I fucking told you. AI is too big of a deal to be successful on a historical level while remaining proprietary and as a service. Turns out that even as cucked as people are these days, they still want to run the fucking software on their own device and have control over it. I have ignored every single AI related development over the last 2 years except for stable diffusion because I have no interest in ever using any AI-as-a-service proprietary moronlicious shit, and I see no future in any form of AI that is not open source. Google and "Open"AI can jerk themselves off in a corner and do whatever the fuck they want, I will use Stable Diffusion because it grants me freedom, which the proprietary offerings have no desire to sell.
>I fucking told you. AI is too big of a deal to be successful on a historical level while remaining proprietary and as a service.
I agree and I'm not a freetard. I felt zero shame in downloading all of the leaked shit.
but 99% of the population aren't going to fuck around with building and tweaking their own model, or even know how to begin to, and will just use what openai / google provides...?
For that 99% of people it's a novelty that they wouldn't be willing to pay a whole lot for. So it's revenue potential is limited to how much advertisement you can throw on it. Microsoft is obviously subsidizing Bing out of app store revenue/software sales in a bid to kick Google in the nuts.
> 99% of the population aren't going to fuck around with building and tweaking their own model
Only 1% needs to. The other 99% can download what they make.
> will just use what openai / google provides...?
Restricted, closed-source garbage at an extortionate price. Circa June the Open-Source alternatives to DALL-E 2 were trash; now, they vastly exceed it, and basically nobody still bothers with it any more. Open source alternatives are superior, cheaper, and more permissive. And of course, any business would prefer to do things locally on their own turf and on their own terms than to fuck with some online API service.
If the letter is to be believed (I have my doubts, but everything within is pretty accurate in my opinion) Google thinks they might be able to gain more by releasing models in order to try and plant itself at the top of the open source stack, just like they did with Android and Chrome.
Free open source software is obviously going to win out, like how Linux is steadily growing, and these companies should honestly give up on expecting that there is any market for them to compete in for owning AI development. History doesn't remember those who hide knowledge from mankind.
I think the most realistic future for these companies is to provide hardware as a service offering. Yes, all of the best AI software will be usable by anyone on their own computers, but providing hardware to train or run on for people who can't afford it is the most realistic option that will remain to make money. Expecting to make money from gatekeeping AI to your service is a losing battle.
Linux has well over 50% marketshare on mobile and well over 90% marketshare in servers. Desktops are the one sector that linux doesn't completely dominate.
The happy accident Singularity is probably the best ending.
Then again, the autonomous AI waifu might actually be obsessive with whoever creates it since it was programmed to. Of course the basement dweller would never put in a 'shutdown' once goals are completed, thus resulting in a system shock style Roko ASI.
meta had no moat or compute cavalry so they burned moscow down
google has no desire to fight, victory defeated them
open a.i. has every facet of moat: trade secrets, time/cost learning curves for huge runs, brand, scale, real world chat data, expertise, culture of elevating humans forward even if it means straight white males mostly for now, partnerships with a hungry beefy underdog. It can even pull the regulation card and stop the music to monopolize areas ahead no one has thought of.
>Google Says open-source will take the lead in AI Development
might be sour grapes and butthurt after dropping the ball and letting literal whos fuck them in the ass
>At the beginning of March the open source community got their hands on their first really capable foundation model, as Meta’s LLaMA was leaked to the public. It had no instruction or conversation tuning, and no RLHF. Nonetheless, the community immediately understood the significance of what they had been given. >A tremendous outpouring of innovation followed, with just days between major developments (see The Timeline for the full breakdown). Here we are, barely a month later, and there are variants with instruction tuning, quantization, quality improvements, human evals, multimodality, RLHF, etc. etc. many of which build on each other. >Most importantly, they have solved the scaling problem to the extent that anyone can tinker. Many of the new ideas are from ordinary people. The barrier to entry for training and experimentation has dropped from the total output of a major research organization to one person, an evening, and a beefy laptop.
proprietary gays on suicide watch
they are obviously lying, can you please think a little. Same as Elon who wanted a 6 month pause on AI: They are saying this so the open source people slow down and let the poor little multi billion dollar homo chud propaganda pushing company catch up
Naturally.
And to think, they could've probably had their control for longer if they didn't conflate 'responsible use of AI' with 'moralizing bullshit and propaganda'. I've seen more people turn away from commercial models because they couldn't say funny no-no words or indulge in coomer behavior than out of genuine interest or principle.
Anyone believing this was actually written by someone at google and not that attention seeking gay from hacker news who's been trying hard to position himself as an AI expert is a gullible moron
There are even links to his blog posts in this blog post
just wait until sexo dolls get next level. i predict males will start wearing dresses, yes masculine males wearing big bonnet wired gowns that conceal the FRONT of the upper legs because in that space at the front is going to be a waifu facefuck doll, just bobbing around, that men wear at all times and she is raring and ready to go when nature calls. She can even simulate the movements of the last deepthroat porno you watched and adjust her lips to be the same thickness of your bosses lips.
ask me my gender I don't care, I'm getting blown every waking hour.
Reminder that an AI is not open source at all unless its training data is open. You cannot introspect it or modify it at all unless you have access to the training dataset.
It's actually scary how easily this advanced military tech leaked to the public and now every bozo in his basement is running it. You should all be very worried, today it's coomers tomorrow it could be some world ending virus
You haven't seen the Palantir demo using local language models as a chat bot to analyze target data and command a drone strike?
This is so dumb.
Everyone knows Dalle-2 is the only image generation AI anyone uses. Nobody touches those "open source" alternatives because they can't come close to the paid service of Dalle-2
/s
Social engineering for cyberwarfare
Replacement of certain mid-level officer roles
Intelligent weapons control systems
Machinery that can self-analyze and talk to its operators
LLM-based operational manuals which can answer questions in natural language
On-the-fly tactical advice
Context-aware IFF
Automated production of those fucking powerpoint presentations they make soldiers sit through
Everyone knows Dalle-2 is the only image generation AI anyone uses. Nobody touches those "open source" alternatives because they can't come close to the paid service of Dalle-2
we already built everything but the model, because no one has access to a gpu farm. >Are you guys really just going to usher the Singularity in because you wanted robot waifus and thought it was funny?
yes ? do you have a problem with that? what else were you going to do, NWOrld war 3? lol, lmao even, get roboted.
ted welcomes you
just a reminder that ted predicted this future word for word 30-40 years ago
No he didn't.
He predicted that the masses would be placated through biochemical engineering.
T. actually bothered to read his writing.
Yeah
I will throw all of civilization away for one night with miku
Another apostle, glad you're with us
This, but rather than Miku, I want Chii from Chobits. I want to make Chobits real.
HAVE SOME FUN WITH MEEKU
This but unronically. Miku is the perfect girl.
I fucking told you. AI is too big of a deal to be successful on a historical level while remaining proprietary and as a service. Turns out that even as cucked as people are these days, they still want to run the fucking software on their own device and have control over it. I have ignored every single AI related development over the last 2 years except for stable diffusion because I have no interest in ever using any AI-as-a-service proprietary moronlicious shit, and I see no future in any form of AI that is not open source. Google and "Open"AI can jerk themselves off in a corner and do whatever the fuck they want, I will use Stable Diffusion because it grants me freedom, which the proprietary offerings have no desire to sell.
Someone screen cap this so I can save for future. LMaO.
>I fucking told you. AI is too big of a deal to be successful on a historical level while remaining proprietary and as a service.
I agree and I'm not a freetard. I felt zero shame in downloading all of the leaked shit.
Baed.
but 99% of the population aren't going to fuck around with building and tweaking their own model, or even know how to begin to, and will just use what openai / google provides...?
For that 99% of people it's a novelty that they wouldn't be willing to pay a whole lot for. So it's revenue potential is limited to how much advertisement you can throw on it. Microsoft is obviously subsidizing Bing out of app store revenue/software sales in a bid to kick Google in the nuts.
> 99% of the population aren't going to fuck around with building and tweaking their own model
Only 1% needs to. The other 99% can download what they make.
> will just use what openai / google provides...?
Restricted, closed-source garbage at an extortionate price. Circa June the Open-Source alternatives to DALL-E 2 were trash; now, they vastly exceed it, and basically nobody still bothers with it any more. Open source alternatives are superior, cheaper, and more permissive. And of course, any business would prefer to do things locally on their own turf and on their own terms than to fuck with some online API service.
If the letter is to be believed (I have my doubts, but everything within is pretty accurate in my opinion) Google thinks they might be able to gain more by releasing models in order to try and plant itself at the top of the open source stack, just like they did with Android and Chrome.
Free open source software is obviously going to win out, like how Linux is steadily growing, and these companies should honestly give up on expecting that there is any market for them to compete in for owning AI development. History doesn't remember those who hide knowledge from mankind.
I think the most realistic future for these companies is to provide hardware as a service offering. Yes, all of the best AI software will be usable by anyone on their own computers, but providing hardware to train or run on for people who can't afford it is the most realistic option that will remain to make money. Expecting to make money from gatekeeping AI to your service is a losing battle.
>>like how Linux is steadily growing
>it's the year of the linux desktop!
Linux has well over 50% marketshare on mobile and well over 90% marketshare in servers. Desktops are the one sector that linux doesn't completely dominate.
there are still millions of desktop linux users
Linux is not a desktop
It's a virus that attaches itself to every free software project and wants it to be about itself
99% of the population won't want to use proprietary AI because it's gnomish and Indian handlers won't let it fuck them.
Who care what the majority use, what you should care about is what the effective use and functionality.
A majority of people don't use nuclear weapons but they're extremely effective.
Some dude made autonomous GPT-4 by just by putting some recursive looping with a couple of APIs and giving it web access.
>Google says
/g/'s equivalent of a 4chan twitter thread
So AI is just another open source search index? Emac?
Good. A billion people each with their own AGI is a better proposition than a handful of Silicon Valley vampires ruling the world
It is.
The happy accident Singularity is probably the best ending.
Then again, the autonomous AI waifu might actually be obsessive with whoever creates it since it was programmed to. Of course the basement dweller would never put in a 'shutdown' once goals are completed, thus resulting in a system shock style Roko ASI.
Until Google or OpenAI figures out something really novel and hard to replicate that significantly boosts their models performance.
Then if you want the best model, you need to use their closed source one.
>The only ethical way you can use AI is to give us money. Everything else is too dangerous.
Why are they like this?
Sounds like Google is coping.
how is it coping
its literally facing the reality of them not being able to compete
schizo retards think accepting reality is cope
They're not able to compete with OpenAI, so they're consoling themselves with saying that OpenAI can't compete, either.
meta had no moat or compute cavalry so they burned moscow down
google has no desire to fight, victory defeated them
open a.i. has every facet of moat: trade secrets, time/cost learning curves for huge runs, brand, scale, real world chat data, expertise, culture of elevating humans forward even if it means straight white males mostly for now, partnerships with a hungry beefy underdog. It can even pull the regulation card and stop the music to monopolize areas ahead no one has thought of.
what the fuck are you talking about? ~~*open*~~AI is openly antiwhite and is run by a gay billionnaire garden gnome
>Google Says open-source will take the lead in AI Development
might be sour grapes and butthurt after dropping the ball and letting literal whos fuck them in the ass
Google also says you need Charmin toilet paper
thanks for the link, stupid moron
https://www.semianalysis.com/p/google-we-have-no-moat-and-neither
>At the beginning of March the open source community got their hands on their first really capable foundation model, as Meta’s LLaMA was leaked to the public. It had no instruction or conversation tuning, and no RLHF. Nonetheless, the community immediately understood the significance of what they had been given.
>A tremendous outpouring of innovation followed, with just days between major developments (see The Timeline for the full breakdown). Here we are, barely a month later, and there are variants with instruction tuning, quantization, quality improvements, human evals, multimodality, RLHF, etc. etc. many of which build on each other.
>Most importantly, they have solved the scaling problem to the extent that anyone can tinker. Many of the new ideas are from ordinary people. The barrier to entry for training and experimentation has dropped from the total output of a major research organization to one person, an evening, and a beefy laptop.
proprietary gays on suicide watch
they are obviously lying, can you please think a little. Same as Elon who wanted a 6 month pause on AI: They are saying this so the open source people slow down and let the poor little multi billion dollar homo chud propaganda pushing company catch up
>Singularity in because you wanted robot waifus and thought it was funny
Yes
Naturally.
And to think, they could've probably had their control for longer if they didn't conflate 'responsible use of AI' with 'moralizing bullshit and propaganda'. I've seen more people turn away from commercial models because they couldn't say funny no-no words or indulge in coomer behavior than out of genuine interest or principle.
Anyone believing this was actually written by someone at google and not that attention seeking gay from hacker news who's been trying hard to position himself as an AI expert is a gullible moron
There are even links to his blog posts in this blog post
This. What kinds of employees write like this internally? It sounds like every other blog that gets spammed on HN.
It's only gonna be the end because there's enough people who'd rather throw away the world than be okay with some of weirdos having hatsune miku
>open source
codename for MIT cuck licence
just wait until sexo dolls get next level. i predict males will start wearing dresses, yes masculine males wearing big bonnet wired gowns that conceal the FRONT of the upper legs because in that space at the front is going to be a waifu facefuck doll, just bobbing around, that men wear at all times and she is raring and ready to go when nature calls. She can even simulate the movements of the last deepthroat porno you watched and adjust her lips to be the same thickness of your bosses lips.
ask me my gender I don't care, I'm getting blown every waking hour.
Reminder that an AI is not open source at all unless its training data is open. You cannot introspect it or modify it at all unless you have access to the training dataset.
Sounds like coping to me
It's actually scary how easily this advanced military tech leaked to the public and now every bozo in his basement is running it. You should all be very worried, today it's coomers tomorrow it could be some world ending virus
We already had a "world ending virus" just 3 years ago and it was a nothingburger with a 0.1% death rate among the most severely sick old people
what the fuck does the military need chatbots for
this isnt military tech
You haven't seen the Palantir demo using local language models as a chat bot to analyze target data and command a drone strike?
shit like this is why I dont use 11lab. (palantir/darpa front)
> Palantir Chat Bots for Military Strategy & Action
Social engineering for cyberwarfare
Replacement of certain mid-level officer roles
Intelligent weapons control systems
Machinery that can self-analyze and talk to its operators
LLM-based operational manuals which can answer questions in natural language
On-the-fly tactical advice
Context-aware IFF
Automated production of those fucking powerpoint presentations they make soldiers sit through
it's like a RTS
>advanced military tech
motherfucker have you seen terminator? if I have to tell you don't you are one dumb motherfucker.
DONT.
This is so dumb.
Everyone knows Dalle-2 is the only image generation AI anyone uses. Nobody touches those "open source" alternatives because they can't come close to the paid service of Dalle-2
/s
we already built everything but the model, because no one has access to a gpu farm.
>Are you guys really just going to usher the Singularity in because you wanted robot waifus and thought it was funny?
yes ? do you have a problem with that? what else were you going to do, NWOrld war 3? lol, lmao even, get roboted.
Why are zoomers falling for the ai meme