And yet people still can't give me a good reason.
I've given you a good reason since my first post in response. Because it's cheaper and more efficient to follow closely than to lead and there's no rush to lead so it's wise to conserve until overabundance causes efficiency to no longer be important.
Again, this is simple logic here. Let me meme it for you, why I'm so worried.
Again, it's simple logic that if you can't make these things, you don't talk trash about the people who do. I don't care why you're freaking out and neither does anyone else.
GTP-2, GTP 3, PaLM, LaMDA, flamingo, openAI codex, Megatron, Github copilot: I sleep
ChatGTP: Real shit, entire A.I sector on fire within a month, dozens of ChatGTP clones coming right up, dozens of companies and startups created just to create LLMs, tens of thousands of business are looking to interagte LLMs into their products and business models
It's not even an issue of why China couldn't come up with their own LLM in the first place, it's that they couldn't even recognize the potential of LLM until it was so mainstream that it couldn't be ignored. This are supposed to be A.I researchers on the forefront. There's clearly something very wrong going in the industry if they could ignore PaLM or flamingo, two very impressive LLM that were on my radar when they came out even though I don't work in the A.I sector. PaLM was impressive in the way it could understand meta-reasoning behind jokes and memes, probably the closest to ChatGTP before ChatGTP came out.
I though that China was already making great progress on the LLM front since GTP-3 rocked the boat and with every development of google/deepmind/facebook's new LLM, but I just never heard about it. But it turns out that most of the major A.I players just didn't care?
It's not like they're doing well on other A.I fronts.
Sure LLM may not lead to AGI. But given their failure to even do recognize the potential of LLM, what hope do they have of actually innovating or even recognizing whatever revolutionary new A.I model that comes out of America before it's too late. Unless something changes, China is going to lose the A.I race for sure. And that's the most important technology race ever.
I would love it if China can actually can innovate instead of outright saying that they need to copy ChatGTP, as have many chinese A.I companies have.
TLDR. Just you complaining why people way smarter than you haven't accomplished even more by overtaking the US even faster.
And A.I was supposed to be the field where China and America were neck and neck. After all, the A.I winter stalled a lot of progress on this field. Deep learning was a pretty new development. It's a pretty even playing field I thought.
LOL AI's not just some chat program you're freaking out about and even then, a few months apart is neck and neck.
I'm just saying how it is. I knew from 2020, that GTP-3 and it's successors were going to be a BIG deal. Chinese A.I sector didn't get the memo until ChatGTP in 2022. Excuse it however you want.
1. I don't see any posts from you in 2020.
2. How do you know who got what memo? Knowing something is coming isn't the same thing as getting ahead of it. There's no reason that China needs to invest massive resources to debut a chat program before the US.
Probably because I don't have billions of dollars, an team of researchers and my own supercomputing center.
Ahahaha, excuses. You're not smart enough to work as an engineer for a top tech firm so you criticize them like an insect biting at a lion's toenails. China's tech giants weren't born with billions of dollars either. Now, they got the money and you got the freak outs LOL. Stop making excuses for your own failures and demanding perfection from others.
The question is, why didn't those who did, start research back in 2020 when GTP-3 was making waves, instead of waiting until 2022 until ChatGTP exploded onto the scene?
No need to invest huge amounts in a race for a few months when you can let others do the heavy lifting and come up with something similar right after. You seem to be obsessed with who's first like the Soviets were obsessed with who have the biggest factories/airplanes/electricity plants/etc... but you don't think about what you need and how to most efficiently get it.
When dealing with AGI, a few months can catapult you a century ahead or behind.
Uh no. A month is a month and a century is a century. Unless you're in a wormhole, time does not distort. Unlike you, actually smart people (or even average people) know that. If someone invents an AI program and you come up with something similar in 5 months, you're behind 5 months, not a few years, or a century or any other nutjob freakout time.
Only if they land on you can they drop you a century behind. Not the same thing as AI.
Or someone gets AGI first, it upgrades itself over and over in a few days/weeks/months, becoming a ASI and proceeds to do whatever it wants? Not something you want your competitor to get first.
It upgrades itself and starts evolving until it takes over and does whatever it wants? And in this fantasy of yours, are there machine tentacles that interrogate the young female captains of the last remaining human rebels? LOL In the real world, whatever lead the US has in tech, China can overtake. First means nothing. Take your medication and stop watching Matrix reruns.
Soft power is very real. Just look at us, typing in English, giving more data for the whatever new AI OpenA.I is doing to dote out next.
That's not soft power. We're speaking in English talking about how China can overtake the US. If soft power existed, that would be a perfect example of its failure.
