GitHub's Devin Competitor, Sam Altman Talks GPT-5 and AGI, Amazon Q, Rabbit R1 Hacked (AI News)

Summary

The video script discusses the latest advancements in AI coding assistance, highlighting GitHub Copilot Workspace and Amazon Q as examples of AI's growing capabilities in software development. GitHub Copilot Workspace is an AI coding assistant that can generate entire methods or pages of code, while Amazon Q is a generative AI assistant for businesses and developers, offering code generation, testing, debugging, and more. The script also touches on concerns about code quality due to AI-generated code reuse and the potential for AI to democratize software development. Additionally, it covers the Rabbit R1 AI box, Sam Altman's insights on building large-scale AI infrastructure, the challenges of productizing advanced AI, and the subtle dangers of AI. The script concludes with the first full music video generated by Sora, an AI text-to-video product, and a mention of Open Voice V2, an open-source voice cloning project.

Takeaways

• 🚀 **GitHub Co-pilot Workspace**: An evolution of the original GitHub Co-pilot, offering a more comprehensive tool that not only auto-completes code but also helps to define tasks, create plans, and generate code based on user specifications.
• 🔍 **Code Quality Concerns**: There's a critique that GitHub Co-pilot may lead to suboptimal code reuse and a potential degradation in the median baseline of code quality due to the model being trained on reused code.
• 🤖 **GitHub Co-pilot Chat**: A less impressive feature compared to the auto-complete tool, described as similar to Chat GPT but for code, which didn't seem to offer significant advantages.
• 📈 **Developer Productivity**: GitHub Co-pilot is reported to have boosted developer productivity by up to 55%, highlighting the potential impact of AI on software development.
• 🔧 **Control Over Process**: GitHub Co-pilot Workspace gives developers full control over the development process, from brainstorming to testing, while assisting with natural language inputs.
• 📱 **Mobile Compatibility**: GitHub Co-pilot Workspace is mobile compatible, allowing developers to contribute to code bases from their mobile devices, offering unprecedented flexibility.
• 🛒 **Amazon Q**: Amazon's generative AI assistant aimed at businesses and developers, capable of generating code, testing, debugging, and providing multi-step planning and reasoning for implementing new code.
• 📊 **Amazon QuickSight**: A business intelligence tool that can analyze business data, with Amazon Q's ability to dynamically generate reports and metrics based on user descriptions.
• 📈 **AI in Business**: Amazon Q positions itself as a business product, enabling employees to use natural language to build their own generative AI applications from company data.
• 🧱 **Rabbit R1 AI Box**: The Rabbit R1 has improved significantly since its release, particularly in terms of battery life, and it has been revealed to be running an Android app.
• 🌟 **GPT Models and AGI**: Discussion around the current state of GPT models, with GPT-4 considered a significant step forward, and the challenges of productizing PhD-level intelligence.
• 🎥 **First Fully Generated Music Video by Sora**: A milestone in AI where a full music video was created using Sora, OpenAI's text-to-video product, showcasing the potential of generative AI in creative fields.

Q & A

What is GitHub Copilot Workspace?

GitHub Copilot Workspace is an evolution of the GitHub Copilot project, which is an AI coding assistant. It allows developers to define a task, such as solving a GitHub issue or defining a new feature, and then it builds the entire solution for the developer, including code generation and specification detailing.

What is a concern with AI coding assistants like GitHub Copilot?

One critique is that developers might reuse code without thoroughly reviewing it, which could be suboptimal or contain bugs. This repeated use of unreviewed code could potentially degrade the overall quality of code in the model's training data.

What is Amazon Q?

Amazon Q is a generative AI-powered assistant for businesses and developers. It is designed to assist with tasks such as coding, testing, and upgrading applications, as well as troubleshooting, performing security scanning, and optimizing AWS resources.

How does Amazon Q Developer assist with code transformation?

Amazon Q Developer can dynamically generate interfaces for code upgrades, such as converting code into Java 8, providing a tool for developers to transform and upgrade their codebases.

What is the purpose of Amazon QuickSight and Amazon Q?

Amazon QuickSight is a business intelligence tool that allows users to analyze business data easily. Combined with Amazon Q, it can generate reports with metrics and graphs based on abstract descriptions provided by the user.

What is the Rabbit R1 AI box, and what has been revealed about it?

The Rabbit R1 AI box is a device that has improved significantly since its release, particularly in terms of battery life. It has been revealed to be running an Android app, and the Rabbit app has been made functional on Android phones.

What are some concerns raised by Sam Altman about AGI?

Sam Altman, the CEO of OpenAI, expresses concerns about the subtle dangers of AGI, such as its potential to cause pervasive but less noticeable changes in society, like increased attention disorders. He also discusses the challenge of productizing PhD-level intelligence and the importance of iterative deployment of AI technologies.

What does Sam Altman think about the current state of models like GPT-4?

Sam Altman considers models like GPT-4 to be 'embarrassing at best,' suggesting that there are more advanced models within OpenAI that have not been publicly disclosed.

What is the significance of the first full music video generated by Sora?

The first full music video generated by Sora, a text-to-video product from OpenAI, signifies a major advancement in AI-generated multimedia content. It demonstrates the potential for AI to create complex, creative works.

What is Open Voice V2 and how does it work?

Open Voice V2 is an open-source voice cloning project similar to 11 Labs. It allows users to create voice clones that can replicate any voice in any style and language, raising ethical concerns about the use of AI in voice replication.

What is the ethical concern surrounding AI and societal impact?

The ethical concerns surrounding AI and its societal impact include issues related to privacy, consent, the potential for misuse, and the broader implications of AI on employment, social interaction, and the nature of human intelligence and creativity.

Outlines

00:00

🚀 GitHub Co-Pilot Workspace: AI-Powered Coding Assistant

GitHub Co-Pilot Workspace is an advanced AI coding assistant that builds upon the original GitHub Co-Pilot project. It allows developers to define a task, such as solving a GitHub issue or creating a new feature, and then generates a comprehensive plan and code for the task. The tool claims to boost developer productivity by up to 55%. It also addresses concerns about code reuse and quality by allowing developers to review and edit the AI-generated code. The workspace offers a natural language interface for planning, building, testing, and running code, with full control over each step of the process. It is designed to augment, not replace, developer creativity and is mobile compatible, enabling coding from anywhere.

05:01

🤖 Amazon Q: Generative AI for Business and Development

Amazon Q is a generative AI assistant aimed at businesses and developers, offering capabilities in code generation, testing, debugging, and planning. It comes in four different use cases: Amazon Q Developer, which assists with coding, testing, and application upgrades; Amazon Q Business, which uses a large language model for tasks like generating job postings; Amazon Quick Site, a business intelligence tool that analyzes business data; and Amazon Q Apps, which allows employees to build simple AI applications using natural language. The tool is positioned as a business product that can transform and implement new code from developer requests and make it easier for employees to access business data.

10:03

📱 Rabbit R1 AI Box and Android App Integration

The Rabbit R1 AI box has been revealed to be based on an Android operating system, and it has been demonstrated that the Rabbit App can run on an Android phone. This suggests that the company's hardware could potentially be replicated on other Android devices. Despite this, the form factor and hardware of the Rabbit R1 are appreciated, and there is a suggestion that an app version could be developed without cannibalizing the hardware market. The device's battery life has seen significant improvement with updates, making it more usable.

15:04

🧠 GPT's Future, AGI, and Societal Impact

Sam Altman, CEO of OpenAI, discusses the future of AI, including the challenges of building very large computers and the commoditization of AI models. He emphasizes the importance of iterative deployment of AI technology, allowing society to adapt and co-evolve with the technology. Altman also expresses concerns about the rate at which society can adapt to new AI technologies and the potential for subtle, overlooked dangers. He mentions that while GPT-4 is considered 'embarrassing' in comparison to future models, it is still important to release and improve upon such models. Additionally, the first full music video generated by Sora, an OpenAI text-to-video product, is highlighted, showcasing the potential of AI in creative fields.

Keywords

💡AI coding assistant

AI coding assistants are tools that utilize artificial intelligence to assist in the process of coding by automating tasks such as writing or suggesting code. In the video, GitHub Co-pilot is highlighted as an example of an AI coding assistant that can generate entire methods or pages of code, significantly boosting developer productivity.

💡GitHub Co-pilot workspace

GitHub Co-pilot workspace is an evolution of the original GitHub Co-pilot project. It is designed to help developers define tasks, create plans, and generate code in a more comprehensive and task-centric manner. The video emphasizes how it allows developers to start from a high-level task definition and systematically build out a full plan, including code generation and testing.

💡Code quality

Code quality refers to the degree to which source code meets the desired criteria for operability, maintainability, and efficiency. The video script discusses a critique of GitHub Co-pilot, where the reuse of code without thorough review could potentially lower the median baseline of code quality due to the propagation of suboptimal or buggy code.

💡Amazon Q

Amazon Q is a generative AI-powered assistant aimed at businesses and developers. It is designed to not only generate highly accurate code but also to test, debug, and implement new code based on developer requests. The video mentions Amazon Q as an alternative to GitHub Co-pilot, suggesting its capabilities in transforming code and assisting with business intelligence tasks.

💡Natural language processing (NLP)

Natural language processing (NLP) is a field of AI that focuses on the interaction between computers and humans through natural language. The video describes how GitHub Co-pilot workspace allows developers to brainstorm, plan, build, test, and run code using natural language, indicating the application of NLP in facilitating a more intuitive and efficient coding process.

💡Developer productivity

Developer productivity refers to the efficiency and effectiveness with which a developer can create and maintain code. The video script notes that GitHub Co-pilot has been reported to boost developer productivity by up to 55%, highlighting the impact of AI tools on the software development process.

💡Large language models (LLMs)

Large language models are AI models that are trained on vast amounts of text data to understand and generate human-like language. The video discusses Amazon's attempt at creating its own large language model with Amazon Q and how it is used in different contexts, such as generating job postings and transforming code.

💡AGI (Artificial General Intelligence)

Artificial General Intelligence (AGI) refers to highly autonomous systems that can outperform humans at most economically valuable work. The video touches on the challenges of productizing PhD-level intelligence and the societal impact of AGI, indicating the ongoing discourse and development in the field of AI.

💡Ethical concerns

Ethical concerns in AI pertain to the moral implications of deploying intelligent systems that may affect human life, privacy, and autonomy. The video script briefly mentions ethical concerns surrounding AI, suggesting the need for responsible development and deployment of AI technologies.

💡Sora

Sora is a text-to-video product developed by OpenAI that can generate full music videos from textual descriptions. The video script describes Sora as a groundbreaking tool that signifies the potential of AI in creative fields, showcasing the ability to create complex multimedia content automatically.

💡Open Voice V2

Open Voice V2 is an open-source voice cloning project that allows users to generate speech in any style and language. The video presents it as an example of AI's growing capabilities in mimicking and generating human-like voices, raising both exciting possibilities and potential ethical considerations.

Highlights

GitHub Co-pilot Workspace is an evolution of the AI coding assistant that can define tasks and build entire code structures.

GitHub Co-pilot has been reported to boost developer productivity by up to 55%.

Critiques of GitHub Co-pilot include concerns over code quality due to the reuse of potentially suboptimal code.

GitHub Co-pilot Chat, a code-based version of Chat GPT, was released in 2023 but is not as powerful as the autocomplete feature.

Developers can now use natural language to brainstorm, plan, build, test, and run code with GitHub Co-pilot Workspace.

GitHub Co-pilot is designed to augment, not replace, developers, requiring human review of generated code and specifications.

Amazon Q is a generative AI assistant aimed at businesses and developers, offering code generation, testing, and debugging.

Amazon Q comes in four different use cases, including developer assistance, business intelligence, and secure AI application building.

Amazon QuickSight, combined with Amazon Q, enables the creation of comprehensive business reports with ease.

Rabbit R1 AI box was found to be based on Android, with the Rabbit App successfully run on an Android phone.

Sam Altman, CEO of OpenAI, discusses the challenges of building supercomputers and the commoditization of AI models.

Altman expresses concerns about the pace of societal adaptation to advanced AI and the development of a new social contract.

OpenAI's GPT-4 is considered a significant step forward, though described by Altman as 'embarrassing' in comparison to future models.

The first full music video completely generated by Sora, OpenAI's text-to-video product, has been released.

MyShell AI released Open Voice V2, an open-source voice cloning project that can replicate voices in various styles and languages.

Ethical concerns and societal impacts of AI are discussed, emphasizing the importance of gradual AI deployment and public input.

The potential dangers of AGI are considered, with a focus on subtle, overlooked risks rather than dramatic, obvious threats.

Transcripts

so it seems like now the major tech

companies are rolling out their AI

coding assistance and they are extremely

comprehensive and very impressive so we

have two examples of that we have one

from GitHub which is Microsoft and they

call it GitHub co-pilot workspace and

then we have Amazon's version which is

Amazon Q which they just rolled out plus

we have a few other news stories that

we're going to get to today so let's get

into it first this is GitHub co-pilot

workspace and this is the evolution of

the GitHub co-pilot project which was

really the first AI coding assistant and

it was incredible you just start typing

some code and it fills out the entire

method or even entire pages of code for

you and now GitHub co-pilot workspace

takes it a step further you start from

the task you start by defining what you

want to accomplish whether that's

solving a GitHub issue or defining a new

feature and it builds the entire thing

for you it's really the next generation

of AI coding assistants so this is the

blog post about it right here they say

in 2022 we launched GitHub co-pilot as

an autocomplete pair programmer in the

editor boosting developer productivity

by up to 55% now one critique of GitHub

co-pilot is that you are reusing code

that you're not really even looking at

and it's potentially suboptimal code or

even code that might have bugs in it and

so many people are reusing the same code

over and over again and all of that is

going back into the model to train it

maybe we're actually degrading the

median Baseline of code code quality

which is something interesting to think

about uh then in 2023 we released GitHub

co-pilot chat which is essentially like

chat GPT but with code honestly it's not

super impressive I never used it it's

kind of just like taking your code based

putting it into chat GPT and then asking

questions about it and it's not even

that good it's certainly not nearly as

powerful as GitHub co-pilot autocomplete

now we have GitHub co-pilot workspace

developers can now brainstorm plan build

test and run code in natural language

this new task Centric experience

leverages different co-pilot powered

Agents from start to finish while giving

developers full control over every step

of the process this is essentially

Microsoft's version of Devin all right

so we have an example of how it works

what we're seeing here is an issue a

GitHub issue create an AI player option

and then we have a description of what

it should be so then we have that we can

open it in workspace then the AI

actually fills out a much more detailed

specification for this issue and then

what we can do is click this generate

plan button we can edit any of the

details along the way as it says here we

can add parts to the spec it's now

generating a plan and then we can

actually just click Implement and it's

going to show us the diff right here so

all we had to do was describe what we

want to build and then it built it with

the diff in a really nice interface and

it's going to go step by step and

accomplish every one of the

specifications in the issue or in the

plan then we can actually open it in a

live preview so you can test it out

before merging the code there it is so

all within this one environment and then

we can actually create the poll request

and merge the code so very very cool and

I think what's interesting about GitHub

co-pilot is they're not claiming to

replace developers they are claiming to

augment developers and I really actually

believe that with this purpose you still

need to review the code you still need

to review the specifications but I think

this is going to give many more people

the ability to contribute to code bases

and right here it says and is expressly

designed to deliver not replace

developer creativity faster and easier

than ever before so again it all starts

with a task then we build a full plan

and it actually builds it for us and we

can edit it then we can create the code

and again completely editable then we

test out the code we actually see it

running and then we can create a pull

request and the code so let me show you

a video of it working end to endend and

this is from the GitHub co-pilot team so

let's watch this video together it's

about 2 minutes getting started with

your work usually takes place looking at

a project board and navigating to GitHub

issues co-pilot workspace brings your

favorite AI assistant into a native Dev

environment designed for everyday tasks

for example it can use the information

in your GitHub issue along with

references from your repository to build

out a specification based on the current

State and proposed state and you can

tailor the spec as you need whether

that's adding editing or removing items

once ready you can progress from spec to

plan and the process feels familiar with

the ability to adjust the plan as you

need by creating modifying or removing

files and adjusting the expected tasks

for each step this leaves you in control

free to solve the higher level problems

and build out your plan before getting

into the finer details of writing code

co-pilot workspace then streams the

suggested changes to our environment

notice that we have a diff view to

easily digest the changes and can easily

make updates within the editor but

that's not all at our fingertips we have

access to an integrated terminal so we

can run the test in our workspace before

committing our changes and creating a

pull request and what if you want to

make use of advanced features like step

through debugging no problem create a

code space and pick up where you left

off when you raise your pull request

with co-pilot workspace it generates a

description for you and automatically

adds a link to your workspace adding a

little bit more context for the reviewer

and supporting their code review

workflow and as it's a pull request our

usual checks trigger including GitHub

action workflows and code scanning

co-pilot workspace leaves you in control

solving the higher level problems and

iterating quickly over the solution and

and it is mobile compatible so you're

going to be able to do all of this from

your mobile phone which is kind of

insane to think about you can be coding

and actually contributing real code to a

project from your phone from anywhere in

the world next we have Amazon q and

Amazon Q is a generative AI powered

assistant for businesses and developers

and is now generally available now I

haven't played with it yet but what's

interesting is this is Amazon's attempt

at their own large language model and I

think they rolled their own but they

don't really give any specifications

about it and it comes in four flavors

and I'm going to talk about that in a

minute so a little bit about it Amazon Q

not only generates highly accurate code

it also tests debugs and have multi-step

planning and reasoning capabilities that

can transform and Implement new code

generated from developer requests Amazon

Q also makes it easier for employees to

get answers to questions across business

data such as company policies Etc so

again four different use cases and we're

going to take a look at each so first is

Amazon Q developer and it's very similar

to what we just looked at Amazon Q

assists developers and it professionals

with all of their tasks from coding

testing and upgrading applications to

troubleshooting performing security

scanning and fixes and optimizing AWS

resources so let's take a look at a

quick demo video and I'll walk you

through what we're seeing so first is

Amazon QBE business and really what this

is is just a large language model so

write a job posting this is essentially

a prompt you can put into any large

language model and it generates a job

posting not super impressive now what

I'm really interested in is Amazon Q

developer so you can ask it questions

and what we're seeing here is

transforming code so it's actually

converting it into Java 8 for example it

doesn't seem quite as impressive as

GitHub copilot workspace but that's

pretty cool it dynamically generated or

I believe it dynamically generated that

interface to upgrade and here we go so

it's doing that upgrade now now we have

Amazon q and Amazon quick site and quick

site by the way is just a business

intelligence tool that just means you

dump a bunch of your business data in

there and you can analyze it really

easily and I think this is a really

underappreciated use case especially as

a business you need data analysts data

scientists to do this work but now you

can essentially just describe the types

of metrics that you want to see and

Amazon quick site with Amazon Q is going

to put it all together for you so here

we go build a story explaining

profitability Trends and again you can

be very abstract like that and it puts

together an entire report with graphs

and everything that's super super

impressive and look at all these

suggestions of different metrics you can

run all generated by AI very very cool

then they have Amazon Q apps which says

it enables employees to use natural

language to securely build their own

generative AI applications and it's all

from the company's data so Amazon Q is

really positioned to be the business

product so employees simply describe the

type of app they want in natural

language and Q apps will quickly

generate an app that accomplish is their

desired task so let's look at this quick

example build an app that takes a ro

title and onboarding plan structure

weekly 6090 day and optionally a new

hire's name it should output a summary

of the role guidelines for the given

role so it's essentially using AI to

build really simple applications to be

used within the work environment and so

you generate these applications and then

they can be used over and over again

kind of cool I've not really seen this

elsewhere uh it's kind of neat and so

that's all for Amazon Q I haven't tested

it myself it looks pretty cool though so

I might go and test it next rabbit R1 so

I have my rabbit R1 right here it's

actually gotten much better since I've

gotten it one of my biggest complaints

was the fact that the battery drained

really quickly and then with one update

it improved probably about 10x for me so

where previously I could literally stare

at the device and watch the battery

drain percent by% now it lasts for 2

hours and only drains about 5% of

battery so a vast Improvement and really

it actually makes it usable now but what

now RS Technica has broke is that the

rabbit R1 AI box is revealed to be just

an Android app however it's not really

breaking anything we kind of knew that

they announced that the underlying

operating system they announced it a

while ago is Android and so yeah of

course it's just kind of an Android app

and so what somebody was actually able

to do is get the rabbit App working on

an Android phone which is pretty

interesting and so here we can actually

see the rabbit App working on a phone

that is the rabbit app and so the

company posted on X that they are aware

there are some unofficial rabbit OS app

website emulators out there but look of

course it can be an app I like the form

factor I like the hardware it's gorgeous

to me it feels nice I know I'm a sucker

for this kind of stuff but I really

appreciate it and of course I don't know

why they're not just also building an

app I I guess that would cannibalize

their Hardware Market but for $200 it

seems worth it to me and of course if

you don't want to buy it just don't buy

it and if you want to load it up on your

own device in a kind of hacked way

you're welcome to do that of course next

Sam Alman said some very interesting

things about GPT 5 and GPT 4 so he talks

about a lot of stuff at this Stanford

talk so let's go through a few of them

one he is asked like what he thinks

about right now and his answer is how to

build really big computers and we

already know he was trying to raise 78

trillion dollar we already know that

Microsoft is building the biggest

supercomputer ever made on behalf of

really open AI so he is thinking about

this deeply I've also said multiple

times models are becoming commoditized

that is not where the value is going to

be what questions then are you wrestling

with that no one else is talking about

how to build really big computers I mean

I think other people are talking about

that but we're probably like looking at

it through a lens that

no one else is quite imagining yeah and

can we continue on that thread then of

how to build really big computers if

that's really what's on your mind can

you share I know there's been a lot of

speculation and probably a lot of here

say too about the semiconductor Foundry

Endeavor that you are reportedly

embarking on um can you share what would

make what what's the vision what would

make this different than others that

just foundies although that that's part

of it it's like if if you believe which

we increasingly do at this point that AI

infrastructure is going to be one of the

most important inputs to the Future this

commodity that everybody's going to want

and that is energy data centers chips

chip design new kinds of networks it's

it's how we look at that entire

ecosystem um and how we make a lot more

of that and I don't think it'll work to

just look at one piece or another but we

we got to do the whole thing okay so

there's multiple big

problems yeah um I think like just this

is the Arc of

human technological history as we build

bigger and more complex systems another

thing he talks about and I'll show you

the clip in a second is the fact that

they're not sure how to productize PhD

level intelligence and how to actually

deliver the most value to the world

which is kind of interesting they're

kind of building the plane in the air

while flying it which makes me a little

bit nervous but I tend to be an optimist

so I have some good outlook on what the

future of AI will be I mean we're we're

definitely

wrestling with how we when we make not

just like grade schooler or middle

schooler level intelligence but like PhD

level intelligence and Beyond the best

way to put that into a product the best

way to have a positive impact with that

on society and people's lives we don't

know the answer to that yet so I think

that's like a pretty important thing to

figure out and he also talks about AGI

the AGI timeline how gp4 is embarrassing

at best which is interesting it kind of

leads me to believe he's seen something

much better internal at open AI let's

watch what he has to say about that open

ey is phenomenal chat gbt is phenomenal

um everything else all the other models

are

phenomenal it burned you've earned $520

million of cash last year that doesn't

concern you in terms of thinking about

the economic model of how do you

actually where's going to be the

monetization source well first of all

that's nice of you to say but Chachi PT

is not phenomenal like Chachi PT is like

mildly embarrassing at best um gp4 is

the dumbest model any of you will ever

ever have to use again by a lot um but

you know it's like important to ship

early and often and we believe in

iterative deployment like if we go build

AGI in a basement and then you know the

world is like kind

of blissfully walking blindfolded along

um I don't think that's like I don't

think that makes us like very good

neighbors um so I think it's important

given what we believe is going to happen

to express our view about what we

believe is going to happen um but more

than that the weight to do it is to put

the product in people's hands um and let

Society co-evolve with the technology

let Society tell us what it collectively

and people individually want from the

technology how to productize this in a

way that's going to be useful um where

the model works really well where it

doesn't work really well um give our

leaders and institutions time to react

um give people time to figure out how to

integrate this into their lives to learn

how to use the tool um sure some of you

all like cheat on your homework with

with it but some of you all probably do

like very amazing wonderful things with

it too um and as each generation goes on

uh I think that will expand and that

means that we ship imperfect products um

but we we have a very tight feedback

loop and we learn and we get better and

it does kind of suck to ship a product

that you're embarrassed about but it's

much better than the alternative um and

in this case in particular where I think

we really owe it to society to deploy

iteratively one thing we've learned is

that Ai and surprise don't go well

together people don't want to be

surprised people want a gradual roll out

and the ability to influence these

systems um that's how we're going to do

it there could totally be things in the

future that would change where we think

iterative deployment isn't such a good

strategy but it does feel like the

current best approach that we have and I

think we've gained a lot um from from

doing this and you know hopefully s the

larger world has gained something too

now here's a clip where he talks about

some of the dangers of AGI and I'll play

that quickly do you think the biggest

Danger from AI is going to come from a

cataclysmic event which you know makes

all the papers or is it going to be more

subtle and pernicious sort of like you

know like how everybody has ADD right

now from you know using Tik Tok um is it

are you more concerned about the subtle

dangers or the cataclysmic dangers um or

neither I'm more concerned about the

subtle dangers because I think we're

more likely to overlook those the

cataclysmic dangers uh a lot of people

talk about and a lot of people think

about and I don't want to minimize those

I think they're really serious and a

real thing but I think we at least know

to look out for that and spend a lot of

effort um the example you gave of

everybody getting addd from Tik Tok or

whatever I don't think we knew to look

out for and the unknown unknowns are

really hard and so I'd worry more about

those although I worry about both and

are they unknown unknowns are there any

that you can name that you're

particularly worried about well then I

would kind of they'd be unknown unknown

um well you can I am worried just about

so so even though I think in the short

term things change less than we think as

with other major Technologies in the

long term I think they change more than

we think and I am worried about what

rate Society can adapt to something so

new and how long it'll take us to figure

out the new social contract versus how

long we get to do it um I'm worried

about that so he actually has an entire

talk it's really interesting I'll drop

the link to that in the description

below I might actually make a full video

just on the talk because he talks about

a lot of interesting stuff let me know

if you want to see that all right next

an article from Venture beat but really

the gist is we have our first full music

video completely generated by Sora and

if you haven't heard of Sora it is the

text to video product that open AI kind

of previewed a few months back really

mind-blowing stuff let me show you some

brief Clips I'm not sure if I'm going to

get copyrighted for the music so I'm

just going to replace it for some

different music so it might seem a

little off but let me just show you the

Sora part cuz that's the interesting

part

Karma a think for me we hold hands make

plans keep seats Karma she's my best

friend One Step Closer here it ends get

a thing for me you see you be digging up

6et Karm never let me down my when I cut

[Music]

you and The Last Story just a quick one

my shell AI released open Voice V2 which

is an open Voice cloning project kind of

like 11 Labs but open source and

apparently it works really well let's

look at the video demo open Voice

instantly call any boys through generous

speech in any style at any

langage we are conflating a lot of

things with the word

intelligence ethical concerns surround

AI societal impact 3 2 1 ethical

concerns surround AI societal impact

wanders and watches with eager ears the

aroma of freshly baked bread filled the

kitchen the aroma of freshly baked bread

filled the kitchen the aroma of freshly

baked bread filled the kitchen we were

conflating a lot of things with the word

intelligence

restly so that works really well if you

want to see a tutorial for how to get

that installed and how to use it let me

know in the comments if you liked this

video please consider giving a like And

subscribe and I'll see you in the next

one