Chatgpt - Seriously good potential (or just some Internet fun)

I think, a subtle but important difference is that algorithms are directing you to people who can influence you. The algorithm itself is merely a middleman.

The people controlling the algorithms are directing people.

We're worried about AGI when humans have been shown to be utter prices of ****. I mean you don't have to look very far, this very forum...
 
The replicant test is for rooting out genetically engineered super human psychopaths, not artificial intelligence
artificial intelligence needs to exhibit elements of human behaviour before people would trust it and be influenced.

Italy says ban is over privacy concerns - can you use chatgpt with genuine anonymity ... or wil your boss tell you tomorrow we've had a tender for your job.
 
The replicant test is for rooting out genetically engineered super human psychopaths, not artificial intelligence
artificial intelligence needs to exhibit elements of human behaviour before people would trust it and be influenced.

It already does!

I'm still amused that artificially created intelligent beings aren't being seen as.... well artificial intelligence.

This is going to be a big mental block for some people as things progress, silicon vs biological etc.
 
I'm still amused that artificially created intelligent beings aren't being seen as.... well artificial intelligence.

So any child conceived through this artificial process, isn't human but is instead AI ? Because replicants are just a more advanced form of this so if they're not human, where do you draw the line on what is human and what is AI ?
 
Hmmm if only the movie had explored something like that...

I don't know why you're getting confused with IVF though, why wouldn't they be human?

A replicant is a genetically engineered humanoid with super intelligence, strength and a limited lifespan, how is it not artificial or not intelligent?
 
Last edited:
Chatgpt wrote me a bash script today that worked first time and does exactly what I want.

What does it do / how advanced?

I got invited to try Google's Bard - but I sat there last night thinking what could I ask it to do, and the only thing I came up with is to ask it to tell jokes.
 
What does it do / how advanced?

I got invited to try Google's Bard - but I sat there last night thinking what could I ask it to do, and the only thing I came up with is to ask it to tell jokes.
ChatGPT is pretty good at a variety of programming tasks. I've used it for generating systems in Unreal and implementing various algorithms.

It still needs an experienced programmer to tidy things up and refactor in places, but it's pretty solid.
 
ChatGPT is pretty good at a variety of programming tasks. I've used it for generating systems in Unreal and implementing various algorithms.

It still needs an experienced programmer to tidy things up and refactor in places, but it's pretty solid.

Interesting. Do you just give it a summary of what you want it to do, or is it a bit more involved?
 
What does it do / how advanced?
Pretty simple, it’s checking a web page for a specific string and if found, runs a couple of commands.

Interesting. Do you just give it a summary of what you want it to do, or is it a bit more involved?
“Write me a bash script for a pi that does [this]”

I’ve also had it write me a python script for the pi to trigger a relay and led on the GPIO pins. Not only did it make sure it imported all the appropriate libraries, it also told me what pins to connect the relay and led to.
 
The replicant test is for rooting out genetically engineered super human psychopaths, not artificial intelligence
artificial intelligence needs to exhibit elements of human behaviour before people would trust it and be influenced.

Italy says ban is over privacy concerns - can you use chatgpt with genuine anonymity ... or wil your boss tell you tomorrow we've had a tender for your job.
It stores user data so be careful what you put in there if it’s work sensitive stuff.
 
It stores user data so be careful what you put in there if it’s work sensitive stuff.

Yeah, that's one issue some firms are waking up to, not just the issue of what it stores but also, potentially, what it's producing and from where.

In some cases, if it's regurgitating large portions of someone else's IP which was in its training data then that could be iffy.
 
yes - that seems to be the concern for living artists/painters;

further - on the replicant test you'd really expect chatgpt to acknowledge the film reference when the blade runner questions are asked ... you quickly identify people from your tribe based on shared culture.

not sure - have they disclosed the hardware/cloud resources (and energy consumption) chatgpt currently has versus the human brain,
tesla reputedly binned off the custom hardware they were using for self drive learning , to use commodity graphics cards - not clear why?
 
tesla reputedly binned off the custom hardware they were using for self drive learning , to use commodity graphics cards - not clear why?

I would hazard a guess that it's cheaper and less faff to just use an off-the-shelf product. I know we've had similar discussions for the product we sell at work, years ago we were going down the route of a custom server build - which caused all sorts of headaches during manufacturing when it wasn't built correctly. Now we pretty much use a bog standard server so we don't have to worry about build issues.
 
I asked Bard to produce a greyscale height image of an area of terrain, and it said it could do it and upload the finished image to a Google Drive folder and send me a link.
24 hours later I asked it if it had finished, and it said it couldn't do it as it's only a text based chatbot.
 
Back
Top Bottom