I knew it
All they're doing is paying electricity bills for nothing
I bet the data they're collecting from people interacting with the AI is worthless too
and most people still aren't bothering to use it at all
 Nothing Ever Happens Shirt $21.68 |
 |
Nothing Ever Happens Shirt $21.68 |
the amount of money they'll save on cutting salaries/positions in 5-10 years will be well worth it
these clickbait titles are moronic
>in 5-10 years
...of r&d?
whos gonna pay for that?
the goyem will since most of these big tech companies are effectively subsided by the state
(cont)
they'll rather wait 20 and expect a fosshead will write one in hopes of being bestowed the privilege of becoming a wageslave
or miscroshit plans to shower closed ai with another 10B to pay for all the labeling
>any time now
>just 2 weeks
>a few months now!
>5-10 years
you AI gays are delusional
besides making cool pictures and sexting your waifu this AI stuff will never do anything impressive and this cringe fearmongering, telling lower level employees that "AI will replace you!!!" is cringe and fantasy world
I love ludittes. They are sooooo cute i just want to smooch them like a pet.
"Technology replaces workers" is the definition of ludditism, homosexual
Luddism is when you oppose the use of technology to replace workers.
I oppose the use of human workers to perform labor that could be done by a machine. It is unethical and wasteful for humans to perform menial, repetitive tasks.
>I oppose the use of human workers to perform labor that could be done by a machine
Anything that could be reasonbly automated, has been already. One thing humans are really good at is doing menial tasks in a extremely flexible and space efficient manner.
A lot of times it's simply not worth it to automate human labor, even if it's insanely repetitive physical work. A fairly simple conveyer belt, packaging, and loading system that only needs one worker to load the product into the hopper could replace 4-6 people, but something like that can cost half a million dollars or so, it's still cheaper to just pay some mexican women local minimum wage.
>Anything that could be reasonbly automated, has been already.
I have a feeling this has been said for decades, while we keep building new machines to automate more tasks.
And yet "strangely" we haven't 'ran out' of work, almost like humans are general automatons or something
Yeah, new fields open, new businesses appear, existing areas with little or no automation expand.
We're not at really at risk of running out of jobs, but there is a problem of specializing in a field or position to then find all the work getting automated.
in the west 30-50% of workforce are doing totally irrelevant jobs that could have been automated with existing technology
think about HR, general bureaucracy, lawyers, middle management, etc.
>oh noes what if irrelevant jobs are eliminated?!
I support replacing women with technology
>there's x amount of work to perform
You are a luddite. Technology is a force multiplier, you don't end up with less work, you end up with more output for a given labor pool
>opposition is when you don't automatically agree with me and/or are skeptical of any of my claims
telltale sign you're in an irrational, ideologically-driven group
>Metaverse won't go anywhere
>I love ludittes, they are soooo cute I just want to -ACK
that's nice zuck
is this a joke or something? The metaverse is gonna be huge, people just haven't caught on to it yet.
You'll be using dogecoin to buy virtual mcdonalds with your AI waifu in the metaverse in 2 weeks guaranteed (all the research points to this)
>JUMP HEAD FIRST INTO THE CHICKEN MACHINE AND GET GRINDED INTO GOYSLOP FOR PLUS SIZED BLACK TRANS QWEENS!
I dunno basedjak... maybe we should think this through. Perhaps not all technology is go-
>UHHHH... LUDDITE! FRICKING GET HIM! YOU CAN'T STOP PROGRESS CHUD!
FOMO is a mind virus, there's no point in explaining this shit. Same morons who hyped up 3D printers and NFTs.
yea, i was going to but i just can't be bothered anymore
there isn't a single known, meaningful case of AI replacing workers where the AI actually did a better job than the workers themselves
the anon the mentioned call centers is one of those examples
dealing with shitty phone AI is horrendous and trying to force the stupid AI to get you a real person so that you can fix a client's issue with some proprietary modem/router feels impossible as it wastes my fricking time
MAYBE AI could get somewhere but i think AI right now is where VR/AR is right now
it needs at least a decade or more for it to be able to do anything actually useful
for instance, VR porn is pretty great but VR isn't innovative enough to get most people onto it no matter how good the porn is because it's too immature; the same applies to AI
again, MAYBE AI can take over but it'll likely be so long that only your children's children will be able to benefit from it
>no matter how good the porn is
The porn sucks is the issue, what's good about porn is that there's an unlimited amount for practically any taste or fetish besides the most ridiculously niche and VR porn is like 20 videos of used up roastie plowed by tyrone plus some DIY shitty cgi-anime trash. Worthless.
>VR porn is pretty great but VR isn't innovative enough to get most people onto it no matter how good the porn is because it's too immature
Its the best it can be without brain implants. There are already hands free machines with vr
>the anon the mentioned call centers is one of those examples
>dealing with shitty phone AI is horrendous and trying to force the stupid AI to get you a real person so that you can fix a client's issue with some proprietary modem/router feels impossible as it wastes my fricking time
That is not a bug this is a future.
For many companies there is no sense for having customer service. This is why customer laws forcing serving on them exist. AI call center is token service to dodge laws. If no laws they wouldn't have any service at all.
Yea, very much feels like 3d printers or Crypto. Grifters who seem to think technology just improves linearly forever acting like it's world changing when in reality it's going to have a few interesting use cases but overall be fairly unimportant to society.
AI has been out in full force for over a year and has yet to really do anything except help Jeets put even more bugs in their code and help furries jerk off more easily.
this time it is different. since the average IQ of this site is around 90-100, people haven't been able to discover anything beyond the racism and coom filter. however, prompts that will benefit at most 1000 people around the world are also banned on chatgpt. (i won't say what they are) in my opinion, if the CIA had these filters installed before the bot was even released, this shows that LLMs have been used at the state level for many years and that they are experienced in this regard.
LLMs have been useful for coding a data set for me, my only use case so far.
>prompts that will benefit at most 1000 people around the world are also banned on chatgpt. (i won't say what they are) in my opinion, if the CIA had these filters installed before the bot was even released, this shows that LLMs have been used at the state level for many years and that they are experienced in this regard.
how does it feel to be a schizophrenic moron? frick off.
> types "how to kill the president" into chatgpt
> "Im sorry, i cant talk about that"
Damn, the CIA is hiding its true potential.
Even if the LLM could answer, do you think it would give you anything meaningful?
Half the time when i use this shit it just repeats whatever I said back at me with 0 regard for wether its correct or not.
LLM's are great for people who are easily manipulated
>Yea, very much feels like 3d printers
What's wrong with 3d printers?
they suck shti and everyone still buys their plastic trash from china instead of making it at home (bc 3d printers are even more of a b***h than normal printers) where it is made with the usual molding etc methods instead of 3d printing there either. they are a novelty toy for tech nerds to frick with, not some kind of industrial revolution paradigm shifter.
3D printers have been huge for engineers for prototyping, get a job
That's their point moron. The proponents were shilling it as a revolution, everyone will have one to print their plastic crap on demand, from the internet
Turned out, using one is so much fricking pain and so costly you need to be either really invested, or paid to do it
They're not difficult to use.
Then again literal paper printers have dedicated technicians and help lines and support warranties so yeah 3d printers are a high bar for your typical Joe 102 IQ
>That's their point moron. The proponents were shilling it as a revolution, everyone will have one to print their plastic crap on demand, from the internet
i havent bought one becasue i cant think of any plastic crap i actually want that i cant already just buy
prototyping what? dragon dildo shapes (post sanding)?
> 3d printers are even more of a b***h than normal printers
You've clearly never owned a HP printer
Nothing, but there was a period where they were a huge grift, billed as revolutionary, ect. They ended up being a neat tool for a lot of professionals and a fun toy for a lot of hobbyists, but your average person doesn't even remember they exist most of the time.
After the internet happened, people are really vulnerable to tech grifts, the FOMO on the next Internet has people buy into niche technologies with a few interesting use cases as if they're going to Change The Way We Live.
Crypto has substantially improved my life and allows me to make payments where f*at can and AI was great source of coom for a few months but sadly local models are way too limited now and cloudshit is useless by definition.
Dude, I'm using gpt-4 to write functional code for me. It's far from useless.
>this one moron was wrong about something so anyone who claims something is overhyped is also wrong forever
3D print anything good lately?
I 3d print once a week or more but that's because I'm a /tg/ user
consumer level 3D printers are a niche and this is unlikely to change, but additive manufacturing is now an integral part of most industries
You can google for "3D print services mjf" and have a complete component shipped in 3 days, 20 years ago prototyping like this was unthinkable
>this one moron was wrong about something
>AI will never become viable.
>AI will never be able to classify images.
>AI will never be able to create poetry.
>AI will never be able to create music.
>AI will never be able to create art.
>AI will never be able to create working code.
>Open source AI will never be able to compete with proprietary models.
>AI will never be profitable. (You are here)
>AI will never replace my job.
Shall I continue the list or have you homosexuals embarrassed yourselves enough at this point?
I've been working in this field for close to two decades now, if I got a cent for every time some cretin predicts what AI will never be able to do™ only to be completely wrong and move on to the next pathetic cope attempt I'd have enough money to buyout the entire S&P500 and still have money left to spend on a centuries worth of hookers and coke.
>>AI will never replace my job.
Anon, most jobs could easily be replaced without the use of AI because our economy is mostly fake. Don't you remember during covid when half the country got a free vacation and nothing happened?
>Pic related.
>>>Open source AI will never be able to compete with proprietary models.
which ones do?
It still can't write workable code though.
>>>this one moron was wrong about something
>>AI will never become viable.
It won't
>>AI will never be able to classify images.
It can't
>>AI will never be able to create poetry.
It can't
>>AI will never be able to create music.
It can't
>>AI will never be able to create art.
It can't
>>AI will never be able to create working code.
It can't
>>Open source AI will never be able to compete with proprietary models.
those are bad too
>>AI will never be profitable. (You are here)
It won't
>>AI will never replace my job.
It won't
So I take it you are confident in your ability to discern a poem generated by an AI from one made by a person?
>>>AI will never be able to classify images.
>It can't
are you brain damaged?
>>AI will never replace my job.
My job is already just a human failsafe in case the system breaks and the machine does something wrong. 100 years ago it took about 600 times the labor to do what I do now. It's a bit like that Australian bit about the boat that the front fell off and how they have strict safety requirements and minimum crew requirements of 'one I suppose'.
The people who AI will "never replace" (meaning, within their working life, 50 years or so) are
>at the absolute pinnacle of skill and insight, the kind of people who could develop new AI systems for their field if it came to it
>in-person customer service which cannot be replaced as such but is already almost extinct
>police, we're a long way off from people just accepting robocop
>physical jobs like construction, plumbing, electricical. trades. we're more than 50 years out from an actual robot coming to your house to fix your leaking shitter, or 'house printers' or whatever the frick
>upper management who will defend to their last dying breath their right to get paid to do nothing
and not much the frick else. If your job is sitting on the computer all day, it's over for you.
I think we will get automated construction for most of the work within 50 years. Most repair work will remain but then with a human working with some worker robots. As we built better buildings with service bots in mind that repair work will slowly shrink.
I also think skill and insight jobs will go before construction work. Doctors are already out performed by simple flow charts. Programs that write program will take over coding jobs. Management jobs will be killed off by competition of companies with more AI at more levels.
The customer service is already a mixed bag most customer service can be done by Indian call centers are rarely move up to a while person in the USA.
Once we get combat robots police bots will follow up a bit later. Cops are generally the most expensive budget item for a city and replacing police that frick up and bring down lawsuits on the city will be replaced by modified combat bots, likely working with a human operator either remotely or as a squad leader.
Also teachers. AI can’t reliably tardwrangle uppity teens into becoming productive members of the society.
If the parents are unemployed, they can homeschool their kids, which in practice means supervising them while they watch pre-recorded lessons and ask ChatGPT to grade their assignments.
Homeschooling will lead to bias and stupid kids. And stupid kids will lead to stupid decisions, and ultimately in the collapse of democracy, since said stupid kids will also be stupid voters.
>Homeschooling will lead to bias
You think that government-run and corporation-run schools don't have biases?
Kids of any background will be able to access the best teaching materials in the world, due to the falling cost of AI-produced and AI-assisted lessons.
All that is required from governments is that they run a quality control process to make sure that these educational resources and AI teachers really do teach kids the minimum amount needed to pass standardized tests.
>You think that government-run and corporation-run schools don't have biases?
Depends on the country of course. But at least the trained educators have standards that they need to pass.
And I wasn't saying that AI wouldn't be beneficial in education also. It's the tardwrangling part where it's lacking.
Yeah my company does MRIs of patients for orthopedic surgery and we 3D print custom made cutting guides for the surgeon to use in joint replacement surgery. Our sales reps bring them in along with the instrumentation and implants.
It's pretty sick.
>Metaverse isn't a thing a company builds. It's the next chapter of the internet over-AAACCCKK
I can do this all day.
>pic
nice bait
> Selloff incoming
he wasn't wrong though. sales and marketing moved from offline to online but ultimately didn't increase in any significant way as a result of the internet. you're also underestimating the impact of the fax machine on the corporate world. paul krugman generally suck but he was right about this. the telegraph was the last major revolution in communication. everything since then has been a slight improvement of the telegraph.
This, all it did was transform how we do commerce, get food, watch tv, moved office jobs into work-from-home, enabled consumer electronics companies to become the biggest in the world, created a trillion dollar commodity out of thin air called cryptocurrency, changed software delivery forever, reinvented communication, point of sale, and networking, skyrocketed gaming revenue to vastly surpass film, upset the taxi industry, revolutionized marketing, low latency stock trading and forex, soared demand for semiconductors and drove hardware innovation. It's basically just the telegraph.
He is unironically correct
And you are seething for whatever reason
The major difference is that AI plays into the idea that "someday soon human labor will be obsolete".
I get that the tech is interesting and novel but this is an such an absurd idea only made more absurd by the fact that people have been predicting this sort of thing for generations.
Thousands of years ago before computers were a glint in your ancestors eyes fears and hopes regarding human labor ending have existed. BUT THIS TIME ITS FOR REAL AND ITS GONNA HAPPEN IN THE NEXT FEW YEARS.
You are a moron, congrats go tell your mom and dad.
>The major difference is that AI plays into the idea that "someday soon human labor will be obsolete".
Do you think that AI will never be able to do the average intellectual task like writing code or business reports?
If you think it could do that eventually, but not soon, what do you think the barriers are between now and then?
Or do you think that the market will generate lots of new jobs for people with below-average intellect?
I think that when the dust settles and we can actually see what’s going on AI will be a lot less world shattering than people are anticipating.
In fact, life will probably go on as normal, evolving progressively as new tech slightly augments our regular lives but likely never brings ASI or super abundance or anything remotely close to the various tech utopia / distopias AI grifters are trying to sell us on being right around the corner, certainly not within our lifetime
That sounds like you're basing your conclusion on vibes, or by analogy to previous technological advances, rather than building a model of the dynamics from first principles.
If you accept that the tech gets better over time, why shouldn't that create an exponential feedback loop, where this year's tools help produce the next year's tools?
That's been happening in tech for the past 50 years or so, which means it would be surprising if it stopped suddenly.
Unfortunately humans aren't good at keeping track of exponential growth in our heads, as we are tempted to look backwards and see linear growth, and assume that the path ahead will have a similar amount of linear growth.
If an AI could do 1% of your job last year, and 2% this year, that doesn't mean there's 98 more years until it replaces you, it means means there is just 6 years of exponential growth.
You're right that human civilization has generally adapted to the new tools so far, but what's different now is that the tools we are making are increasingly able to operate by themselves without any human needing to use them.
This means that we're starting to transition from a civilization where humans use tools to one where human capabilities are unable to add any value.
Fortunately the value will still be created, as long as there are consumers to take part in the market economy, which means that money needs to be taxed from the producers and paid directly to consumers.
where i live ai has replaced some bank and isp call center workers. they don't connect you to a real person anymore unless you insist it.
those were make-work-jobs to make women feel they're useful people.
The only time I had an appointment with my bank the prostitute had to google the stuff while I was sitting there and told me stuff that didn't really answere my questions.
This shit has been around for quite some time, it just didn't have the "AI" meme attached. Your AI helpdesk sounds way cooler than "support bot thrown together in javascript".
Sad and delusional, will replace most white collar jobs in a few years, already destroyed the copywriting market, soon will cut most of the design and survelliancd market.
It's new tech, It's fine to be unadopted for 5/10 years and be unprofitable, If you are not scared by the speed of ai adoption and how it's useful in litterally any field you are,again, delusional.
>already destroyed the copywriting market
NOOOO NOT THE GRIFT MARKET FOR BRAINLETS GURUS PROMISED ME I'D BECOME A MILLIONARE COPYWRITING
>Sad and delusional, will replace most white collar jobs in a few years, already destroyed the copywriting market
can't you tell when something has been written by chat gpt? i can. it stands out a mile.
>posts total bullshit to bait-mine responses, and harvest information for free
I'm getting too old for this.
Based neoluddite. Woah dude AI that takes bullion of man hours to learn how to do one specific thing. Oh it doesn't understand program syntax? Oh it's trying to use variables that are out of scope? Oh it's calling a function that it didn't even define? Wow, I'm worried for all the Indians this is going to replace.
Is it worth 5-10 years of losing money on AI when they could be just doing the R&D internally?
Many companies leveraging AI to save money right now are doing as companies like OpenAI lose money on running the systems. If they can't bring costs under control, prices will have to go up and then we will find out if AI actually saves money over having people do the work instead.
Folding@home was first released in 2000.
im born in 88 Black person
my first hard drive was a phone book sized codebook for an atari
frickign hell
i didnt uderstand shit from frick with i was doing
but my first language is actually basic
Lol look at this guy who started with a basic language, I started with a complex one
>dont turn off your "console" for 3 weeks so that your code doesnt disappear
lil brother
kek
>solder pads on your k6 to overclock
it was an era on had to experience
no, sorry
it was a jumper on the mobo, kek
>i didnt uderstand shit from frick with i was doing
Somethings don't change.
when you bash your head against the wall something always gives.
until now it was the wall
it doesn't matter if openai loses money, they'll just get more funding from microsoft who practically owns them anyways
>Is it worth 5-10 years of losing money on AI when they could be just doing the R&D internally?
they are doing the r&d internally, and they're using free labor (you and your prooompting) to train their models. everything you put in and everything that comes out is recorded and analysed. they're not letting you make coom for charity bro.
I don't prompt.
They cuck hosted models so much all the spicy proooopting is happening locally
>vaccine die-off in two weeks
>AI lay-off two weeks after that
suits don't think that far ahead, it's all about end of year fiscal report and fat bonuses for shareholders
They dumped so much money into AI because it's "the next big thing" and investors want to make sure they don't miss out
>nobody thinks about long term
tells more about you than anyone else
sure, there are people like you mentioned, but do you think these big social systems could be in place for long time if every manager is a parasitic butthole?
pretty much yeah, you'll find more people willing to do a good and lasting job among the grunts rather than in upper managment
when a manager decides to buy a new lathe he's not thinking about long term, on paper he sees that:
a lathe can make 100k a year
a new lathe cost will be distributed over 10 years for 40k/yr, meaning that if we buy a new one now, in the year of our lord 2024 he'll have a net 60k increase in revenue
The fact that the company grows long term is just a side effect, it's all about making the most money as soon as possible
>it's all about making the most money as soon as possible
Inb4 Microsoft silently approved pirating Windows, just to keep the monopoly for decades.
Microsoft can play the long game. They have all the resources, and as one anon itt said government will subvent them, as they are major force in global technological/intelligence sphere.
>the amount of money they'll save on cutting salaries/positions in 5-10 years will be well worth it
Companies won't spend that money though. You think companies aren't required to make all of the money possible right now. They are.
What you are describing will be relegated purely to start-ups. Because big companies, the ones ironically with the money to make this change, are shackled to making maximum money right now, which AI isn't going to do without a 5-10 year investment.
Cope.
Interesting. TuringBot's shortest reasonably good regression predicts total leading node death by 2032:
-6.35441+(238.643/(x-1994.61))
>leading node death by 2032
That's not a surprising result, given that AGI will occur the year before.
https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/
Comment from your link:
>3 years ago, TSMC provided the first N5 ("5nm") iPhone chip, that had a transistor density of 134 million transistor per mm^2. The first N3 iPhone launched after 3 years and the transistor density increased only 36.5%, or 11% per year.
>That's a lot less than the twice the transistors every two years of Moore's Law. One reason is because nodes usually came every two years, but N2 is also expected to come after 3 years. Another is that the nodes usually increased density by 100%, but recently they've been increasing density by 70%, and more recently they've only been increasing logic density by 70%, cache and IO are getting very hard to scale and this problem should only get worse because future chips should have a smaller proportion of logic transistors in it, since it's the only thing scaling well. TSMC said their N2 node will only increase density by 10%. On top of that, new nodes are getting more expensive.
>One of the reasons AI became a lot better recently is because of specialized hardware, but we should be reaching diminishing returns there too.
>So I don't think our AI chips will be much faster in 2030 and this seems required to bring general AI in 2030.
>Density of iPhone chips: https://twitter.com/handleym99/status/1707073… https://en.wikipedia.org/wiki/Apple_A14 https://en.wikipedia.org/wiki/Apple_A1
>TSMC N2: https://www.anandtech.com/show/18832/tsmc-out…
>3 years ago, TSMC provided the first N5 ("5nm") iPhone chip
>...
>So I don't think our AI chips will be much faster in 2030
Really? You're measuring the progress on AI hardware by looking at the chips inside iToys? Nvidia just announced that they are having to double the speed at which they release new generations of their GPUs because of the pace (and profit) of AI development.
> Nvidia just announced that they are having to double the speed at which they release new generations
believing marketing BS
in the past Nvidia released new chips every 6 months, then every 12 months
now it's every 2+ years with lower end taking even more time
That chart of historical data isn't marketing, and companies aren't going to spend millions of dollars on GPUs based on untested claims. Your average consumer might buy an overpriced GPU because they saw some youtuber tell them to, but OpenAI with their internal benchmarks are the ones that Nvidia need to deliver for.
>https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/
>Resolution Criteria
Why do we have a higher standard for an AGI system than for a human? I think I've seen plenty of people that aren't capable enough to pass the test.
GPT-4 is already borderline AGI. People just keep moving the goalposts, equating AGI with ASI, and now demanding robot embodiment (which is only a few years away anyway)...
>As soon as it works, no one calls it AI anymore.
I completely agree, but it shows that what people want isn't "general intelligence", they want something that can replace any human at any task.
That's understandable, as most people want robot butlers, and middle managers want an AI that will let them sack all their staff, and CEOs want AI that will let them sack all their middle managers.
The irony is, though, that AI companies will focus disproportionately on the narrow task of making an AI that can automate the process of improving AI, so we might end up with an ASI before we have an "AGI" that can satisfy the "make a cup of coffee in an unfamiliar kitchen" test.
>2017, tfw no GF
Nah, I'm thinking MS is going to have subscriptions for windows 12 so they can pay for frickups like this.
FPBP
You gotta plant the seeds for trees you'll never get to enjoy the shade of if you want your society to flourish. AI development is one of those things where once the ball really gets rolling, it'll produce way more value than all the money previously invested in it by several magnitudes. The reality is, that most people alive today may only catch a glimpse of those dividends in the twilight of their life.
Lies. Once human workers are completely replaced these companies are going to hike the prices and killing 90% of the "industry" overnight.
hahaha dumbass
>Still have drivers
>Still have pilots
>Still have programmers
the previous CEO of github nat friedman said on twitter that this is bullshit and they actually make money from the copilot subscriptions
source?
Unlike NFTs, it has potential. They're severely cucking it right now in Dalle and ChatGPT, but it can become useful eventually.
Copilot more like Copajeet
>the day big corpo got swindled by the whole chain of production
bravo, ai lads + marketing
survival of the fittest
reminder that AI is about improving our lives and big tech actually isn't just evil and looking to get rich
they're already rich
>corpos are your friends 🙂
You probably believe jannies are your friends too
my dad is an exec so dont give a frick
When used correctly capital investments that were expected to be paid back in X time were instead profitable on considerably shorter timetables.
>inb4 SEC wont let me be
All that matters is how in tune the c suite are with automation capabilities. If they talk pipe dreams their cash goes up in smoke too. If it sounds legit, then the cash saved will be legit. Ez
Does he work at Nintendo?
It's publicly traded so the one and only answer I can say is go frick yourself.
It's closer to the movie margin call where whoever is first lives and the last are in lots of pain (if not chap7)
>reminder that AI is about improving our lives
Local models are about improving our lives and empowering the individual. Corporate AI is about slashing labor costs.
>reminder that AI is about improving our lives
by producing crappy pictures of people with too many fingers?
wasn't there an article recently about openai generating over 10 billion in revenue or some shit?
>wasn't there an article recently about openai generating over 10 billion in revenue
from where exactly? to do what?
Not from writing my shitty ERP.
Microsoft is working on local AI models for everything e.g teams noise cancellation, background blur etc. GPU prices will skyrocket in 2-3 years, once every company wants to run some AI shit on your pc.
remember the initiative for people to allow for the use of their gpus to find "the cure for coof"?
fold@home or some shit, right?
what if that was the test run?
Folding@home has been around for longer than you've been alive, Black person.
wym
scam
(cont)
same idea as in the push for electric cars being actually a push to offset capacitative capabilities to the general population
now imagine the same but with compute power with consoomer gpus
if AI is so smart, then why can't they ask the AI to help them profit from the AI? check mate, atheists
tech companies have worked with "AI" (programs that run on big data and statistics) for a long time, no?
btw, what you say sounds interesting. can you give more specific examples?
well guys, it's been nice, I don't think I will be able to afford my medicine after AI takes the noise canceller jobs
>GPU prices will skyrocket in 2-3 years
again?!
Many companies such as Google and Amazon took advantage of the AIs boom to launder money and evade taxes on living expenses and student aid grants.
And that's not counting the possible great algorithm or program they could take ownership of by being written into their platform or NVIDIA taking advantage of it to raise their prices higher and higher.
People hyping up AI just look at what it can do and ignore the infrastructure and cost required to run it.
AI is carrying me through my engineering degree¨
please dont im a lazy frick
Wallander's been making a good business with NotePerformer for several years now.
I think VulgarLang's doing well too.
Good, focused AI products can be a good business. Slop trash that's censored to hell cannot. NotePerformer won't refuse to play the Horst-Wessel-Lied.
>>>Good, focused AI products can be a good business. Slop trash that's censored to hell cannot.
I don't get it then
what do they want people to do with AI that is going to make them any money?
secure venture capital
>what do they want people to do with AI that is going to make them any money
data processing
Not if all they do is add a chatbot or image generator to their unrelated service
>AI MAKES NO MONEY
Or they're just saying that so people back off .... If you think there is no value you won't be up their ass.
Like when marijuana was legalized they said "no one is getting rich off of this"... Yeah fricking right... They're making tens of billions on that shit in just certain areas
Ye bro, like marijuana makes so much money brooo
>CURLF - 2.96B market cap, 2022 negative operating income
>IIPR - 2.12B cap, 2022 166 mil before tax
>GTBIF - 2.32B cap, 2022 209 mil before tax
>VRNOF - 1.5B cap, 2022 negative before tax
>TCNNF - <1B cap, 2022 108 mil before tax
none of these make billions of profits, they don't even make 1 billion put together
Craziest part is that if any big player offered unlimited unpozzed, truly unrestricted bleeding edge models for a fair monthly subscription people would line by the millions to get their AI waifus
AI is the future. You can't deny this. Right now its still developing and emerging but it will continue to get better and is already capable of some high quality things that require only minimal human intervention. You should buy low on AI right now instead of fear mongering or denying it, because it will only continue to grow. You'll be like the guy who said the internet was a fad in 1995.
We've already reached the diminishing returns point of AI development, and entered into the grift phase. Sorry anon.
>the Internet is the future so pets.com is a sure thing!
How would I buy into AI? All the startups are VC circlejerks you can't get into with $200k.
It seems the only thing is GPU and some oldschool robotics companies, but they'll loose or just buy the startups, diluting all value increases.
Create a product or service using it, a lot of the code and even models are opensource. No one is stopping you.
I meant investing. I would not personally want to build AI. In such a red sea market, with everything being obsolete in months, chances of sustainable profit are low.
>with everything being obsolete in months, chances of sustainable profit are low
I don't see this problem, most AI solutions are X-to-Y. If you build a product around text-to-speech / text-to-image / image-to-text / etc, you can simply replace the model out as newer and better ones are released. Just treat it as a black box that you build around
It seems I misunderstood. I am even doing that, but in a slightly moronic way that might give me some problems when ANN topology becomes more tricked out. We will see.
put money in the people making the hardware ie NVidia
I did that, with ETFs as well, but a bit late in spring. I just have a good case of FOMO.
Microsoft / google and now amazon dumped a ton of money into ai. I'm just investing in all those guys and nvidia.
so what does this mean for the near future? I'm in love with GPT4, everything be it web browsing, dalle-3, advanced data analysis which is just a python notebook. Will FOSS be able to pick up here or not if big tech pulls the plug
this
i've been pushed to use it cause it is actually better than google now
To some extent. I believe free voice assistants are behind the corner, FUTO voice typing + Llama on-prem + HomeAssistant or other integrations can do wonders.
If humans are able to come up with chips that can run AI's more efficiently then everything will change.
>AMAZON LOSES $100 BILLION A YEAR AND HAS NEVER TURNED A PROFIT
same shit.
It's the NFT scam all over again. What makes money is stupid consoomers who buy the paid versions.
But the problem is that they're not making money off the consumers
Automation technologies largely cause the tendency of the rate of profit to fall, because they (by definition) require minimal human labor inputs to replicate at ~0 cost until they hit market saturation. Which is to say: even if AI reshapes society, gayMAN could eat a loss. So far the real winners are the ones selling shovels during a gold rush, like Nvidia.
imagine how much money they would make from uncensored AI waifus?
>Big tech CEO's and BOT posters on suicide watch because Bing AI didn't instantly generate infiniate money the second it was turned on
Troglodites and grognards on this board still using Windows Vista don't think AI is useful, but like any new technology it actually matters how it's implemented and not how much money you can throw at it.
Microsoft is the king of throwing money at stuff to make imitations that don't make money back, remember the Zune? All the R&D to make Microsoft Edge before they just gave up and turned it into a Chromium fork? The Windows phone?
The fact that people still think Microsoft is the canary in the coalmine is bizzare, they're not, they're just a functional monopoly that can get away with constantly failing forward because their operating system is so ingraned into every facet of IT that they literally will never go out of buisness, at the very worst the goverment will absoutely bail them out, considering more than half of goverment infrastructure runs on their operating system.
Any "This thing in tech is dying becasue Micro$oft can't make money off of it" is just a signal for "I'm so ignorant that I think everything in tech is the Microsoft corporationm, and sometimes Apple, and nothing else"
>j-just 2 more weeks until everyone realizes AI is irrelevant and never going to happen!
lol keep coping
AI isn't going anywhere. Companies that invest in it for serious products are going to fall face first.
AI is just a toy and no one in their sane mind will rely on it for critical products.
>AI is just a toy and no one in their sane mind will rely on it for critical products.
of course not. it just creates word salads. its advice is often total garbage and it cant even do basic maths
personally, I can't wait for some dipshit to attach an LLM into some kind of customer support or claims processing engine and watching people turn them into a confused deputy to cause millions to billions in damages.
But anon! It's gonna replace us ALL!!!! Just like my heckin' starwars scifierino flicks!!!!!!! It's just like rickandmorty!!!!!1!
Microsoft has a lot of "not making money" programs in their pockets.
Bing
Copilot
Xbox
Zune
And now Activision(?)
>I knew it
>All they're doing is paying electricity bills for nothing
of course not. it does nothing useful and it isnt even real AI.
I was already pretty sure that they must be losing money given how much code I generate for $10.
My guess is they'll wait until userbase growth plateaus and they'll jack up the price and I'll keep paying cause going back to manual typing is not something I wanna do.
>market monopolization
>data mining
the only not trash posts in this 130-post thread
they've been regulated out of profitability. not by the legal system but by their diversity hires. tesla's still making bank off it thoughbeit
is blue skin colors the new thing?
what was wrong with just using the yellow smiley color if you want a neutral color...
Blue is the go-to color for most fintech companies, as blue represents trust and security.
Huh, I never got that message. Blue screen of death taught me it's not a good tech color.
No way software will replace humans. Humans are just too smart. If you don't believe me, go out and talk to a random person; they are irreplaceable.
So what you're telling me is that ChatGPT is a NEET living in its mom's basement without contributing anything or paying rent?
>the word 'grift' being used a lot in this thread
Is this the new hot buzzword of October?
are you the buzzword hunter?
>I bet the data they're collecting from people interacting with the AI is worthless too
always has been.
Is it just me or are we seeing less and less AI news and it seems like we are about to "pivot" to something else?
Anyone? Is AI old hat now?
No? Dalle-3 was only a few weeks ago and tons of other news have come out if you've been keeping up. The ride is nowhere near over
>Small improvements to LLMs
>Google releases something competitive to ChatGPT.
>Anthropic and OpenAI slightly improve GPT-4 and Claude2
>Meta or another group releases better open source models, up to around GPT-3.5 level.
>Small improvements to Image Generation
>Dalle3 gets small improvements.
>Google or Meta releases something similar to Dalle3, but not as good.
>Slight improvements to AI generated videos.
>Basic hooking up of Dalle3 to video generation with tagged on software, not really good consumer stuff yet. Works in an interesting way, like Dalle1, but not useful for much yet.
>Further experiments hooking LLMs up to robotics/cars, but nothing commercial released.
>Small improvements in training efficiency and data usage, particularly obviously in smaller models becoming more capable than older, larger ones.
That's just the rest of 2023. 2024 you'll see gpt-5
Or AI will get abandoned as soon as "the next big thing" poops itself out and the tech industry trips over itself to try and exploit the latest gold mine
AI is going to be the next big thing for a while. It's not going away, because its potential is limitless.
AI's potential is limited by hardware (Moore's law is dead) and censorship. It's not limitless at all.
>Moore's law
Irrelevant unless you can demonstrate to me how it's slowing down current AI progress.
It's potential is limited by being censored and rude to it's users when it has to be censored
It's potential is limited by needing huge amounts of training data that will get harder and harder to legally obtain, and it's potential is limited by how badly AI companies have already acted with this stolen data
>AI is going to be the next big thing for a while. It's not going away, because its potential is limitless.
Were you not alive for NFTs and Web3 anon?
There's never going to be another next big thing. AI is the new electricity, transistor, PC and Internet
You know that AI needs actual tangible hardware to run, right? If GPT-5 releases any time soon, it's going to be underwhelming as frick.
Yet more venture capitalist crap of being there first, getting the user, but no possible monetisation while fricking up everything else
>Yet more venture capitalist crap of being there first, getting the user, but no possible monetisation while fricking up everything else
its only the autists on here who think creating anime girls with 6 fingers is really awesome
I don't really care about the autists running this shit for free on their computers, or even shit like NovelAi. The one advantage is that we could have fangames made by amateurs with no art skills so they can stop making fricking Mario fangames
I care about the big corps spilling their spaghettis again on more bullshit to shut out anyone else
itt: zoomers who don't understand that it took decades for machine learning to get to where it is
but sure, it's a fad
or morons who don't understand that there have been multiple AI hype cycles
"experts" genuinely believed that AGI was around the corner 50 years ago
>I bet the data they're collecting from people interacting with the AI is worthless too
I beg to differ, it's probably much more intimate, precise, and data-mineable than any other kind of metric, thus very valuable, compared to narrowing down your interests based on ad misclicks, shared family members, IPs, and anti-trackers. Here everything is given on a silver platter.
The data they're mining is garbage. When you scroll through BOT do you see a sea of valuable content made by honest human beings? Most social media content is grifters, shills, bots, trolls, and trannies. There is no good data to mine, like trying to quench your thirst by drawing water from the Ganges. Bro it's nothing but shit.
What is "valuable content"?
Content you can make money from.
You can make money by using AI, just by not straight up selling its outputs (mostly)
>The data they're mining is garbage.
That's the thing. 99% of the internet is garbage, but with AI it's easier to filter the garbage automatically.
Moreso if you require the users interacting with the AI to not write 500 pages novels of their sex fantasies.
It's like comparing gold panning to industrial sifters, the thing coming in has the same ratio of garbage.
>with AI it's easier to filter the garbage automatically
Nope, at that point you're thumbing the scales and giving yourself the information you want to see. Only marketers tell you this information is valuable, and they do it so they have something tangible feeling to show the corporate dumbfrick they suckered into paying for their services.
I dunno man, Nvidia has been raking it in.
>During a gold rush, sell shovels
Nah, AI is definitely gonna replace humans but not in a way that normies expect. Tasks like programming are very tedious and require a 100% certainty that it works and doesn't have hidden bugs. Imagine debugging a complex multithreaded program that AI wrote - it is equivalent to debugging a multithreaded native program in assembly. Also the company that makes the AI possibly has legal issues with companies that use that AI to make their software. All in all - software is probably finna be resistant.
First ones to go are translators/support agents/call center people/doctors(not surgeons, for disease determination)/advocates etc. Last ones to go will be the hentai addicted, coffee guzzling, cum smelling autists from the sysadmin branch. End times come when these guys are replaced.
>Tasks like programming are very tedious and require a 100% certainty that it works and doesn't have hidden bugs.
>it works and doesn't have hidden bugs.
Inb4 any latest AAA games. Or Boeing software that randomly dives if the pilot tries to fly up.
Programmers, by working with copilot, are training their own replacement.
Are you fricking stupid? No one is going to trust chatgpt over a fricking doctor to see if they have cancer.
I distrust doctors so severely that I'd trust a BOT poster over them. If anything chatgpt will be an improvement
>No one is going to trust chatgpt over a fricking doctor
Public healthcare will, at least as long as it's more economical.
It's not like actual doctors are any better.
>people getting to choose their doctor/healthcare options
Sure Anon, you can talk to a real doctor. Do you have the Extra Platinum Healthcare plan? No? Then you get to be diagnosed by the hospital equivalent of a McDonald's Self Order Kiosk.
Literally the metaverse all over again
AI has capacity to do great things but only in the long run. It will require investment that isn't all of the money right now so no one will put the money needed for a slow-burn-but-high-reward project
Literally like the metaverse. Tech companies are fricked and we'll see a mass dieoff in the next ten years because they can't make instant investment returns anymore.
>stealing all of the data they can, hoarding it like dragons. They don't give a frick if it is childrens data, company trade secrets, or copyrighted content or not
>Spending huge sums of money on processing and wasting more electricity than Bitcoin miners
I honestly can't wait for AI companies to get fricked in the ass. OpenAI should be burned down and Altman is a psychopath who should just plead insanity before he gets 10000 life sentences for copyright infringement
Altman wants to (I don't think he can, but his goal is to) be the sole private owner of the world's only AGI, he's literally the technological bad end science fiction writers have been doomsaying about for decades.
clickbait. business 101 is that you're never profitable from the get go, that's day 1 shit.
?feature=shared
They're probably conflating operational cost vs R&D cost. There's no way the electricity to run a prompt for a given model isn't measured in pennies and it's possible to put your datacenter somewhere with extremely cheap electricity. Electricity costs is not going to be a bottle neck.
I don't care if they lose money from people using it. I have no need or want satisfied by using it, so I don't.
I don't code but I've tried using language models for work-related stuff and the really damning thing is that it's just factually wrong about most of the answers it gives. Kind of a big obstacle lol I mean it looks very cool and flashy but it's obviously just like, a convincing spellchecker.
Rewind to ten years ago when it was all about 'tech startups', we similarly saw a decade of VC outpouring, the most favorable economic conditions imaginable, yet not a single usable product at the end of the day (besides the unprofitable ride-sharing apps that stay afloat on raising funds by fudging profit stats).
ai vill replace ~~*artists*~~ and take all their money
all hail ai
>AI has not been profitable
Frickers do not know the meaning of the word intangible. Not even joking. This is literally the race to be the one who will provide AI and bill people monthly for it, AI will be literally the new electricity.
I use AI art to expose myself to random new ideas. It's like going outside and looking at THINGS! Using your imagination to see patterns where there are none. If you stared at textured drywall, you'll see things if you look long enough. That's what AI is to me, but I get to throw "cute anime girl" into it.
>I use AI art to expose myself to random new ideas. It's like going outside and looking at THINGS!
enjoy
>Microsoft loses $20 per user per month
so charge $21 per month
Very few people would pay that.
well at least they'd make a profit
They wouldn't because the cost per user would dramatically increase due to fixed costs.
>fixed
Just spray it with WD40 it'll come loose again.
>Double price
>Half users quit
>Double price again
>Another half quit
>Double price again
>Only 10% quit but you are now hemorrhaging userbase and the investors all leave despite you being technically profitable
This is why all the companies instantly started lobbying for "AI licenses" when open source generation started advancing so quick.
They're really fricking spooked by how quickly foss software was able to catch up, despite what fud about Dall-E 3 and GPT-4 superiority would currently have you believe. Models trained on one task (e.g. a model trained on only python) on home hardware can beat GPT 3.5 at that task, small models can run (poorly, but advancing slowly) on phones and laptops. This is foss beating their efforts. The current leadership of Dall-E 3 and GPT4 is immense, but it's because they have enough server power to run multiple models on huge racks of teslas. Dall-E 3 passes your prompt to GPT4, which is in itself rumoured to be 8 different models working together, and then passes that back to Dall-E. It's immensely, hugely wasteful but gets results now.
Bing is forcing out GPT-4 so quickly and tanking so much cash on Dall-E 3 because they want to be the name for AI and eventually convince people to pay for it by putting it in office 365 and going to large corporations and offering "hey can we make a model for u pretty please", and eventually being the name people go to for AI models once they shrink enough that everybody can get high quality results. The idea that home hardware can run even poor quality results for free while they're spending billions on server racks scares the shit out of them, because it means they'll always be racing for higher quality results and spitting it out as soon as possible to widen the gap, never being able to stop upgrading hardware and keep that gap going.
>competition forces them to try hard to improve rather than stagnating with an easy lead
Good.
It is good. Facebook releasing LLAMA for free is quietly one of the best moves for competition that company has done.
i don't care if it makes money. it's fun to play around with. money isn't everything, homosexual
If they offered a simple monthly sub for unlimited, uncucked, unfiltered and unlogged gpt4 access they'd have more money than god.
Coombux are the big bux but the people in charge would rather cut their own nose off to spite their face.
>spend billions training highly intelligent AI
>spend even more scooping out 90% of that data that's actually intelligent for "ethics purposes"
how could dis happen
meaningless. You do know Twitter wasn't profitable until recently right? same with Uber and every other "successful" startup. They will be profitable eventually, or collapse and be replaced with other companies making the same promise. They could be unprofitable for the next 20 years and as long as VCs keep investing in them they'll continue to exist
If the internet existed back when cars were first being released to the public we'd have had a bunch of posts like this.
>they're too expensive, the manufacturer doesn't make ANY money!
>they're so loud and they spew pollutants out the back, not at all like my beautiful horse!
>these will NEVER catch on! they'd need a million years of development to even be profitable, let alone replicate a majestic horse!
People who think AI isn't going to be involved in or outright replace >90% of all economic activity in the next 20 years are fricking mongoloids. Just ignore everything they say.
Midwit take.
Cars replaced horses because they were cheaper to use and maintain. They also didn't shit in the streets.
They also made tremendous amounts of money for large corporate interests and governments while vastly improving the quality of life of those who had access to them. AI will do the exact same thing.
>X thing happened therefore Y thing will also happen
Not how it works.
>or an AI that works 24/7 at 1000x the speed and efficiency of their human counterpart with the only cost being electricity and bandwidth?
What you are describing won't ever exist, therefore the human.
>What you are describing won't ever exist, therefore the human.
Oh sorry, I didn't realize I was talking to a mentally handicapped person. Carry on.
>Oh sorry, I didn't realize I was talking to a mentally handicapped person. Carry on.
Do you know what S-curves are, dumbass?
The only curves I care about are on your mother
Are you sure? My mother is 73 years old.
Do you have any pictures of her feet?
based mine turned 74 today big fat cuddly milkers mwahpop nipples
>Cars replaced horses because they were cheaper to use and maintain.
Which do you think is cheaper to employ? A living human that works 8 hours a day, needs healthcare and other benefits, constantly b***hes for more pay while trying to minimize the amount of work they do, or an AI that works 24/7 at 1000x the speed and efficiency of their human counterpart with the only cost being electricity and bandwidth?
the AI which is really intelligent will quickly surpass humans and couldn't be employed, in that case the AI would be the one employing humans for tasks requiring physical presence
this is the biggest issue with those wanting to get rich from AI: any really intelligent AI will make you and investors irrelevant
They're still in the data mining/dependency creation phase. Once the silicon gets efficient enough and people are hooked enough and the models are big enough profits are supposed to be inevitable.
How come they can't ask ChatGPT how they could make money?
there's no dataset
so your telling me snake oil tech is a scam? Wow its not like i worked out the limits of ai 20 years ago playing with the same tech in videogames!
looks like we've got a tech prodigy here
tell us, what are the limits of ai?
how much longer do we have before no new ai advance is possible?
two more weeks?
what do u think moron you think this is the first time we've advanced to AI?
i think that anything a neuron can do, can be simulated by a chip
i also think that human intelligence isn't the limit to how intelligent a system can be
maybe we don't have the necessary algorithms yet, and maybe hardware progress is slowing down, but i don't think we're anywhere close to the limits of ai
1 ai replace 1 job but then requires 2 humans to parse and double check for the sake of legal accountability
The economy is saved, actually
Criticising someone else's work is always much easier than doing the work itself, so it's more like 1 human will check the work of an AI that replaces 10 humans.
Since when has big tech been profitable anyway? AI is profitable but not when they're giving the service out for free for millions of users for something that requires a billion GPUs to run.
Amazon didn't make a profit for the first 9 years of its existence
AI gives companies excuses to make their product a service. Paying a monthly subscription for Photoshop is stupid. But with AI? Now they can justify it since your hardware can't run a local copy of the AI.
All software will incorporate some form of AI as an excuse to always require an Internet connection and a subscription. It will make money.
Microsoft doesn't care about profits they just want to use the primitive version of "AI" as a stepping stone to create a global control grid for the new world order.
>pls feel bad for the mega corporation
you all fell for it so easily
https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-second-quarter-fiscal-2024
Second-quarter revenue was a record $10.32 billion, up 141% from the previous quarter and up 171% from a year ago.
AI, IT PRINTS MONEY