The bing version is still the search engine version. It does seem like if you wanna play around with true GPT4 chat you'd need to get the sub.
Don't think I can upload any image at the momentSo who here has Chat GPT4 to get its review on the $30 Italian lunch or that sad NY BBQ plate?
oo, that's a great use for it. will have to try it out.I use ChatCPT all day, it's actually a invaluable resource for coding, and providing insight into other people's code. It's part of my workflow.
It's like having an extremely knowledgeable companion, I can't go back.
IDE on one screen, ChatGPT in another. I get it to write my regex for instance. Unbelievable stuff.
Only some AP. It did terrible on quite a few
For context, GPT 3.5 scored in the bottom 10% for a simulated bar exam, while 4 gets in the top 10%.
Its facility with language (outperforming state-of-the-art English models when given tests translated into other languages) is impressive.
I'm most interested in what the larger context window unlocks. 32k tokens is about 50 pages, imagine something this powerful maintaining coherency for that long. You could auto-summarize long essays, synthesize novellas, generate more than just toy programs. And it's surprisingly affordable.
In only half a generation it went from a 1 on AP Calc to a 4. That massive leap is stunning.Looking at where it performs best, it works best in known problem spaces (simple standardized tests) and performs worst on things requiring awareness and problem solving (see writing only being 54th percentile and USNCO local section and AP calculus)
It would require exact coordination between multiple countries and massive corporations with big profit incentives so basically no, unfortunately not lol
Ah, jesus. Welp here we go!Huge. MS invested $10 billion in OAI (49% stake iirc), custom-built them a supercomputer for training, and have exclusive access to the codebase to build products like Bing Chat and various Windows integrations that need more than just the API.
I need this stuff officially supported on digital voice assistants like Google Assisstant
Yeah and the fact that Microsoft laid off their entire AI Ethics department the day before they announced GPT-4 does not instill confidenceUS Dept of Commerce recently released a report estimating that "virtually every business and government agency" will be using generative AI technology in the next 10-20 years, which may well be a pretty conservative estimate. This thing is moving pretty quickly.
Why they chose this as an example of the new model's impressive capabilities is beyond me
View: https://twitter.com/Confessant/status/1635699468531281920
Is this available to the public now? I have been using the shit out of bing chat gpt while coding.
Just watched the livestream demo. Dude doodled a website mockup in a notebook, took a photo, submitted it, and got working HTML code back.
edit:
View: https://www.youtube.com/live/outcGtbnMuQ?t=1h10m40s
It can be a bit buggy but it can also save a lot of time I think. I definitely cross-reference its answers if it's something I don't already know. I think it makes sense if you're using it for pretty boilerplate stuff that you'd just spend time writing out anyway.I'm surprised at all the people saying they use it to code. In my experience like half of it is wrong, maybe another 25% is suboptimal and the rest is correct and usable. It has given me some good answers, but most of the time it's faster to search on Stack Exchange.
provider "aws" {
region = "us-west-2"
}
locals {
container_port = 80
}
resource "aws_ecr_repository" "example" {
name = "example-repo"
}
resource "aws_ecs_cluster" "example" {
name = "example-cluster"
}
resource "aws_iam_role" "ecs_execution_role" {
name = "ecs_execution_role"
assume_role_policy = jsonencode({
Version = "2012-10-17"
Statement = [
{
Action = "sts:AssumeRole"
Effect = "Allow"
Principal = {
Service = "ecs-tasks.amazonaws.com"
}
}
]
})
}
resource "aws_iam_role_policy_attachment" "ecs_execution_ecr_policy" {
policy_arn = "arn:aws:iam::aws:policy/AmazonEC2ContainerRegistryReadOnly"
role = aws_iam_role.ecs_execution_role.name
}
resource "aws_ecs_task_definition" "example" {
family = "example-task-family"
requires_compatibilities = ["FARGATE"]
network_mode = "awsvpc"
cpu = "256"
memory = "512"
execution_role_arn = aws_iam_role.ecs_execution_role.arn
container_definitions = jsonencode([
{
name = "example-container"
image = "${data.aws_caller_identity.current.account_id}.dkr.ecr.${var.region}.amazonaws.com/${aws_ecr_repository.example.name}:latest"
portMappings = [
{
containerPort = local.container_port
hostPort = local.container_port
}
]
}
])
}
data "aws_caller_identity" "current" {}
resource "aws_vpc" "example" {
cidr_block = "10.0.0.0/16"
tags = {
Name = "example-vpc"
}
}
resource "aws_subnet" "example_1" {
vpc_id = aws_vpc.example.id
cidr_block = "10.0.1.0/24"
tags = {
Name = "example-subnet-1"
}
}
resource "aws_subnet" "example_2" {
vpc_id = aws_vpc.example.id
cidr_block = "10.0.2.0/24"
tags = {
Name = "example-subnet-2"
}
}
resource "aws_security_group" "example" {
name = "example"
description = "Example security group for ECS service"
vpc_id = aws_vpc.example.id
}
resource "aws_security_group_rule" "example" {
security_group_id = aws_security_group.example.id
type = "ingress"
from_port = local.container_port
to_port = local.container_port
protocol = "tcp"
cidr_blocks = ["0.0.0.0/0"]
}
resource "aws_ecs_service" "example" {
name = "example-service"
cluster = aws_ecs_cluster.example.id
task_definition = aws_ecs_task_definition.example.arn
desired_count = 1
launch_type = "FARGATE"
Yes, so can GPT-3, it just may take multiple prompts.Can it create a executable programs instead of just code snippets?
Yeah, this is both incredible and horrifying at the same time.I am starting to feel a hint of anxiety about this. The rate of progression and the people at the helm are not inspiring any confidence in me.
I am starting to feel a hint of anxiety about this. The rate of progression and the people at the helm are not inspiring any confidence in me.
Yeah, this is both incredible and horrifying at the same time.
Fuck it. Let's go. To the bottom, to its nightmarish use in modern societies. I can't even fathom what unhinged capitalism will conjure up.
It can be a bit buggy but it can also save a lot of time I think. I definitely cross-reference its answers if it's something I don't already know. I think it makes sense if you're using it for pretty boilerplate stuff that you'd just spend time writing out anyway.
I did a test asking it to generate some terraform code for ECS and ECR, at a glance it seems like a time saver, though I'd def go in and adjust a lot of it.
Code:provider "aws" { region = "us-west-2" } locals { container_port = 80 } resource "aws_ecr_repository" "example" { name = "example-repo" } resource "aws_ecs_cluster" "example" { name = "example-cluster" } resource "aws_iam_role" "ecs_execution_role" { name = "ecs_execution_role" assume_role_policy = jsonencode({ Version = "2012-10-17" Statement = [ { Action = "sts:AssumeRole" Effect = "Allow" Principal = { Service = "ecs-tasks.amazonaws.com" } } ] }) } resource "aws_iam_role_policy_attachment" "ecs_execution_ecr_policy" { policy_arn = "arn:aws:iam::aws:policy/AmazonEC2ContainerRegistryReadOnly" role = aws_iam_role.ecs_execution_role.name } resource "aws_ecs_task_definition" "example" { family = "example-task-family" requires_compatibilities = ["FARGATE"] network_mode = "awsvpc" cpu = "256" memory = "512" execution_role_arn = aws_iam_role.ecs_execution_role.arn container_definitions = jsonencode([ { name = "example-container" image = "${data.aws_caller_identity.current.account_id}.dkr.ecr.${var.region}.amazonaws.com/${aws_ecr_repository.example.name}:latest" portMappings = [ { containerPort = local.container_port hostPort = local.container_port } ] } ]) } data "aws_caller_identity" "current" {} resource "aws_vpc" "example" { cidr_block = "10.0.0.0/16" tags = { Name = "example-vpc" } } resource "aws_subnet" "example_1" { vpc_id = aws_vpc.example.id cidr_block = "10.0.1.0/24" tags = { Name = "example-subnet-1" } } resource "aws_subnet" "example_2" { vpc_id = aws_vpc.example.id cidr_block = "10.0.2.0/24" tags = { Name = "example-subnet-2" } } resource "aws_security_group" "example" { name = "example" description = "Example security group for ECS service" vpc_id = aws_vpc.example.id } resource "aws_security_group_rule" "example" { security_group_id = aws_security_group.example.id type = "ingress" from_port = local.container_port to_port = local.container_port protocol = "tcp" cidr_blocks = ["0.0.0.0/0"] } resource "aws_ecs_service" "example" { name = "example-service" cluster = aws_ecs_cluster.example.id task_definition = aws_ecs_task_definition.example.arn desired_count = 1 launch_type = "FARGATE"
I guess there's a response limit, because it cut off the answer here.
I'm surprised at all the people saying they use it to code. In my experience like half of it is wrong, maybe another 25% is suboptimal and the rest is correct and usable. It has given me some good answers, but most of the time it's faster to search on Stack Exchange.