-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
ChatGPT and the (near) future of AI
- Thread starter 8BiTw0LF
- Start date
diffusionx
Gold Member
lol, a moderation bot would be more expensive than jannies though
ThisIsMyDog
Member
That test was passed years ago, for GPT-4 it would be easy as fuck to pass. Its useless now.
Guilty_AI
Member
would probably depend on the magnitude. I believe for a huge social media site like twitter or facebook an AI would be cheaper.
Last edited:
diffusionx
Gold Member
jannies do it for free
ThisIsMyDog
Member
Nuts. I wonder when old farts politicians are going to start realising that something strange is going on?
Last edited:
Vestal
Junior Member
I am down the rabbit hole of GPT4. This thing is going to put Junior Devs out of work lol.
After spending the day building scripts and code snippets to use at my job I can honestly say that I won't go a day without using GPT4 for my job going forward. The ability to easily generate usable code and logic is unreal. It is so easy to give it a query and let it do most of the grunt work in creating a basic skeleton for a function or code snippet. All it takes is a couple of additional queries to fine tune them to your needs.
After spending the day building scripts and code snippets to use at my job I can honestly say that I won't go a day without using GPT4 for my job going forward. The ability to easily generate usable code and logic is unreal. It is so easy to give it a query and let it do most of the grunt work in creating a basic skeleton for a function or code snippet. All it takes is a couple of additional queries to fine tune them to your needs.
E-Cat
Member
I disagree, it still doesn’t have expert human-level coherence and factuality over long contexts. It’s also not good at spatial reasoning, arithmetic or variable binding.
Last edited:
Pagusas
Elden Member
Waiting for this to hit the multimedia/video area, soon as they open it up to start allowing it to edit videos, create motion graphics, edit testimonials and such, well, I’ll need to find a new side business.
The only current thing I think holding that back is the bandwidth in dealing with raw video.
The only current thing I think holding that back is the bandwidth in dealing with raw video.
Last edited:
PuppetMaster
Member
So like 90% of humans.
Omega Supreme Holopsicon
Banned
PSYGN
Member
AI will churn through all the PR, Slack, discussions and be able to form context and reasoning and map similarities before and behind its changes. Senior and junior devs will have to correct it and tell it why it's incorrect - this would be the AI in the training stage just as any junior developer trying to find its way through a large codebase, but unlike starting from ground 0 with every new junior dev, companies will just be able to spawn more instances with this contextualized knowledge pool. But in the meantime I don't think it'll put junior devs out of work. If anything, these tools will let junior devs work faster and perhaps with less obvious mistakes, and allow them to get feedback as to why solution X is the best way without having to take time away from a senior dev or spend as much time researching online. To be honest, I think much of what senior devs do in most coding roles will be easily handled by AI, and context can be trained over time. The "safe" devs are the ones working on novel ideas or has some physical aspect to it (robotics, networking).
Last edited:
E-Cat
Member
Not all humans can pass a proper Turing test.
SJRB
Gold Member
I'm 30 minutes into this interview and it is borderline unbearable. As always Lex's heart is in the right place but man is he a poor interviewer. You'd think someone gets more experienced and skilled at interviewing over time, especially 367 podcasts in.
Lex, my dude. Please formulate concrete questions before you start an interview. Literally every question he asks Sam starts off as a basic question but then he spirals into some quasi-philosophical word salad. I know you are super smart and I know the subject is super complex, but jesus christ just ask the man a coherent question.
PaddyOCanager
Member
Yeah, I watched the whole thing because it really is an important interview, but Lex is almost always a tough watch. I rarely sit through an entire podcast of his, typically the clips are easier to bear.
Hugare
Member
My feelings exactly
Also his monotone voice puts me to sleep.
About the topic: I'm using ChatGPT daily on my work now.
I've programed many little scripts that can extract information from .pdf files, rename files and etc.
It helps a ton with repetitive tasks. I've automated a task that required more than 10 hours from a coworker each month.
I'm almost paying for the upgraded version
CrankyJay™
Member
Yeah baby, let's go!
https://www.msn.com/en-us/money/oth...1&cvid=5173883bbbe146c19186537df09bc25a&ei=33
https://www.msn.com/en-us/money/oth...1&cvid=5173883bbbe146c19186537df09bc25a&ei=33
sendit
Member
You're fighting a losing battle. If you're not augmenting your workflow without ChatGPT, you're already behind the curve.
E-Cat
Member
You need to brush up on your reading comprehension. I'm not "against" ChatGPT, and am using it. It has nothing to do with whether I believe it is currently capable of passing a Turing test.
Fools idol
Banned
I believe we are a stones throw away from this being used to stage major cyber attacks.
My lead developers have been testing this out constantly since release as we were lucky enough to get an early access pass to the API. It's mind blowing how far advanced this is, especially the leap over from GPT3 to GPT4. It's like if the original Gameboy came out and then 6 months later the PS5 arrived.
My lead developers have been testing this out constantly since release as we were lucky enough to get an early access pass to the API. It's mind blowing how far advanced this is, especially the leap over from GPT3 to GPT4. It's like if the original Gameboy came out and then 6 months later the PS5 arrived.
Sybrix
Member
Interesting article
I concur.
Elon Musk among experts urging a halt to AI training
Elon Musk among those calling for training in powerful artificial intelligence to be suspended.
www.bbc.co.uk
I concur.
Last edited:
Omega Supreme Holopsicon
Banned
each year agi is delayed we get closer to collapse, and tens of millions of avoidable deaths occur.Interesting article
Elon Musk among experts urging a halt to AI training
Elon Musk among those calling for training in powerful artificial intelligence to be suspended.www.bbc.co.uk
I concur.
E-Cat
Member
Concur with what? The West cucking itself and signaling to rogue states that AGI is really close? Sounds like a recipe for disaster.Interesting article
Elon Musk among experts urging a halt to AI training
Elon Musk among those calling for training in powerful artificial intelligence to be suspended.www.bbc.co.uk
I concur.
Wildebeest
Member
I got google Bard to admit that Yoda is a goblin.
Cyberpunkd
Member
Let's not forget the important part - the code is PERFECTTTTTTTTTTTTTTTT, no more fucking corrections because you forgot to put something at the end.
This thing is legit, thinking already how I can use it in Sales. The LinkedIn content creation I was fooling around with is insane, down to stupid motivational hashtags.
NecrosaroIII
Ultimate DQ Fan
I tried out Bard today. It seems pretty similar to ChatGPT. I do feel that ChatGPT is a bit more articulate though.
Omega Supreme Holopsicon
Banned
North korea builds true ai and then...
Grey Goo: The Disturbing Way Our Civilization Could End
Last edited:
E-Cat
Member
From the viewpoint of employment, this thing's like the pleasant breeze of fresh air against your cranium moments before your head gets smashed in by a sledgehammer.
Hugare
Member
Dude is salty.
I bet one arm that if he was still in OpenAI's board (he jumped out of the company some years ago) he wouldnt be saying this
We have some long years ahead until we see an Ultron. For starters, GPT is not connected to the internet. It's trained offline.
You cant fight against progress. They may suspend it to the public, but companies will keep evolving it on the background. And people are already sharing code all over the internet.
The genie is out of the bottle. Its over.
Its like this all over again
Omega Supreme Holopsicon
Banned
I think the plugins are connecting it to the internet, as one of the examples is of putting items in a shopping cart for someone to make the final decision.Dude is salty.
I bet one arm that if he was still in OpenAI's board (he jumped out of the company some years ago) he wouldnt be saying this
We have some long years ahead until we see an Ultron. For starters, GPT is not connected to the internet. It's trained offline.
You cant fight against progress. They may suspend it to the public, but companies will keep evolving it on the background. And people are already sharing code all over the internet.
The genie is out of the bottle. Its over.
Its like this all over again
Also with Llama having leaked, you have a chatgpt3.5 level model essentially out in the wild being trained and potentially connected to the net by various individuals.
But these models are still too weak to be any serious threat.
Vestal
Junior Member
The code "CAN" be perfect. But you still need to know how to construct the prompt correctly and how the logic works.. Ive seen quiet a few times that while the code looks good, the ordering of the code can be a bit off.
PSYGN
Member
Must be the damn best spaghetti in the world!
HoodWinked
Member
what the actual fuck... Guy takes dog to Vet, dog doesn't get better, puts blood tests into chatGPT gets correct diagnosis. goes to 2nd vet, dog cured.
Kenpachii
Member
Interesting article
Elon Musk among experts urging a halt to AI training
Elon Musk among those calling for training in powerful artificial intelligence to be suspended.www.bbc.co.uk
I concur.
We need AI and we needed it a 100 years ago.
We need far faster progress in basically everything if we want to get anything done in a single lifetime. Humans are far to limited in making progress.
We need a huge jump into space travel, we need a huge jump into medicine and medicale solutions, we need a huge jump in renewable energy etc etc.
A person like elon musk should know that without AI progress is going to stand still for many generations.
Now the question is is it real AI or is it just a advanced google search machine what we got right now. That's good at fooling you. Or are we actually going to see some progress and jumps in the coming decades.
Last edited:
clem84
Gold Member
I just want a learning assistant. I don't care if they halt progress on more advanced AI, there's no reason to block learning assistants. A voice controlled AGI that can understand and respond well, and fast, will revolutionize learning. Learning languages, math, economy, law...
It's an exciting time to be alive.
It's an exciting time to be alive.
Last edited:
StreetsofBeige
Gold Member
Probably because he's got a team working on AI too and doesn't want to be left in the dust.Interesting article
Elon Musk among experts urging a halt to AI training
Elon Musk among those calling for training in powerful artificial intelligence to be suspended.www.bbc.co.uk
I concur.
Vestal
Junior Member
Yes and no I guess? He may have ulterior motives, but the letter itself is grounded in some pretty good arguments that should be discussed. I don't want AI advancement to stop, but I can't deny the geo-political implications of AI advancement reaching a certain threshold. There needs to be open discussion on AI, its uses and ethics, along with continued research into the inner workings of the neural network.
The problem is. Unless everyone pauses Model advancement, no one will.
Raonak
Banned
In a world of constantly overstressed and overworked humans, AI is gonna be a godsend. I can understand the fear of all jobs going extinct and stuff like that.
But the reality is this is gonna make it so much easier to learn... Literally anything.
The ability to ask an infinite amount of questions and clarifications for anything is gonna be revolutionary for learning. We have a massive teacher shortage with 1 teacher having to teach 40 kids. Augmenting a classroom with AI chatbots that can be used by each student simultaneously, used to clarify any questions they are too self conscious to ask a teacher about.
In terms of other jobs, I keep thinking how every time we employ a new developer. It takes so much of my time to get them up to shape. Fresh employees are so fucking hard to train and plan for. But with AI we can train them up so much faster and allow them to become so much more productive.
Yeah. The world will completely change.
But the way the world is going, it NEEDS to change.
But the reality is this is gonna make it so much easier to learn... Literally anything.
The ability to ask an infinite amount of questions and clarifications for anything is gonna be revolutionary for learning. We have a massive teacher shortage with 1 teacher having to teach 40 kids. Augmenting a classroom with AI chatbots that can be used by each student simultaneously, used to clarify any questions they are too self conscious to ask a teacher about.
In terms of other jobs, I keep thinking how every time we employ a new developer. It takes so much of my time to get them up to shape. Fresh employees are so fucking hard to train and plan for. But with AI we can train them up so much faster and allow them to become so much more productive.
Yeah. The world will completely change.
But the way the world is going, it NEEDS to change.
Dirk Benedict
Member
AIDS Shrrek, Mickey, Will Smith and Obama+ Mario
Elon Musk says "Hold-a-My-Pasta-Bowl" to Will Smith
Elon Musk says "Hold-a-My-Pasta-Bowl" to Will Smith
Last edited: