What do you mean "think of ways to solve it"? Do you think government officials or corporate executives read this forum? Do you think you're some kind of consultant?
...are you going to write letters to Ren Zhengfei again?
News on China's scientific and technological development.
- Thread starter Quickie
- Start date
...are you going to write letters to Ren Zhengfei again?
There’s lots of research going into using Neural Nets for machine control law algorithms right now that will matter a whole lot more for industrial applications than stuff like ChatGPT, but because it’s not consumer facing it gets no attention. Yet that stuff is going to be way more important examples of AI applications than a textbot that regurgitates text libraries from the collective history of the internet. There’s definitely utility in what ChatGPT does but what it means for AI development is being way oversold. It’s a very impressive tool for certain knowledge access use cases. It’s not going to turn into Skynet in the next five years.
Remember when Google's Deepmind was supposed to solve AGI and take over the world? And what happened? Jack Squat happened after their one trick pony show of AlphaGo and now Google had to issue a code red after it got caught with pants down when M$ft pushed OpenAI to go out of the gate with ChatGPT...
ChatGPT is nothing special, even OpenAI CEO admitted its just a clever parlor trick at this point... And the underlining GPT model its based off of has existed for a while now, its just that unlike IBM Watson that won at Jeopoardy or Google's AlphaGo bot, this was the first time something like this was made freely available to the masses to test and play around with...
But now research is being made on how to run mini versions of it locally, without cloud or supercomputers, see:
Not sure where, why or how you think this chatbot crap makes US ahead of China in AI
We are nowhere close to AGI or singularity
.....
If China's A.I sector actually reacted to GTP-2
Stop trolling::
and that so called text to graphics thing you were talking about? Its open source, model and all, and runs on a single computer with medium gpu
Same for text to voice, including Chinese audio models
None of this stuff is "EUV" level of hard... Im already running all this on my rtx4090 gaming pc
Last edited:
FWIW, I work in the business services sector, and I've noticed ChatGPT has really streamlined some of my colleagues' work and made them much more efficient (Nowhere near taking my job though 
). It's going to see a LOT of use in services but I'm not to sure about everywhere else. I think ChatGPT gets a lot of attention because it's consumer-facing with obvious applications. It also definitely gets a lot of training it because of how far-reaching English is.
). It's going to see a LOT of use in services but I'm not to sure about everywhere else. I think ChatGPT gets a lot of attention because it's consumer-facing with obvious applications. It also definitely gets a lot of training it because of how far-reaching English is.
I tried telling ChatGPT to write a business/RD proposal. It came out with something plausible looking that I could revise.FWIW, I work in the business services sector, and I've noticed ChatGPT has really streamlined some of my colleagues' work and made them much more efficient (Nowhere near taking my job though
Turns out I revised every word by the end.
Still waiting for someone to try to do a whole coding project with ChatGPT only to find themselves stuck in debugging hell.
Yeah it's pretty rough. It can and will give you fake tax codes for instance. I mostly use it to draft emails and other super simple stuff lol.
What are you talking about? The issue is quality here. GTP-4 already talks like a literature professor and can provide sources, a vast step up from a braindead youtuber script who has trouble with basic english.
What an underestimation of the potential uses.
??? Better than pretending that everything is fine. It's not like some people here don't work in A.I sectors or related industries that can give some feedback. Otherwise, what the point of this entire forum?
Moving the goalposts? You were the one that said that A.I can't form human relationships. And this is for manipulating people. That's the whole point. One day your A.I GF is just sex-texting you, then one day she's asking you to buy more Apple products, than she's dropping hints that all chinese products are spywares, that you should work out more, that you should join the army etc etc. Imagine facebook, reddit discord, but with an army of bots subtly just dropping subliminal messages everywhere, and you can't tell if they're real or bots. Hell, with A.I text to voice, an A.I can legit just talk to you in a real human's voice.
So you really think that LLMs are useless or the array of media generation models are useless? Fine. Time will prove me right in the end. Let's see GTP-4 and other media generation model change the digital landscape while China is once again left in the dust.
Probably because they're not as useful?
One reason is that it exposes just shortsighted Chinese A.I sector is. This is the most important technology in a century and they choose to flip their shit over ChatGTP? Why not GTP-2? Or GTP-3? Or PaLM? I was flipping my shit over GTP-3 when it came out 2 years ago, but the chinese A.I sector didn't raise nearly as much fuss as they did over ChatGTP. Funny how the only LLM to gain mainstream attention was the only LLM that finally kicked the Chinese A.I sector into high gear. It's doesn't bode well when your top A.I researchers seem to be blind to the progress in their field of study and has to rely on MSM to decide what projects to invest in.
If we had seen ChatGTP level of interest and activity over GTP-3 or GTP-2, then maybe Chinese A.I sector would have been the first to release a ChatGTP-tier product, instead of currently, where more than a dozen companies and labs are scrambling to catch up.
You seem to be flipping your shit over a lot of stuff... go back and reread my post where I already refuted your misplaced concerns. Instead right now you seem to be intentionally ignoring what I posted, then double down on your factually wrong statements, then flip out about flipping out...