Where We Go From Here with OpenAI's Mira Murati

Summary

In this insightful discussion, the guest shares their journey from Albania to working at OpenAI, emphasizing the importance of math and science in their upbringing. They delve into the challenges of building products with AI models, exploring the evolution from theoretical concepts to practical applications in fields like aerospace and automotive engineering. The conversation highlights the significance of AI safety, alignment with human values, and the potential of large language models to transform how we interact with technology, suggesting a future where AI systems could autonomously perform intellectual tasks, raising questions about reliability and the economic implications of AI advancements.

Takeaways

• 🌟 The speaker emphasizes the difficulty of building good products on top of AI models, suggesting that while model creation is a focus, the real challenge lies in application.
• 📚 Born in post-Communist Albania, the speaker's early education was heavily focused on math and sciences, which influenced his career path in mechanical engineering and AI.
• 🚀 With a background in mechanical engineering and aerospace, the speaker's interest in AI was sparked during his time at Tesla, particularly with the development of autopilot systems.
• 🕶️ The speaker explored augmented reality and virtual reality, noting the importance of understanding the practical applications and limitations of these technologies.
• 🧠 The shift in focus from theoretical knowledge to building practical applications is highlighted, with an emphasis on the generality of AI over domain-specific competence.
• 🔮 OpenAI's mission attracted the speaker due to its focus on building AGI (Artificial General Intelligence), which he sees as a fundamental and inspiring goal for technology development.
• 🤖 The speaker discusses the importance of AI safety, particularly the development of reinforcement learning with human feedback to align AI systems with human values.
• 🔗 The dialogue capability of AI models like Chat GPT is seen as a special form of interaction that allows for the expression of uncertainty and correction of errors, contributing to AI safety.
• 🛠️ The speaker suggests that we are at an inflection point in redefining human-digital interaction, moving towards a model where AI systems act more as collaborators than tools.
• 📈 The transcript discusses the concept of scaling laws in AI, indicating that there is still significant room for improvement as models scale up in terms of data and compute power.
• 💡 The future of AI is envisioned with models that encompass multiple modalities (text, images, video) for a more comprehensive understanding, akin to human perception.

Q & A

What was the educational focus in post-Communist Albania and how did it influence the speaker's interests?

In post-Communist Albania, there was a significant focus on math and physics, which influenced the speaker to develop a strong interest in these fields. The humanities were considered questionable due to the ambiguity of information sources and truthfulness, leading the speaker to pursue math and sciences relentlessly.

How did the speaker's career evolve from mechanical engineering to working with AI?

The speaker started as a mechanical engineer working in Aerospace and later joined Tesla, where they initially worked on Model S dual motor and then led the Model X program. This experience sparked an interest in AI applications, particularly autopilot, which led them to explore AI in different domains such as augmented reality, virtual reality, and eventually joining OpenAI to focus on general AI (AGI).

What was the speaker's initial attraction to OpenAI?

The speaker was drawn to OpenAI because of its mission, which they felt was more important than any other technology they could contribute to. They believed that building intelligence is a core unit that affects everything and is inspiring for elevating and increasing the collective intelligence of humanity.

How does the speaker view the role of physics and math backgrounds in the field of AI today compared to 15 years ago?

The speaker observes that many influential contributors to the AI space today have a physics or math background, which is a shift from 15 years ago when the field was more dominated by engineers from lexical, mechanical, and other engineering backgrounds.

What does the speaker mean by 'building an intuition' in the context of problem-solving in math and theoretical spaces?

The speaker refers to the process of deeply engaging with a problem over an extended period, often requiring time for reflection and incubation, which leads to the development of a new idea or solution. This process builds an intuition for identifying the right problems to work on and the discipline to persevere until a solution is found.

How does the speaker perceive the current state of AI systems in terms of their capabilities and limitations?

The speaker sees AI systems as having made significant progress, particularly in representation and understanding of concepts, but acknowledges that there are still substantial limitations, especially regarding output reliability, such as the issue of hallucinations and the need for models to express uncertainty.

What is the speaker's perspective on the future of programming and AI interaction?

The speaker believes that we are at an inflection point where the way we interact with digital information is being redefined through AI systems. They envision a future where natural language interfaces become more prevalent, allowing for collaboration with AI models as if they were companions or co-workers.

What was the original intent behind creating Chat GPT, according to the speaker?

Chat GPT was initially created as a means to get feedback from researchers on the dialogue modality of AI, with the goal of using this feedback to improve the alignment, safety, and reliability of more advanced models like GPT-4.

How does the speaker define AGI (Artificial General Intelligence)?

The speaker defines AGI as a computer system capable of performing autonomously the majority of intellectual work, indicating a level of capability that spans a wide range of tasks beyond specific domains.

What are the speaker's thoughts on the future of AI in terms of economics and the workforce?

The speaker anticipates that AI systems will take over more tasks, allowing humans to focus on higher-level work and potentially reducing the amount of time spent working. They also suggest that AI could lead to a more powerful and accessible platform for building applications and products.

What is the speaker's view on the potential future of AI models in terms of their capabilities and roles?

The speaker expects AI models to become incredibly powerful, with the ability to understand and interact with the world across multiple modalities. They also highlight the importance of addressing the challenge of super alignment to ensure these models are safe and aligned with human intentions.

Outlines

00:00

🌟 From Albania to AI: A Journey of Intellectual Curiosity

The speaker reflects on their journey from being born in post-communist Albania to developing a deep interest in math and sciences. This early education in a politically isolated country with a strong emphasis on math and physics led to a career in mechanical engineering and eventually to Tesla, where they contributed to the development of the Model S and X. The speaker's focus on applying knowledge to build practical applications, especially in AI, led them to explore various domains including augmented reality, virtual reality, and eventually to OpenAI, where they became deeply involved in the development of AGI (Artificial General Intelligence).

05:01

🤖 The Convergence of Physics and AI: A New Era of Problem Solving

The speaker discusses the increasing intersection between physics and computer science, noting that many influential figures in AI have a background in physics or math. They highlight the unique problem-solving approach required in theoretical fields, which involves deep contemplation and patience. The conversation delves into whether AI development is becoming more of a systems or engineering problem, with the speaker suggesting that while systems and engineering challenges are significant, there is still much to explore in terms of fundamental advancements in AI.

10:02

🛠️ The Evolution of AI Interfaces: From APIs to Natural Language

This section explores the evolution of how humans interact with AI, moving from traditional APIs to more natural language interfaces. The speaker shares their personal experience using AI models and the feeling of using a 'supercomputer with an abacus', indicating the disparity between the power of AI models and the simplicity of current programming interfaces. They speculate on the future of programming, suggesting that it may become less abstract and more collaborative, with AI models acting as partners in the creative process.

15:06

🚀 The Birth of Chat GPT: A Focus on Safety and Alignment

The speaker recounts the origins of Chat GPT, which was not initially intended as a product but emerged from research into safe AI systems. They discuss the development of reinforcement learning with human feedback to align AI systems with human values and prevent misuse. The speaker also touches on the challenges of hallucinations in AI and the efforts to make models more reliable and safe through fine-tuning and user feedback.

20:06

🔮 Prognosticating AI's Future: Scaling Laws and Economic Implications

In this paragraph, the speaker addresses the question of AI's future, particularly the concept of scaling laws and whether the pace of progress will continue. They express optimism about the potential for AI to become more capable as models scale up, despite the historical pattern of diminishing returns in AI. The speaker also considers the economic impact of AI, drawing parallels with the silicon industry and speculating on whether general AI models will dominate or if there will be a fragmentation of specialized models.

25:09

🌐 The Expanding Role of AI: From Text to Multimodal Understanding

The speaker envisions a future where AI models encompass not just text but also images, video, and other modalities, providing a more comprehensive understanding of the world. They discuss the importance of pre-training models to have a broad sense of the world and the challenges of making AI outputs reliable through reinforcement learning with human feedback. The speaker also hints at the potential for AI to solve issues like hallucinations and the importance of building a platform that facilitates collaboration between humans and AI.

🌌 The Challenge of Super Alignment: Ensuring AI Serves Human Intentions

In the final paragraph, the speaker discusses the challenge of super alignment, ensuring that increasingly powerful AI models are aligned with human intentions. They acknowledge the fear associated with powerful misaligned AI and the importance of addressing this technical challenge. The speaker also reflects on their personal stance, positioning themselves as something other than a doomer or accelerationist, indicating a nuanced view of AI's potential and risks.

Keywords

💡AI

AI, or Artificial Intelligence, refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. In the video, AI is central to the discussion, with a focus on its applications and the challenges of building good products on top of AI models. The speaker's journey from studying mechanical engineering to working with AI at Tesla and OpenAI exemplifies the practical application of AI.

💡Autonomy

Autonomy in the context of AI refers to the ability of systems to operate independently without human intervention. The video discusses the progression towards AGI, or Artificial General Intelligence, which is characterized by an AI system's capacity to perform autonomously across a broad range of intellectual tasks, much like humans.

💡Model

In AI, a 'model' is a system that has been trained on data to make predictions or decisions. The script mentions building models and the difficulty of creating good products on top of these models. It also discusses the evolution from theoretical models to practical applications in fields like aerospace and automotive engineering.

💡OpenAI

OpenAI is a research and deployment organization focused on creating and disseminating safe and beneficial AI technologies. The speaker mentions being drawn to OpenAI because of its mission, indicating the organization's role in advancing AI and its emphasis on safety and alignment with human values.

💡AGI

AGI, or Artificial General Intelligence, is the concept of a machine that possesses the ability to understand, learn, and apply knowledge across a wide range of tasks at a level equal to or beyond that of a human. The video discusses the pursuit of AGI and the challenges in achieving a system that can autonomously perform the majority of intellectual work.

💡Reinforcement Learning

Reinforcement Learning is a type of machine learning where an agent learns to make decisions by taking actions in an environment to maximize some notion of cumulative reward. The video mentions reinforcement learning with human feedback as a method to align AI systems with human values and improve their ability to follow instructions.

💡Hallucinations

In the context of AI, 'hallucinations' refer to the generation of false or misleading information by a model, which is a significant issue for AI reliability. The script discusses the challenge of addressing hallucinations in AI models and the potential solutions, such as incorporating browsing capabilities for models to cite sources and reduce misinformation.

💡API

API stands for Application Programming Interface, which is a set of rules and protocols for building software applications. The video mentions the use of APIs to make AI models accessible, allowing developers to integrate AI capabilities into their applications without needing deep expertise in machine learning.

💡Alignment

Alignment in AI refers to the process of ensuring that an AI system's actions and goals are consistent with human intentions and values. The script discusses the importance of alignment for safety and the development of methods like reinforcement learning with human feedback to achieve it.

💡Economics of AI

The economics of AI pertains to the costs, benefits, and market dynamics associated with the development and deployment of AI technologies. The video script alludes to the potential for AI to reshape the economic landscape, with general AI models possibly consuming specific functionalities and creating new industry structures.

💡Productization

Productization is the process of creating a product out of a service, technology, or concept. The video emphasizes the difficulty of building good products on top of AI models, indicating that while creating AI models is a significant focus, turning these models into successful products that meet market needs is a complex challenge.

Highlights

Building good products on top of AI models is incredibly difficult.

The speaker's background includes a focus on math and sciences, influenced by the educational environment in post-communist Albania.

Interest in applying mathematical knowledge to build things, leading to a career in mechanical engineering and aerospace.

Experience at Tesla contributed to an interest in AI applications, particularly autopilot technology.

The transition from theoretical to practical applications of AI in domains like augmented reality and virtual reality.

The importance of generality in AI and the appeal of working on AGI at Open AI.

The prevalence of physics and math backgrounds among influential figures in the AI space.

The value of a theoretical approach to problem-solving in AI development.

The difference in user engagement between APIs and more accessible AI technologies like Chat GPT.

The potential for natural language interfaces to replace traditional programming in AI interaction.

The concept of AI as a collaborative partner rather than just a tool to program.

The strategy of making AI technology accessible to discover new use cases and applications.

The development of Chat GPT as a means to gather feedback from researchers on alignment and safety.

The challenge of addressing AI hallucinations and the potential of reinforcement learning with human feedback.

The decision-making process at Open AI regarding what to work on, focusing on safety and alignment.

The economic implications of AI, drawing parallels to the evolution of the silicon industry.

The future of AI as potentially capable of performing the majority of intellectual work, defined as AGI.

The current state of AI capabilities, with a focus on reliability and expanding to new tasks.

The importance of emerging capabilities in AI, even if they are currently unreliable.

The potential trajectory of AI systems becoming autonomous and the need for human oversight.

The vision for AI platforms, allowing people to build on top of pre-trained models.

The challenge of building good products on top of AI models, which is more difficult than model creation.

The future of multi-modal AI models, incorporating text, images, and video for a comprehensive understanding of the world.

The focus on reliability in AI outputs and the introduction of methods to reduce hallucinations.

The concept of AI as a collection of collaborative agents, providing a platform for others to build upon.

The long-term challenge of super alignment in AI as models become incredibly powerful.

The speaker's perspective on AI's future, neither a doomer nor an accelerationist, but something else.

Transcripts

there is a lot of Focus right now on

building more models but you know

building good products on top of these

models is incredibly difficult

[Music]

thank you

I would love it if you're comfortable

it's giving kind of the longer form of

your background what brought you to open

AI you know just kind of bring us up to

speed and then we'll go from there yeah

so um

it was born in Albania just after the

fall of Communism very interesting times

in this you know very isolated countries

sort of similar to North Korea today and

I bring that up because it was I think

very Central to sort of my

education and focus in math and sciences

because there was a lot of focus on

maths and physics

in in Coast communist Albania and you

know the humanities like history and

sociology and these type of topics they

were a bit questionable like the source

of information and truthfulness was uh

was hard it was ambiguous so anyway I I

got very interested in math and sciences

and that's what I pursued relentlessly

until you know still still working

fundamentally in mathematics

and over time my interests grew more

from this theoretical space into

actually building things and figuring

out how to

apply that knowledge to build stuff and

I studied mechanical engineering

and went on to work in Aerospace as an

engineer and then joined Tesla shortly

after where I spent a few years

initially I joined to work on Model S

dual motor and then I went on to model X

from the early days of the initial

design and eventually led the whole

program

um to to Launch

and this is when I got very interested

in applications of AI and

specifically with autopilot and so

I started thinking more and more about

different applications of AI okay what

happens when you when you you're using

Ai and computer visually in a different

domain instead of autopilot

and after Tesla I went on to work on

augmented reality and virtual reality

because I just wanted to get experience

with different domains and I thought at

the time that it was the right time to

actually work on special computing

obviously in retrospect too early

back then but anyways I learned a lot

about the limitations of pushing this

technology to the practicality of using

it every day and at this point I started

thinking more about what happens if you

just focus on the generality like forget

the competence in different domains and

just focus on generality and there were

two places at the time there were laser

focused on this issue and open Ai and

deepmind

uh and I was very drawn to open AI

because of its Mission and I felt like

there's not going to be a more important

technology that we all built than than

AGI

um

back then I certainly did not have the

same conviction about it as I do now

but I thought that fundamentally if

you're building intelligence it's such a

it is such a core unit then the

university it affects everything and so

you know what what else is there is

there to do more inspiring than than

Elevate and increase Collective

collective intelligence of humanity

whenever I meet somebody that's a like a

real

um uh influencer and has done major

contributions to the space they almost

invariably have a physics background or

a math background which is actually very

different than it was 15 years ago like

15 years ago I was like you know the

kind of you know it was like engineers

and you know they came from lexical

engineering mechanical engineering but

it does feel like

um you know there's something and I

don't know if it's like some like quirk

in the network or like it's it's more

fundamental like systemic and I mean do

you think that this is kind of the time

for the physicist to step up and kind of

contribute to computer science and

there's something about that or do you

think it's just more of a coincidence

so I think maybe one thing to draw on

from the theoretical space of math

but also the kind of the natural

problems with math is that you know you

kind of need to sit with a problem for a

really long time and you have to think

about it sometimes to sleep and you wake

up and you have a new idea and over the

course of maybe a few days sometimes or

weeks you get to the final solution and

so it's not like a quick reward

and sometimes it's not this iterative

thing and and I think it's almost like a

different way of thinking where you're

building an intuition but also a sort of

discipline to sit with the problem and

have faith that you're going to solve it

and over time you build an intuition on

what problem is the right problem to

actually work on so do you think it's

now more of a systems problem more like

kind of more of an engineering problem

or do you think that we still have a lot

of like kind of pretty real kind of

signs to unlock

um both I think the systems and the

engineering problem is massive

um is as we're deploying these

Technologies out there

um and we're trying to scale them we're

trying to make them more efficient we're

trying to make them easily accessible so

you don't need to have you know to know

the intricacies of ml in order to use

them and actually you can see sort of

the contrast between making these models

available through an API and making the

technology available through child GPT

it's fundamentally the same technology

maybe with with a small difference with

reinforcement learning with human

feedback for chat GPT but it's

fundamentally the same technology and

the reaction and the ability to to grab

people's imagination and to get them to

just use the technology every day is

totally different I also think the API

for example PPT is such an interesting

thing so it's my program against these

models myself for fun right and it

always feels like whenever I'm using one

of these models in a program I'm like

I'm wrapping a super computer with an

abacus it's like the code itself just

seems so kind of flimsy compared to the

model that it's wrapping sometimes I'm

like listen I'm just going to give the

model

like a keyboard and a mouse and like and

let it do the programming and then

actually the API is going to be English

and I'll just tell it what to do and

it'll do all the programming and I'm

just kind of curious as you designed

stuff like chat GPT do you view that

over time the actual interface is going

to be like the like natural languages or

do you think that there's still a big

role for programs the programming is

becoming less abstract where we can

actually talk to computers in high

bandwidth in natural language but

another Vector is one where we're using

the technology and the technology is

helping us understand how to actually

collaborate how to collaborate with it

versus program it and I think there is

definitely the layer of programming

becoming easier more accessible because

you can program things in natural

language but then there is also this

other side which we've seen with child

GPT that you can actually collaborate

with the model as if it was a companion

a partner Co worker you know that's the

interesting thing like it'll be very

interesting to see what happens over

time like you've made decision to have

an API and whatever but like you don't

like have an API to a co-worker right

like you talk to a co-worker so it could

be the case that like over time these

things evolve into like you just speak

natural languages or do you think it

will always be a component of a finite

State machine a traditional computer

that's it yeah I think this is right now

an inflection point where we're sort of

you know redefining how we interact with

with digital information and it's it's

through you know the form of this AI

systems that we collaborate with and uh

maybe we have several of them and maybe

they all have different competences and

maybe we have a general one that kind of

follows us around everywhere knows

everything about uh you know my context

what I've been up to today what my goals

are

um sort of in life at work and kind of

guides me through and coaches me and so

on and you know you can imagine that

being super super powerful so I think it

is we are right now at this inflection

point of redefining what this looks like

um but you know

but there is also

we don't know exactly what the future

looks like and so

we are trying to make these tools

available and the technology available

to a lot of other people so they can

experiment and we can see what happens

it's a strategy that we we've been using

from the beginning and also with child

GPT where you know the week before we

were worried that it wasn't good enough

and we also what happened you know we

put it out there and then people told us

it is good enough to discover new use

cases and you see all this emergent use

cases that I know you've written about

um and that's what happens when you make

this stuff accessible and easy to use

and put it in the hands of everyone so

this leads to my my next question which

is

um so you invent cold fusion and then

part of it you're like okay listen I'll

just give people like electrical outlets

and they'll use the energy but like I

think when it comes to AI people don't

really know how to think about it yet

and so like there has to be some

guidance like you have to make yeah some

choices and so you know you're an

opening you know you have to decide what

to work on next well if you could like

walk through that decision process like

how do you decide like what to work on

or what to focus on or what to release

or how to position it if you consider

how Chad GPT was born it was not born as

a product that we wanted to put out

there in fact the real roots of it go

back to like more than five years ago

when we were thinking about how do you

how do you make this safe AI systems

um and you know you don't necessarily

want humans to actually write the the

goal functions because you don't want to

use proxies for complex call functions

or you don't want to get it wrong it

could be very dangerous this is where

reinforcement learning with human

feedback was was developed where you

know

but what we were trying to really

achieve was to align the the AI system

to human values and get it to receive

human feedback and based on that human

feedback it would be more likely to do

the right thing less likely to do the

thing that you don't want it to do

and uh you know after we developed gpt3

and we put it out there in the API this

was the first time that we actually had

Safety Research become practical into

the real world and this happened through

instruction following models so we use

this method

to basically take

prompts from customers using the API and

then we had contractors

generate feedback for the model to learn

from and we fine-tuned the model

on on this data and build the

instruction following models there were

much more likely to to follow the intent

of the user and to do the thing that you

actually wanted to do and so this was

very powerful because AI safety was not

just this theoretical concept that you

sit around and you talk about but it it

actually became you know was sort of

going into AI Safety Systems now like

how do you integrate this into into the

real world

and obviously with large language models

we see great representation of Concepts

ideas of the real world but on the

output front there are a lot of issues

um and one of the biggest ones is

obviously hallucinations

and so

we we had been studying the issue of

hallucinations truthfulness

um how do you get these models to

express uncertainty the precursor to

child GPT was actually another

project that we called Web GPT

and it used retrieval to be able to get

information and cite sources and so this

project then eventually turned into

child gbt because we thought the

dialogue was really special because it

allows you to sort of you know ask

questions to correct the other person to

express uncertainty there's just

something down the error because you're

interacting exactly there is this

interaction and you can get to a deeper

truth

um and and so anyway we we started going

down this path and at the time we were

doing this with gpt3 and ngpts 3.5

um and and we were very excited about

this from a safety perspective

um but you know one thing that

people forget is that actually at this

time we had already trained gbd4 and so

internally at open AI we were very

excited about gbt4 and sort of put

chagibit in the rear view mirror

and uh then you know we kind of realized

okay we're gonna take six months

to focus on alignment and safety of gbt4

and we started thinking about things

that we could do and uh one of one of

the main things was actually to put

charge ubt in the hands of researchers

out there that could give us feedback

since we had this dialogue modality and

so this was the original intent to

actually

get feedback from researchers and use it

to make contributive form or aligned and

safer and more robust more reliable I

mean just for clarity when you say align

and safety do you actually

do you include in that like correct and

does what it wants or do you mean

safety like actual like protecting from

some sort of harm by alignment I

generally mean that it aligns with the

user's intents so it does exactly the

thing that you wanted to do

but safety includes other things as well

like misuse where the user is you know

intentionally trying to use the model to

generate to create harmful outputs

um

so yeah you can we were trying in this

in this case with charge GPT we were

actually

um trying to make the model more likely

to do the thing that you wanted to do to

make it more lines and uh we also wanted

to figure out the issue of

hallucinations which is obviously an

extremely hard problem but I do think

that with this method of

to reinforcement learning with human

feedback maybe that is all we need if we

push this hard dinner so there's no

grand plan it was literally like what do

we need to do to like get to AGI and

it's just one step after right yes and

it's you know all the little decisions

that you make along the way but

maybe what made it more likely to happen

is the fact that we did make a strategic

decision a couple of years ago to pursue

products yeah and we did this because we

thought it was actually crucial to

figure out how to deploy these models in

the real world and it would not be

possible to just you know sit in a lab

and develop this thing in a vacuum

without feedback from users from The

Real World so there was there was a

hypothesis and and I think that that

helped us along the way make some of

these decisions build the underlying

infrastructure so that we could actually

eventually deploy things like

I would love if you would Riff on

scaling laws I think this is the big

question that everybody has like I mean

like the pace of progress has been

phenomenal and you would love to think

that the the graph always does this but

like the history of AI seems to be that

like you hit diminishing returns at some

point and it's not parametric it kind of

like tapers off and so

from your standpoint was probably like

the most informed vantage point in the

entire industry do you think the scaling

laws are going to hold and we're going

to continue to see advancements or do

you think we're hitting diminishing

returns

so there isn't

any evidence that we will not get much

better much more capable models as we

continue to scale them across the access

of data and compute whether that takes

you all the way to AGI or not that's a

different question there are probably

some other breakthroughs and

advancements needed along the way but I

think there's still a long way to go

in in the scaling laws and to really

gather a lot of benefits from from these

larger models how do you define AGI

um in our chart Opening Our Charter we

we Define it as a computer system

basically that is able to perform

autonomously the majority of

intellectual work okay I am was that I

was at a lunch and Robert nishihara from

any scale was there and um and he asked

what I called it Robert nishihara

question which I thought was actually

very good characterization he said okay

so like you've got to continue in

between like say a computer and Einstein

so you can go from a computer to a cat

you know from a cat to an average human

and you go from an average human to

Einstein then they ask a question of

okay so where are we on the continue

what problem have we solved and the

consensus was we know how to go from a

cat to an average human like we don't

know how to go from like a computer to a

cat because like that's you know that's

the general perception problem or very

close but we're not quite there yet and

then we don't really know how to do the

Einstein which is kind of set to set

reasoning with fine tuning you can get a

lot obviously

um but in general I think we're sort of

the most tasks kind of like in turn

level I would say that's what I I

generally say the issue is reliability

right of course you know you can't fully

rely on the system to do the thing that

you wanted to do all the time and you

know how do you increase that

reliability over time and then how do

you obviously expand to the the

capabilities

um the new the emergent capabilities the

new things that these models can do I

think though that it's important to pay

attention to these emerging capabilities

even if if they're highly unreliable and

especially for people that you know are

building companies today you really want

to think about okay what what's somewhat

possible today

um what do you see glimpses of today

because you know very quickly this could

actually become

these models could become reliable so

I'd love I've been asking just a second

to prognosticate on what that looks like

but before very selfishly I've got uh

I've got a question

on on how you think the economics of

this are going to pencil out which is

I'll tell you what it reminds me of it

reminds me very much of the Silicon

industry

so I remember in the 90s when you buy a

computer there are all these weird

co-processors there's like here's like

string matching here's a floating Point

here's crypto and like all of them got

consumed into basically the the CPU it

just turns out generality was very

powerful and that created a certain type

of economy one where like you had you

know Intel and AMD and like you know it

all went in there

and of course because a lot of money to

build these chips and so like you can

imagine two Futures there's one future

where like you know generality is so

powerful that over time the large models

basically consume all functionality

and then there's another future where

there's going to be a whole bunch of

models and like things fragment and you

know different points of the design

space do you have a sense of like

is it open Ai and nobody or is it

everybody it kind of depends what you're

trying to do so obviously the trajectory

is one where these AI systems will be

doing will be doing more and more of the

work that we're doing and they'll be

able to operate autonomously but we will

need to provide Direction and guidance

and oversee but I don't want to do a lot

of the repetitive work that I have to do

every day I want to focus on other

things and maybe we don't have to work

you know

10 12 hours a day and maybe we can work

less and Achieve even higher

outputs and so that's sort of what I'm

hoping for but in terms of like how this

how this works out with with the

platform

um

you can see even today you know we make

a lot of models available through our

API and from the various from the very

small models to to our Frontier models

and people don't always need to use the

most powerful the most capable model

sometimes they just made the model that

actually fits for their specific use

case and it's far more economical so I

think there's going to be a range

um but yeah in terms of how we're

Imagining the platform play

um we definitely want people to build on

top of our models and we want to give

them tools so to make that easy and give

them more and more access and control so

you know you can bring your data you can

customize these models and you can

really focus on the layer beyond the

model and defining the products which is

actually really really hard there is a

lot of Focus right now on building more

models but you know building good

products on top of these models is

incredibly difficult okay we only have a

couple more minutes sadly I would love

for you to prognosticate a little bit

unlike where you think this is all going

like yeah like three years or five years

or ten years

I think that um the the foundation

models today obviously have this great

representation of the world in text and

we're adding other modalities like

images and video and various other

things so these models can get a more

comprehensive sense of the world around

us similar to how we understand and

observe the world the world is not just

in text it's also in images so I think

that will certainly expand in in that

direction and we'll have these bigger

models that will have all these

modalities

um and that's kind of the pre-training

part of the work where we really want to

get

these pre-trained models that understand

the world like we do and then there is

the output part of the model where we

introduced reinforcement learning with

human feedback and we want the model to

do the actually the thing that we ask it

to do and we want that to be reliable

and there is a ton of work that needs to

happen here and maybe introducing

browsing so you can get fresh

information and you can cite information

and solve hallucinations I don't think

that's impossible I think that's

achievable on the product side I think

we want to put this all together in this

collection of agents that people

collaborate with and you know really

provide a platform where people can

build on top of and you know if you

extrapolate really far out these models

are going to be incredibly incredibly

powerful and with that obviously comes

fear of them being misaligned having

this very powerful model that are

misaligned with our intentions so then a

huge challenge becomes the the challenge

of of super alignment which is a

difficult technical Challenge and we've

we've assembled an entire team at open

AI um

to just focus on on this problem so very

very very last question are you a Doomer

an accelerationist or something else

let me say something else all right

perfect thank you so much fantastic

[Applause]

[Music]