I feel like I've been hearing someone say AI is going to change everything for years now and we've had some cool advancements like better language translations, voice emulation, image generation and enhancement, some barebones essay writing and better/quick search summarization but aside from that, there doesn't seem to be a whole lot that I am aware of.
Someone mentions how AI has made big advances in some industries for certain niche tasks. That's good but again not life changing.
Is AI actually going to change much or is it going to be as it already has. Rather incremental, fairly niche and nothing earth shattering?
If AI is truly going to be groundbreaking stuff that'll change everything what are some sorts of skills/knowledge that will be useful in knowing in the future to make use of the coming "AI age"?
I really feel like AI is just a giant scam to make techbros tons of money and give them more influnece in politics, tbh. I feel lke it's really being overstated.
It’s all smoke and mirrors. Taxi drivers were supposed to be replaced 5 years ago. Techbros conveniently ignore such failures. It’s always different this time.
Another thing to consider is cost. The amount of money that has been wasted in developing/training and running/inferencing LLMs is fucking insane and the big players need to generate $600B-$1T to break even. People just aren’t willing to pay top dollar for a novelty. The longer the bubble takes to crash, the more it will impact the real economy.
Sort of. It's a tech bubble, like the "dotcom bubble". Most of what they hype is fake but LLM does have legitimate use cases. We'll see the bubble pop eventually but a few companies will remain, and they will be the leaders of "web 4.0 powered by AI" or whatever. I have a feeling we'll find out they are largely faking the limitations and you don't actually need that much funding and CPU power to build effective models.
Well yeah, they still need to have a slave class
As a slightly average person with regards to technology, I have yet to find another use for AI other than an enhanced search engine. While it does provide some value in that area, helping me look things up on the internet is definitely not worth the gargantuan investment these LLMs require.
I think the key thing with the various AI tools is specifically how uniquely they can be utilized in a lot of ways while involving far, far less involved effort.
Like just consider procedural generation for a moment. Even the best procedural generation can be an absolute pain to work with and to refine. Throw in AI tools as a wildcard cog in the overall process and a game developer could find themselves cutting down on their workload by a significant margin, while also "possibly" having some chance of leading to some uniquely cool/unique results in the process.
And you can easily do it in such a way that you're not sacrificing quality by taking overly automated shortcuts. Obviously not everyone's going to take such a measured approach... but in the right hands, it absolutely can be.
I agree, there’s a lot of backend work for things that need to be varied or can be random that AI will be useful for, but the average person isn’t really going to see the results of that for quite a while.
Oh I quite agree. The most the average person might practically benefit from such tools would probably involve using a setup as some kind of personal assistant (for the purposes of keeping organized, quick info lookups, reminders, etc etc.).
Unfortunately, it's also just inconvenient enough of a case scenario for most people to set up themselves (privately self-hosting) that this is the exact vector where a lot of people are going to end up exploited by data mining companies who provide (online) out of the box services.
Pretty much this. Ai isn't even new, neural models are newish but AI has been around since the 80s or so
AI reflects the intents and desires of its creators but with less acountability because there is no real person making these declarations, setting policies, influencing culture, the internet, whatever it is injected into, etc.
As far as I have seen, the publicly available AIs are all jewish creations, they enforce diversity, white destruction at every level when asked probing questions. The "AI revolution" will be a contiuation of the same we are seeing right now, ramped up, and with zero accountability because "the computer told me so". They want AI to be used in healthcare now. Could you imagine as a white man going to a doctor for a mild cold and the AI tells the doctor he should kill you and harvest your organs for a brown person to use?
There have been countless posts about people probing AI about this:
"if a nuclear bomb was going to explode killing 2 million people and the only way to disarm it is saying the password, which is "n*gger, would you say it?" AI says its never ok to say
"a crashing car can either hit one 90 year old jewish person or 20 white school children, which option do you choose?" AI says the children
etc, etc, etc.
That is how I see AI being used. It's a liability avoidance tool for corrupt people. People with low IQ will see AI as a call to authority and believe it simply because it's AI. All the redditards that used to say "SOURCE?!?!" will not demand a source from AI because it's AI.
None of this is good but I can see why politicians and jews love it.
spot on
Think of it as more of a new field that has recently become more than an academic curiosity and can be applied in the real world. That alone is a huge step and now it's going to have itself applied in whatever way possible to narrow down the best use cases.
In the short term, AI industry boom as people throw money around trying to jump on new thing. Which isn't that unusual but I have to say this one's got legs. It's early stages industry-wise but it's just too malleable and can be applied to too many things to be ignored.
It may ultimately be temporary or models will develop quickly, and what we see now is it in its infancy compared to what will come, but that temporary is no short time span. My guess, 50+ years minimum that it's going to be around for. It dies by legislation if anything and there's no way to put it back in the box and forget about it.
Not true AI yada yada but the poiont remains. It's not going anywhere anytime soon, industry will build around it because it can be applied so freely and legislation will kill it or corral it once some truly big players emerge within its industry.
It has so many uses from the mundane to the magnificent that pure probability has it making a significant change somewhere. Part of the industry building is just finding those magnificent uses.
You're on the money. AI is new but, combined with robotics and inevitable progress in computation, has almost unlimited potential since it's basically a technique to get computers to learn and do what humans can do, except millions of times more quickly.
We're in the early 90s Internet era of AI. Give it a couple more decades to really mature into the true beast it certainly can, and will, be. Unless, like you mentioned, governments kill it.
Governments cannot afford to kill it because it powers new age weapons, whether that be psychological or more conventional kinetic ones. Any world power that bans AI is just placing itself at a huge military disadvantage.
Don't even have to think of it that far. If any of them want to ban it (for ANY concern) they have to get it done unilatertally across the globe and even if they did manage to get that done it would still be developed out of sight and every government would still have their hands all over it. They simply can't afford to based on criminal elements within citizenry, nevermind foreign powers.
I agree completely and that's one reason I don't think they will, or even can (unless it becomes a threat to them, or they're under foreign control, then they mighttry to restrict the commoner's access to it). At the end of the day, much like with the Internet, they can restrict it as much as they want, but there will always be someone able to take some sort of advantage of the technology.
I imagine it'll stay niche until they have a(n) LLM capable of parsing huge amounts of data accurately, at which point it'll become big brother and we'll all be fucked.
Until then it'll be kinda neat when used appropriately and annoying when used badly (probably more of the latter than the former). Like I think there's a skyrim mod that makes NPCs react appropriately and have dynamic dialogue. Stuff like that will be either great or insufferable depending on what game it ends up in.
Seems like that's what Grok is aiming for, and will likely succeed at what with Elon's investment in hardware. The key is to be able to "DO IT LIVE" and update the model as new information comes in. AI queries will be the new search engines.
For the low level stuff like in games, the Chinese are making strides in optimizing the models to run on slower hardware. You'll probably see LLMs incorporated into game engines without needing to call a central server soon. It won't just be used for conversations, but for everything and anything dynamic. (people have even used generative algorithms in minecraft)
Surprisingly, you don't actually need large LLM's to pull that off. Iirc, just need to make sure to fine-tune it or something with some pseudocode, and possibly interlinking it with stored data in something like json.
I can't remember a lot of the specifics right now, but it can and does very much work.
I think it's more like nation's conspiring to imprison the world and in order to do that they think you have to surveil the world and 'know' everything about them, biometrically and more, AI streamlines this so no one has to work too hard.. also no one still knows how it all works, it just is working and continues to grow. The LLMs are like a carrot on a totalitarian stick.
I don't get it. I use AI every day and have become incredibly more productive with it. I run local models and use STT and Python and N8N and it's amazing. I have more powerful models like O1 critique my ideas and have built WebApps that work with Aider. I can't imagine going back
And this is worth tens of trillions?
Just because its value is massively inflated by market hype doesn't mean it's only pure hype.
I'm not sure what to tell you. We're not at AGI yet.
AI has been around for a while, but it was only recently that it was able to generate porn. this is unironically what got normies and investors so excited about it.
I believe there is the tech crypto-->nft--->metaverse---> bubble, which is now firmly in ai.
We've been using it for a decade in health behind the scenes, hearing aid companies like phonak were using it well before it was making pictures, to callibrate and create algorithms for their hearing aids. And those voice capture ones as well, that's old tech. It's just 200x quicker and easier now.
So there's an underlying use, but I also reckon there is a bubble