking a side on trademark or copyright. You guys are the comical ones, turning a cheek one way to applaud a bigger thief in your own words, LMAO. Irnoically Open AI which I'm not a fan of, is one of the few that has started to license their use of data.

DeepSeek pretended to be a new model made on the cheap. That has proven to be all a lie. Even the starting amount of Nvidia gpus for training is a lie. Not disclosing the tens of thousands of gpus they got circumventing embargos was another hint. We know now they got to where they are by distillation, basically running the logic question thru Open AI's model then copying the answer. No wonder huh?

Instead of talking about the facts yall go into this delusional, but they stole first. Lmao, I'm not a Open AI supporter so I'm well aware of what transpired before, but as I wrote that's a separate issue to what DeepSeek has done. And yall are ignoring that it really is a politcal attack on the US and some kind of warning or play by Xi to counter the chip sanctions which is a whole other can of worms.

Proven, where? By whom?


Hilarious...

 

[thesmokingman]

Proven, where? By whom?


Hilarious...

Do you homework?

 

[Assimilator]

Their techbro entitlement is astonishingly offensive.

Their existence is astonishingly offensive.

 

[TumbleGeorge]

Do you homework?

The semianalysis cooking your "homework". I read their gpt generated bs. Huh.

 

[the54thvoid]

You guys are ironic and kind of lol worthy that a mod would take this stance. A lot of AI used trademark and copyrighted stuff early on and it was sort of a wild west until the law and regulations caught up. The question of what, where, and how much is a fight in the courts yet to be finished. By posting the fact that DeepSeek is actually a DeepFake does in no way mean I am taking a side on trademark or copyright. You guys are the comical ones, turning a cheek one way to applaud a bigger thief in your own words, LMAO. Irnoically Open AI which I'm not a fan of, is one of the few that has started to license their use of data.

DeepSeek pretended to be a new model made on the cheap. That has proven to be all a lie. Even the starting amount of Nvidia gpus for training is a lie. Not disclosing the tens of thousands of gpus they got circumventing embargos was another hint. We know now they got to where they are by distillation, basically running the logic question thru Open AI's model then copying the answer. No wonder huh?

Instead of talking about the facts yall go into this delusional, but they stole first. Lmao, I'm not a Open AI supporter so I'm well aware of what transpired before, but as I wrote that's a separate issue to what DeepSeek has done. And yall are ignoring that it really is a politcal attack on the US and some kind of warning or play by Xi to counter the chip sanctions which is a whole other can of worms.

I posted as a member, not a mod. I rarely post these days, but for this it's a fair shout. You don't share that opinion, that's also fair.

I've never supported generative AI that's used other people's copyright or IP. That's why I'm indifferent to this. And frankly, I don't give a brown stink what you think about it.

 

[Shihab]

Even the starting amount of Nvidia gpus for training is a lie. Not disclosing the tens of thousands of gpus they got circumventing embargos was another hint.

Is this the future? When did the investigation publish its findings (or get things rolling in the first place)?
I recall these things taking months, if not years. But I guess if it's China, suspicions are always assumed true, eh?

Nothing was proven about their training costs. But as one of their own people said: you can test it it yourself. You'd need your own data, of course, but that wasn't a major part of the equation to begin with. And if it was, I can name several companies whose logos resemble the sphincter of one's posterior who are much more worthy of being questioned about the "true" cost of their training materials.

Smearing competition as "political attack" is the lol worthy part. Especially since the true, relevant political attack started before that when a certain party tried to kill said competition with underhanded regulations. Another thing that does go to Deepseek's credit.

 

[Assimilator]

I posted as a member, not a mod. I rarely post these days, but for this it's a fair shout. You don't share that opinion, that's also fair.

I've never supported generative AI that's used other people's copyright or IP. That's why I'm indifferent to this. And frankly, I don't give a brown stink what you think about it.

Let's be honest here, whether it's OpenAI or DeepSeek or Llama or Gemini or whatever, they're all garbage products created by garbage human beings for garbage reasons. And yet just like GPUs, people feel the need to pick one hill or another and die on it. It's the primary reason that humanity needs a strong AI to rule us, because most people are such easily manipulated fools.

 

[Shihab]

humanity needs a strong AI to rule us, because most people are such easily manipulated fools.

Boy! Do I have some bad new for you, my friend!

 

[InVasMani]

The thing about copyrights and patents is people born a 100 years ago were in a much better deposition than today to come up with new and unique copyrights and patents because the low hanging fruit has more and more already been plucked from the tree. We're born as copies and taught to copy at birth and naturally perform inference from the time of inception at the DNA level and when born that alongside observations around us.

Ultimately those systems protect people with power and clout more than your average internet user or person within society. It's more for corporations and people working with or for them than individuals in the end is the honest reality.

The way AI was trained is the heart of the issue even though it's not so different from what human's do much of the time anyway. Like if I wanted to do music more like Metallica I'd probably lean more heavily towards emulating the same types of instruments and effects. I mean look at the industry as a whole everyone is mimicking everyone pretty much deliberately or otherwise.

 

[Caring1]

You guys are ironic and kind of lol worthy that a mod would take this stance. A lot of AI used trademark and copyrighted stuff early on and it was sort of a wild west until the law and regulations caught up. The question of what, where, and how much is a fight in the courts yet to be finished. By posting the fact that DeepSeek is actually a DeepFake does in no way mean I am taking a side on trademark or copyright. You guys are the comical ones, turning a cheek one way to applaud a bigger thief in your own words, LMAO. Irnoically Open AI which I'm not a fan of, is one of the few that has started to license their use of data.

DeepSeek pretended to be a new model made on the cheap. That has proven to be all a lie. Even the starting amount of Nvidia gpus for training is a lie. Not disclosing the tens of thousands of gpus they got circumventing embargos was another hint. We know now they got to where they are by distillation, basically running the logic question thru Open AI's model then copying the answer. No wonder huh?

Instead of talking about the facts yall go into this delusional, but they stole first. Lmao, I'm not a Open AI supporter so I'm well aware of what transpired before, but as I wrote that's a separate issue to what DeepSeek has done. And yall are ignoring that it really is a politcal attack on the US and some kind of warning or play by Xi to counter the chip sanctions which is a whole other can of worms.

Are you under the mistaken assumption we condone the actions of either side in their petty power struggle?
This situation is hilariously like a burglar calling the person that took stolen goods from him, a thief. Neither party is innocent.

 

[Aleksandar_038]

Do you homework?

I did.

There are zero proofs, only good old propaganda from OpenAI side.

And a bit of fear... For no reason, because Deepseek is same junk as ChatGPT, but it is still funny to watch.

 

[TPUnique]

It's all overhyped, a solution in search of a problem. A gigantic waste of money and energy consumption. Another dot com bubble waiting to burst. The sooner the better I say.

Can you sift through tens of thousands of scientific articles to make sense of them ?
Can you sift through millions of molecule combinations to come up with potential drug candidates ?
You sure can't. But maybe you think these aren't problems. The scientific community thinks otherwise.

 

[TumbleGeorge]

Can you sift through tens of thousands of scientific articles

No

to make sense of them ?

For LLM this part of task is impossible.

Can you sift through millions of molecule combinations to come up with potential drug candidates ?

No.
But many drug candidates despite their similar chemical structure and geometric shape, the molecules do not actually interact in a therapeutically useful way. So testing the old-fashioned way is necessary. In fact, it's not just necessary, it's still mandatory.

Let me add some things that you may have intentionally not asked. Could this screening work be done with computers, and in particular supercomputers of the existing generations? The answer is unequivocal: Yes!
Would computers have stopped developing if there were no crypto and AI bubbles and the Covid crisis? The answer is: No! Computers have been developing for decades before this and without these circumstances.

 

[Aleksandar_038]

Can you sift through tens of thousands of scientific articles to make sense of them ?
Can you sift through millions of molecule combinations to come up with potential drug candidates ?
You sure can't. But maybe you think these aren't problems. The scientific community thinks otherwise.

Actually, discovering new drugs doesn't have anything with siftimg through millions of molecules.

Some of best discoveries in last 30 yeara in pharmacy were due to misdirection or simply random. And those would not be possible without huge and expensive clinical studies.

So, there iz zero chance that LLM will change that.

That much I know, i have several biochemists in family...

 

[Shihab]

no one in their right mind would OSS the training set, and draw a bullseye on their backsides for every copyright lawyer from every sorry corner of this earth).

”Torrenting from a corporate laptop doesn’t feel right”: Meta emails unsealed

Meta’s alleged torrenting and seeding of pirated books complicates copyright case.

arstechnica.com

But, damn! 82TBs? I know a guy driven to suicide for pirating less than a terabyte!

 

[Veseleil]

”Torrenting from a corporate laptop doesn’t feel right”: Meta emails unsealed

Meta’s alleged torrenting and seeding of pirated books complicates copyright case.

arstechnica.com

But, damn! 82TBs? I know a guy driven to suicide for pirating less than a terabyte!

Those are rookie numbers.

 

[Dragam1337]

While ai can be helpful tools, make alot of stuff easier, it's obviously going to be exploited, and it already is big time... we are already at a point where you basically can't trust anything you see on the internet... and it's only going to get worse.

 

[cal5582]

the whole things a race to the bottom for people to outsource their thinking to another device and just consoom.

 

[AusWolf]

Let's be honest here, whether it's OpenAI or DeepSeek or Llama or Gemini or whatever, they're all garbage products created by garbage human beings for garbage reasons. And yet just like GPUs, people feel the need to pick one hill or another and die on it. It's the primary reason that humanity needs a strong AI to rule us, because most people are such easily manipulated fools.

Garbage reasons? What are they? I can't imagine any (legitimate or garbage) reason why I'd want a paragraph-long bland, generic answer to my question instead of a link pointing to a site where experts answer it.

 

[dragontamer5788]

Garbage reasons? What are they? I can't imagine any (legitimate or garbage) reason why I'd want a paragraph-long bland, generic answer to my question instead of a link pointing to a site where experts answer it.

All of these LLMs are deeply useful at extracting wealth from shareholders, as shareholders seem to throw more money at you when you say the word "AI".

100% legitimate "garbage reason". But its a reason and "getting more money" is the point of companies. And until the typical investor starts thinking about how and where their money is going, this pattern will continue as more and more investors throw money at stupid things (which itself causes more investors to throw more money at things because its fun to move in crowds / memes-stonk trades / etc. etc.).

 

[AusWolf]

All of these LLMs are deeply useful at extracting wealth from shareholders, as shareholders seem to throw more money at you when you say the word "AI".

100% legitimate "garbage reason". But its a reason and "getting more money" is the point of companies. And until the typical investor starts thinking about how and where their money is going, this pattern will continue as more and more investors throw money at stupid things (which itself causes more investors to throw more money at things because its fun to move in crowds / memes-stonk trades / etc. etc.).

I still don't see how that works. Mr Jensen steps onto stage, says "AI" and money magically appears? Why? How? What's the point of all this?

Things like new cars or phones are at least feeding an illusion that you're benefiting from them somehow (even though you're really not), so the common Joe will want it. But why would the common Joe want AI? Sure, there's no feature, but there's no illusion, either. Just buzzwords and thin air.