BREAKING: OpenAI's Going Closed (yes really)

Theo - t3․gg
26 Sept 202419:34

Summary

TLDRThe video discusses major changes at OpenAI, focusing on its shift from a nonprofit to a for-profit structure, which could have significant implications for AI safety and accessibility. The script covers OpenAI’s original mission to ensure AI benefits all, the introduction of profit-driven goals, and leadership departures. It highlights tensions between nonprofit control and investor-driven interests, raising concerns that OpenAI may prioritize profits over safety. The restructuring includes Sam Altman now having equity, sparking concerns from the AI safety community about the company's future direction.

Takeaways

  • 🤔 OpenAI has historically not open-sourced its models, despite providing significant research documentation.
  • 😲 OpenAI's recent shift from nonprofit to a for-profit benefit corporation marks a dramatic change in its structure.
  • 💼 Investors will now have fewer restrictions on profits, making OpenAI a more traditional investment opportunity.
  • 🔄 The nonprofit entity will remain, but it will only hold a minority stake in the new for-profit structure.
  • 🚨 The change raises concerns about how OpenAI will manage AI risks in a more profit-driven environment.
  • 👨‍💼 Sam Altman now has equity in OpenAI, a significant shift since he previously didn't own any part of the company.
  • 📉 The disbanding of the Super Alignment team, which focused on AI safety, has raised fears about OpenAI's commitment to its original safety goals.
  • 🤑 OpenAI's for-profit shift could lead to increased focus on profits over safety, sparking concerns about the future of AI governance.
  • 👋 Several leadership changes have occurred, including the departure of CTO Mira Murati and other key leaders.
  • 📊 OpenAI's valuation has skyrocketed, with investors like Microsoft and others keenly watching this transition into a more traditional corporate structure.

Q & A

  • What changes has OpenAI made to its business structure recently?

    -OpenAI is restructuring from a nonprofit to a for-profit benefit corporation, which will no longer be controlled by its nonprofit board. This change is aimed at attracting more traditional investors and enabling higher returns.

  • Why was OpenAI originally established as a nonprofit organization?

    -OpenAI was originally founded to ensure AI development was safe and beneficial to all of humanity, rather than being controlled by a few large corporations with potentially unsafe practices.

  • What is the significance of Sam Altman now owning equity in OpenAI?

    -Previously, Sam Altman did not own any equity in OpenAI to maintain impartiality. With the restructuring, he now owns a significant stake, which could be worth up to $150 billion, aligning his financial incentives more closely with the company's profitability.

  • How did OpenAI’s relationship with investors change due to the restructuring?

    -Investors can now invest in OpenAI in a more traditional way without caps on returns. Previously, OpenAI had a hard cap on investment returns, limiting them to a maximum of 10x the initial investment.

  • What concerns have been raised about OpenAI's shift to a for-profit model?

    -The shift to a for-profit model could lead to reduced emphasis on AI safety and accessibility, as profit incentives might overshadow the original mission of developing safe and beneficial AI for everyone.

  • What impact did the restructuring have on OpenAI's leadership team?

    -Several key figures, including CTO Mira Murati, announced their departure. This suggests significant internal changes and a shift in leadership as OpenAI transitions into a new business structure.

  • How did the boardroom drama in 2022 affect OpenAI's structure?

    -In 2022, the nonprofit board ousted Sam Altman due to a loss of trust and communication issues, but he was reinstated within five days due to overwhelming support from employees and investors. This incident highlighted the tension between the nonprofit board's goals and the company's direction.

  • What role did Microsoft play in OpenAI's recent changes?

    -Microsoft, a significant investor in OpenAI, now has a seat on the board. Their involvement has raised concerns about the influence of large corporations on the direction of AI development at OpenAI.

  • What happened to OpenAI’s super alignment team?

    -The super alignment team, which focused on keeping AI development aligned with human values and safety, was disbanded, raising concerns about OpenAI's commitment to long-term AI safety.

  • How does the new for-profit structure of OpenAI compare to other AI companies?

    -The new structure resembles that of its rivals, such as Anthropic and Elon Musk's X.AI, which are also for-profit benefit corporations that aim to balance profitability with social responsibility and sustainability.

Outlines

00:00

🔍 OpenAI Shifts Away from Nonprofit Roots

The video opens by addressing a long-standing critique: despite its name, OpenAI is not truly open in terms of its models being closed-source. The company has maintained a nonprofit status, focused on benefiting humanity, but that structure has recently changed. This shift marks the end of an era for OpenAI, as it transitions towards a for-profit benefit corporation. The narrator expresses shock at the scale of the changes, especially given the past controversies around OpenAI. The transition is surprising, with the CTO leaving and Sam Altman set to earn more, reflecting a significant shift in OpenAI's core structure and goals.

05:03

📈 OpenAI's New Corporate Structure

OpenAI is restructuring into a for-profit benefit corporation to attract more traditional investors. Previously, OpenAI capped returns for investors, limiting potential profits. Now, with this cap removed, OpenAI becomes more appealing to investors, allowing greater financial gains. Despite this shift, the original nonprofit will remain, but with a minority stake in the new structure. Concerns arise about how this change might affect AI safety governance, one of OpenAI's founding principles. The restructuring may allow investors to focus more on profit than safety, raising questions about the company’s future direction.

10:05

⚖️ Tensions Between Nonprofit and For-Profit Structures

The video describes the tension between OpenAI's nonprofit board and its for-profit division. Initially, the nonprofit controlled decisions, but the for-profit section has grown more dominant, creating friction. This shift flips the original governance structure, with the corporation now leading while the board takes a backseat. The restructuring has sparked leadership changes, including the CTO's departure, further signaling the company's transformation into a profit-driven entity. This change mirrors moves by other AI companies, like Anthropic, as OpenAI seeks to maintain its competitive edge in the booming AI market.

15:05

🏢 Leadership Changes Amid OpenAI's Growth

OpenAI's leadership is undergoing major changes as it restructures. The company's longtime CTO, Mira Murati, and other key executives have left, raising questions about OpenAI’s future direction. These departures coincide with OpenAI's transition from a nonprofit to a for-profit model. The changes follow the company’s explosive growth, driven by ChatGPT's success, which attracted significant investor interest. OpenAI is shifting away from its original mission of making AI safe and accessible to all, prioritizing a more traditional business model that appeals to investors and aligns with its rapid expansion.

🚨 AI Safety and the Dismantling of Core Goals

The narrator highlights that OpenAI was founded to prevent AI from being monopolized by a few companies or developed unsafely. However, the company’s new for-profit direction risks undermining these original goals. OpenAI no longer offers open models, forcing users to rely on its closed infrastructure. Additionally, safety concerns are mounting as OpenAI disbands its super alignment team, which was dedicated to ensuring AI safety. With the nonprofit’s control reduced, OpenAI now resembles a typical startup focused on maximizing profits, sparking concerns among AI safety advocates.

⚠️ Safety Concerns as OpenAI's Mission Evolves

In the final part of the script, the narrator addresses concerns over OpenAI's mission. The company, which once aimed to make AI both safe and accessible, is now largely profit-driven. The nonprofit board, once a safeguard, has been sidelined, and Microsoft, one of the very companies OpenAI was wary of, now holds significant influence. The disbanding of the super alignment team raises additional fears about AI safety. The narrator acknowledges these concerns but remains cautiously optimistic, awaiting further developments as OpenAI continues to evolve in its new for-profit structure.

Mindmap

Keywords

💡OpenAI

OpenAI is an AI research organization that has been at the forefront of artificial intelligence innovation. Originally founded as a nonprofit aimed at advancing safe and widely accessible AI, the video explains how OpenAI is shifting its structure from nonprofit to for-profit, which may have significant implications for its mission and control.

💡Nonprofit to For-profit Transition

This refers to OpenAI's shift from being a nonprofit organization to becoming a for-profit entity. The video highlights how this change, which includes removing the nonprofit board’s control, may impact the company's priorities and safety considerations, leading to concerns over whether the original mission of OpenAI will be preserved.

💡Sam Altman

Sam Altman is the CEO of OpenAI, and central to the video’s discussion about leadership and the company's recent structural changes. Initially, Altman had no equity in OpenAI, but the video describes how the recent transition gives him a potentially massive stake in the company, which could be worth billions after restructuring.

💡AGI (Artificial General Intelligence)

Artificial General Intelligence (AGI) refers to a type of AI that can understand, learn, and apply intelligence across a wide range of tasks at a level equal to or exceeding human intelligence. OpenAI’s mission was to develop safe AGI that would be beneficial for all, but the video raises concerns about whether profit-driven motives might undermine safety and accessibility.

💡Investor Caps

Investor caps were a unique feature of OpenAI's original for-profit subsidiary, where profits for investors were limited to a 10x return. The video describes how OpenAI is now removing these caps, making it more attractive to traditional investors but also signaling a move toward a more typical for-profit structure, which may prioritize profits over mission.

💡AI Safety

AI safety refers to the measures taken to ensure that AI technology is developed responsibly, avoiding potential harm to society. OpenAI's original mission was deeply rooted in ensuring AI safety. The video expresses concerns that with the disbanding of OpenAI's super-alignment team and leadership changes, safety might take a backseat to profit motives.

💡Super-alignment Team

The Super-alignment Team was a group within OpenAI dedicated to aligning AI development with human values and ensuring that AI remains safe. The video mentions that this team has been disbanded, sparking concerns about whether OpenAI still prioritizes long-term safety in its development of AI technology.

💡Microsoft

Microsoft is one of the major investors in OpenAI and holds significant influence over the company. The video points out that Microsoft now has a seat on OpenAI's board, which raises concerns about how much control large corporations will have over the development and use of OpenAI's technology, potentially going against the original mission.

💡Nonprofit Board

The nonprofit board initially controlled OpenAI to ensure the organization adhered to its mission of safe and equitable AI development. The video discusses how the board’s influence has diminished with the restructuring, leading to worries that the for-profit company's goals may now outweigh the original nonprofit's mission.

💡AI Accessibility

AI accessibility refers to making AI tools and models available to the broader public, not just powerful companies. OpenAI was founded on the belief that AI should be accessible to everyone, but the video highlights how the company's current trajectory—such as not open-sourcing models—may restrict AI use to a small group of large corporations, counter to its founding vision.

Highlights

OpenAI is transitioning from a nonprofit organization to a for-profit benefit corporation.

Historically, OpenAI was a nonprofit with a mission to ensure AI development benefitted humanity.

OpenAI's new structure is intended to attract more investors by removing the profit cap that previously limited returns.

The nonprofit branch will still exist, but will own a minority stake in the new for-profit corporation.

The restructuring has implications for AI safety governance, a key concern that led to the formation of OpenAI.

Sam Altman, previously without equity in OpenAI, will now own a significant stake potentially worth up to $150 billion.

This shift in OpenAI's structure echoes moves by its competitors, like Anthropic and Elon Musk's xAI, which are also for-profit entities.

Mira Murati, OpenAI's CTO, has announced her departure amid the restructuring.

OpenAI's superalignment team, focused on ensuring AI safety, has been disbanded, raising concerns about future governance of AI risks.

Microsoft, a significant investor, now holds a seat on the OpenAI board, increasing concerns about corporate influence over AI development.

OpenAI's nonprofit board originally controlled decisions to ensure AI safety, but this is no longer the case after the restructuring.

The board's decision to fire Sam Altman in late 2023 highlighted tensions between profit motives and AI safety concerns.

Despite concerns about safety and accessibility, OpenAI has continued to make advances in AI development and open-source research.

Sam Altman tweeted his gratitude towards Mira and other departing leaders, noting leadership changes as natural in fast-growing companies.

OpenAI is experiencing significant internal leadership shifts, with new leaders like Mark, Josh, and Matt Knight taking on key roles in the restructured company.

Transcripts

play00:00

we've been making memes about how open

play00:01

AI is not really open for a long time

play00:03

now and it's true open AI is one of the

play00:06

major providers of models that doesn't

play00:08

open source their models so you can't

play00:10

run them yourself they've open sourced a

play00:12

lot of info about how the models work

play00:14

they've been great about documenting

play00:15

their research but open AI itself has

play00:18

been providing closed Source software

play00:20

through their services for quite a while

play00:22

that all said they still were a

play00:23

nonprofit the actual company structure

play00:26

was that with a board that was

play00:28

specifically dedicated to bettering

play00:30

Humanity as a nonprofit charitable org

play00:34

that just changed and I really didn't

play00:37

think it would happen it's an

play00:38

unbelievable shift especially after all

play00:41

the drama last year that Prime and I

play00:42

covered with like 20 videos it's come to

play00:45

an end and here's the results I didn't

play00:48

think this was actually going to happen

play00:50

holy I saw earlier today that the

play00:52

CTO Mira was leaving but I didn't

play00:55

realize it was this big oh God I think

play00:58

it's only fair if Sam Alman is going to

play01:00

make more money that I am too so let's

play01:01

hear from our sponsor real

play01:06

quick what's all that on your

play01:09

desk remember how I mentioned we were

play01:11

working on that new

play01:13

app oh yeah I remember the iPhone app

play01:16

right yeah it was supposed to be but our

play01:18

competition just shipped the Android app

play01:20

and now we're in catchup mode and uh so

play01:23

sorry I'm getting a

play01:24

call

play01:26

yeah you you want the app on Windows as

play01:31

well okay man that sucks doesn't your

play01:35

app have to support a bunch of platforms

play01:37

too how are you so calm oh we just

play01:39

worked with infinite red they built the

play01:41

whole app for us what yeah they're

play01:43

industry leaders in react native they

play01:44

made it so anyone at our company can

play01:46

contribute to the code base like whole

play01:47

of the react web devs they're shipping

play01:49

features on iOS Android and Windows with

play01:52

just one code

play01:53

base H are you not surprised they work

play01:56

with Microsoft all the time they know

play01:57

what they're doing what about Mac Mac

play02:01

that was fast said you can have it by

play02:03

the end of the day that's pretty cool

play02:06

what thank you infinite red for

play02:08

sponsoring today's video check them out

play02:10

at infinite. R make sure you tell them

play02:11

Theo sent you chat gbt maker open AI is

play02:14

working on a plan to restructure its

play02:16

Core Business into a for-profit benefit

play02:18

corporation that will no longer be

play02:20

controlled by its nonprofit board people

play02:22

familiar with the matter told Reuters in

play02:24

a move that'll make the company more

play02:25

attractive to investors the way that

play02:27

open AI handles investors right now is

play02:30

interesting most companies investors buy

play02:33

a percentage with the hopes of being

play02:34

able to sell that percentage at some

play02:36

point in the future open a had an

play02:38

interesting rule where profits on

play02:41

investment were capped they had a hard

play02:42

cap I think it was 10x so once if you

play02:44

invested a million dollars and you could

play02:46

sell it for 10 million you couldn't make

play02:48

more than 10 million on that sale so

play02:50

this is a huge change because now

play02:52

investors can invest in a much more

play02:53

traditional way without these weird caps

play02:55

and whatnot on the Investments huge

play02:58

shift very strange and unex expected the

play03:00

open AI nonprofit will continue to exist

play03:02

and own a minority stake in the

play03:04

for-profit company the move could also

play03:06

have implications for how the company

play03:07

manages AI risks in a new governance

play03:09

structure that's the main concern a big

play03:11

part of why open AI was formed in the

play03:13

first place was to make sure that AI

play03:15

development and advancement wasn't a

play03:17

thing that like two big companies made a

play03:19

bunch of money off of and didn't care to

play03:21

bother with the safety side it was a

play03:23

combination of AI should benefit all and

play03:26

the best AI should be safe and those two

play03:29

things were huge core parts of why open

play03:31

AI was formed in the first place and

play03:33

then they got so far ahead of everyone

play03:35

else so early that it was hard for them

play03:37

to not make money and now they're in the

play03:39

weird position of having a CEO that

play03:41

knows how to make money at businesses in

play03:44

the original nonprofit board not liking

play03:47

the direction of the company all the

play03:48

drama that happened and now this it's

play03:50

weird for a nonprofit to become the

play03:52

industry lead in something and it makes

play03:55

sense they're no longer a nonprofit but

play03:57

God this is a weird scenario also have

play03:59

not is that Sam Altman for the first

play04:01

time now has equity in open AI crazy

play04:04

enough Sam Altman didn't own any part of

play04:07

open AI before so if they did really

play04:08

well or poorly he might make a bit more

play04:10

money but that was about it open aai was

play04:13

not a company he owned stock in versus

play04:15

like me I own a huge portion majority

play04:18

portion of the things that we built with

play04:20

ping Labs so upload thing marker thing

play04:22

pick thing all those things as part of

play04:24

ping I own over half of he owned zero

play04:28

now he owns quite a bit more in fact the

play04:30

amount he owns could be worth up to $150

play04:32

billion after the restructuring as it's

play04:34

also trying to remove the cap on returns

play04:36

for investors which I mentioned before

play04:38

there's that weird cap where you can

play04:39

only make a certain multiplier on top we

play04:41

remain focused on building AI that

play04:43

benefits everyone we're working with our

play04:45

board to ensure that we're best

play04:46

positioned to succeed in our mission the

play04:48

nonprofit is core to our mission and it

play04:49

will continue to exist because the

play04:51

structure right now before these changes

play04:53

go in place for those who haven't kept

play04:54

up with all of this is the open aai

play04:56

board has stayed a fully nonprofit

play05:00

entity and the open AI company where all

play05:02

the engineers work and all the building

play05:04

happens that is a company that reports

play05:06

to the board underneath it so the board

play05:09

is the ones who say what goes they're

play05:11

the actual owners and decision makers

play05:13

they have all the the voting shares of

play05:16

the openai corporation but the open aai

play05:18

corporation that can make money that

play05:20

does all the work that does all these

play05:21

things operates separately and that

play05:24

separation was starting to cause tension

play05:25

between the groups and now they are

play05:28

separating entirely it actually seems

play05:29

like it flipped where the corporation's

play05:30

on top now and the board reports to them

play05:33

as like a voter but they don't make the

play05:35

decisions anymore the details of the

play05:37

proposed corporate structure for

play05:38

reported by Reuters highlight

play05:39

significant governance changes happening

play05:41

behind the scenes at one of the most

play05:43

important AI companies they even call

play05:45

out the Mira leaving thing that I

play05:46

mentioned before the plan is still being

play05:48

hashed out with lawyers and shareholders

play05:49

and the timeline for completing the

play05:51

restructure remains uncertain the

play05:53

restructuring also comes amid a series

play05:55

of leadership changes at the startup

play05:56

open ai's longtime Chief technology

play05:58

officer Mira mad abruptly announced her

play06:01

departure from the company on Wednesday

play06:02

Greg Brockman open ai's president has

play06:04

also been on leave founded in 2015 as a

play06:06

nonprofit AI research organization

play06:08

openai added the for-profit openai lp

play06:11

entity in 2019 as a subsidiary of its

play06:14

nonprofit securing capital for Microsoft

play06:16

to fund its research as I mentioned the

play06:18

openai profitable part was a subsidiary

play06:21

of the nonprofit the company captured

play06:23

Global attention with the launch of chat

play06:24

gbt in late 2022 which was a generative

play06:26

AI app that spit out human-like

play06:27

responses to text queries which became

play06:29

one of the fastest growing applications

play06:30

in history with over 200 million weekly

play06:32

active users igniting Global race to

play06:34

invest in AI along with chat gbt success

play06:37

open ai's valuation is skyrocketed from

play06:39

14 billion in 2021 to 150 billion in the

play06:42

new convertible debt round under

play06:43

discussion which is attracting investors

play06:45

such as Thrive capital and apple the

play06:47

company's unusual structure which gives

play06:49

full control of the for-profit

play06:51

subsidiary to the open AI nonprofit was

play06:53

originally set to ensure the mission of

play06:55

creating safe AGI that is broadly

play06:56

beneficial referring to artificial

play06:58

general intelligence that is at or

play07:00

exceeding human intelligence as I

play07:02

mentioned before the goal of open AI was

play07:03

to create safe artificial general

play07:06

intelligence that was available for

play07:08

anyone to use so it wasn't just

play07:09

Microsoft Apple and Facebook that had it

play07:11

and no one else could the structure came

play07:13

into focus last November during one of

play07:14

the biggest boardroom dramas in Silicon

play07:16

Valley members of the nonprofit board

play07:18

ousted Altman over a breakdown in

play07:19

communication and a loss of trust he was

play07:21

reinstated after 5 days with

play07:22

overwhelming support from employees and

play07:24

investors what a crazy 5 days that was I

play07:26

had a friend visiting and I had to spend

play07:28

the entire time in my my room filming

play07:30

and editing videos about the open AI

play07:32

drama I really hope that's not about to

play07:33

start again don't make me do this again

play07:35

guys please

play07:37

please since then open ai's board has

play07:39

been refreshed with more Tech Executives

play07:41

chaired by Brett Taylor former

play07:42

Salesforce co-ceo who now runs his own

play07:45

AI startup any corporate changes need

play07:47

approval from its nine-person nonprofit

play07:49

board the removal of nonprofit control

play07:51

could make open AI operate more like a

play07:52

typical startup a move generally

play07:54

welcomed by its investors who have

play07:55

poured billions into the company however

play07:57

it could also raise concerns from the AI

play07:59

safety Community about whether the lab

play08:00

still has enough governance to hold

play08:02

itself accountable in its pursuit of AGI

play08:05

as it has dissolved the super alignment

play08:06

team that focused on the long-term risks

play08:08

of AI earlier this year I will say for

play08:10

the conspiracy theorists and the people

play08:12

who are really terrified of AI and AI

play08:14

safety stuff you have quite the bullet

play08:16

list to be scared about right now like

play08:18

let's just go through first we have open

play08:21

AI formed to make safe AGI for all the

play08:26

reason it was formed is because there

play08:28

was concern from these really smart

play08:30

scientist folks including people like

play08:32

Alman Elon Musk and many more that doing

play08:35

AI safely was not a guarantee and that

play08:39

lots of people developing I stuff could

play08:40

do it in a way that would destroy

play08:41

humanity and even if they didn't but it

play08:44

was only made by a couple companies then

play08:47

it wouldn't benefit all of humanity just

play08:49

the bottom lines of those few companies

play08:51

because of those concerns they formed

play08:53

open AI to fight those concerns the

play08:56

concerns being two parts the two major

play08:58

concerns we one AI made for profit might

play09:03

not be safe enough and the safety could

play09:06

potentially destroy Humanity if people

play09:08

are just trying to make money they might

play09:10

leave safety aside and then get

play09:12

destroyed point two is that powerful AI

play09:14

should not be restricted to a small

play09:15

number of Rich companies because if the

play09:17

only people in the world that have ai

play09:19

are like two to three big businesses

play09:21

that's a massive unfair Advantage for

play09:23

them that could destroy the economy

play09:25

imagine if only three companies could

play09:26

make websites terrifying so the two

play09:29

things that open AI was formed to

play09:30

prevent were unsafe Ai and a small

play09:34

number of companies owning the best AI

play09:36

these are the two reasons openai was

play09:38

formed so let's go down the rabbit hole

play09:40

here openai makes the LP for profit

play09:46

section to make development faster for

play09:50

profit part takes investments from the

play09:54

very companies they were scared of and

play09:58

here's where I think the conspiracy

play10:00

theorists are going to start feeling

play10:02

things because the original goal of

play10:04

forming open AI to prevent these things

play10:06

from happening started to feel like it

play10:08

was being lost with the change of the

play10:11

for-profit section but that was okay

play10:14

because they still reported to the

play10:15

nonprofit so anybody saying like oh open

play10:18

AI is now going to be the problem they

play10:20

were built to prevent well the open AI

play10:22

nonprofit board is still the one in

play10:24

control so is it that big a deal yeah it

play10:27

is but it wasn't that bad at the time

play10:29

time but then that for-profit section

play10:31

started to take huge investments from

play10:33

companies like Microsoft wait what about

play10:36

that second part is Microsoft going to

play10:38

get the thing that we didn't want them

play10:40

to get well once we started releasing

play10:42

the models open AI starts releasing

play10:46

models with no way to run them yourself

play10:51

and that was a big deal because as we've

play10:53

MD many a time open AI has closed models

play10:57

you cannot use their models the way you

play10:59

want want to you have to use them via an

play11:00

API on their infrastructure compared to

play11:03

a lot of open models like llama from

play11:05

meta or all the crazy things mistol is

play11:07

doing there's a lot of open model

play11:09

options that you have access to if you

play11:11

want to run a model on your own Hardware

play11:13

maybe embedded on a device stuff like

play11:14

that open AI nope they do not allow that

play11:18

which means the other goal of powerful

play11:20

AI should not be restricted to a small

play11:21

number of Rich companies we went from

play11:23

believing that to not very quick here

play11:27

almost alarmingly so on the other hand

play11:29

to lightly defend them they did open

play11:32

source all of their research they

play11:34

published a lot of papers and a lot of

play11:35

theories in the concept of an llm which

play11:37

is a huge part of how these other open

play11:39

source models could be created they

play11:41

followed in the footsteps of the work

play11:42

open AI did make public so they did

play11:45

Advanced the ecosystem and they allowed

play11:46

for other models to exist but their own

play11:49

model their own best and of the best is

play11:52

not something others have access to

play11:53

which hurts 0 2 a lot but at least the

play11:56

nonprofit still runs things right at

play11:58

least they'll keep it safe

play12:00

right well nonprofit Board gets

play12:04

concerned about safety and incentives

play12:07

Sam doesn't respond so board fires Sam

play12:13

because at this point in time the board

play12:14

was still in control and the board was

play12:17

still people who were aligned with these

play12:19

two goals and this is the point of the

play12:20

board this is the board doing its job if

play12:22

this is where it ended there' be no

play12:24

conspiracy cuz they did their job they

play12:26

kept AI at open AI safe and accessible

play12:28

to all and instead of letting the

play12:30

incentives of profit drive them in a way

play12:32

that makes these two things problems but

play12:35

then outrage brings Sam back OG board is

play12:42

largely fired so the board didn't get to

play12:45

do its job the board didn't get to have

play12:49

its ultimate control where they decided

play12:51

hey if the things are going in a

play12:52

direction that we feel as unsafe or

play12:53

potentially problematic we're going to

play12:55

step in they can't and a large part of

play12:57

why they were forced out as the board

play13:00

was the investors at places like

play13:02

Microsoft putting their weight behind it

play13:04

so the whole point of keeping Rich

play13:07

companies and money out of the best of

play13:09

the best of AI has now flipped and that

play13:12

same money they were concerned about

play13:14

could be argued to be how the open AI

play13:16

board collapsed which is insane but at

play13:19

the very least they promised they would

play13:22

continue to maintain the existing goals

play13:24

but they promised the goals would not

play13:27

change so again the goals are safety and

play13:30

everyone has access we haven't touched

play13:32

too much on the safety but now we have

play13:33

to because the super alignment team gets

play13:36

disbanded this was the team at open AI

play13:39

That's whole thing was just keeping the

play13:43

models in the AI in alignment with what

play13:46

humans needed so their job was just

play13:49

everything they had to do to keep AI

play13:51

safe and from going haywire and killing

play13:53

all of humans and that team has been

play13:54

disbanded Ilia quits as well Ilia was

play13:57

the Russian lunatic Super Smart Genius

play14:01

who invented a lot of the techniques

play14:02

that became modern Ai and it's also

play14:05

really into the security and safety side

play14:08

of AI he was heavily involved with the

play14:10

super alignment team I think he actually

play14:11

ran it and he was also originally on the

play14:13

board before he got removed because he

play14:15

was one of the ones pushing to get Sam

play14:17

removed so now it looks like that safety

play14:18

goal is starting to crumble too but to

play14:21

go back to the goal two The Profit goal

play14:23

now the super alignment team is

play14:24

disbanded oh also I should mention here

play14:27

that uh Microsoft now has a board seat

play14:31

Ilia quits super elment team gets

play14:33

disbanded and now much later open

play14:36

AI is moving the nonprofit to a minority

play14:42

shareholder

play14:43

position so through this absurd process

play14:48

we went from open AI being formed to

play14:50

maintain these two goals to open AI

play14:52

being traditional for-profit business

play14:55

that has a lot of financial incentives

play14:57

to up these two goals

play14:59

so despite the initial structure of the

play15:02

fully open company with the nonprofit

play15:05

board so we've managed to get position

play15:06

everybody was concerned about where the

play15:08

company with the best AI is driven by

play15:10

profit and might not have the right

play15:12

incentives around safety and is making

play15:14

things that are restricted to just the

play15:16

best companies in the world so I think

play15:18

it's fair to say the original two goals

play15:19

open AI was founded on might not be as

play15:22

deep in the blood of open AI anymore

play15:24

especially goal two here but we need to

play15:27

be keeping our eyes out for goal one I'm

play15:29

not a conspiracy theorist I haven't seen

play15:31

anything convincing enough to have me

play15:32

scared about the future of AI tools

play15:35

trying to kill humans but if you're one

play15:36

of the people that is an extremist on

play15:38

that here you go you you have your you

play15:41

have your order of events to feed your

play15:43

conspiracy brain and I fully understand

play15:45

and now that they have a much stronger

play15:47

for-profit drive safety is one of those

play15:49

things that gets left behind when you're

play15:50

driving profit pretty often so uh this

play15:53

might be bad back to the article it's

play15:55

unclear how much Equity Altman will

play15:57

receive Alman already a billionaire from

play15:59

his multiple startup Investments has

play16:00

previously stated that he chose not to

play16:02

take an equity stake in the company

play16:03

because the board needed a majority of

play16:05

disinterested directors with no stake in

play16:06

the company he's also said he has enough

play16:08

money and is doing it because he loves

play16:10

the work the new structure of open AI

play16:11

would resemble that of its major rival

play16:13

in thropic as well as elon's XII which

play16:15

are registered as benefit corporations a

play16:17

form of for-profit that aims to promote

play16:19

social responsibility and sustainability

play16:21

in addition to making

play16:22

profits we'll see how that goes and with

play16:25

everything going as fast as it does Sam

play16:29

has just tweeted once again literally

play16:32

minutes ago we need to cover this oh boy

play16:35

this seems to be a comment more on mea

play16:37

leaving but this will certainly give us

play16:40

more context on everything going on

play16:42

Meera has been instrumental to open ai's

play16:44

progress and growth for the last 6.5

play16:46

years she has been a hugely significant

play16:48

factor in our development from an

play16:49

unknown research lab to an important

play16:51

company when mea informed me this

play16:53

morning that she was leaving I was

play16:55

saddened but of course I support her

play16:57

decision for the past years she has been

play16:59

building out a strong bench of leaders

play17:01

that will continue our progress I also

play17:03

want to share that Bob and Barrett have

play17:04

decided to depart open AI mea Bob and

play17:06

Barret made these decisions

play17:07

independently of each other and amicably

play17:09

but the timing of Mirror's decision was

play17:11

such that it made sense to now do all of

play17:13

this at once so that we can work

play17:14

together for a smooth hand over to the

play17:16

next generation of leadership I am

play17:18

extremely grateful to all of them for

play17:20

their contributions being a leader at

play17:21

open AI is all consuming on one hand

play17:24

it's a privilege to build AGI and be the

play17:26

fastest growing company that gets to put

play17:27

our Advanced research in the hands of

play17:29

hundreds of millions of people on the

play17:31

other hand it's Relentless to lead a

play17:32

team through it they've gone above and

play17:34

beyond the Call of Duty for the company

play17:36

Mark's going to be our new SVP of

play17:37

research and will now lead the research

play17:39

org in partnership with Jacob as Chief

play17:41

scientist this has been our long-term

play17:43

succession plan for Bob someday although

play17:45

it's happening sooner than we thought I

play17:46

couldn't be more excited that Mark is

play17:48

stepping into the role Mark obviously

play17:49

has deep technical expertise but he's

play17:51

also learned how to be a leader in

play17:52

manager in a very impressive way over

play17:54

the past few years Josh is going to take

play17:56

on a new role as head of mission

play17:58

alignment working across the company to

play18:00

ensure we get all pieces in culture

play18:02

right to being a place to succeed at the

play18:04

mission Kevin and three will continue to

play18:06

lead the applied team Matt Knight will

play18:08

be our chief information security

play18:10

officer having already served in this

play18:11

capacity for a long time spent our

play18:13

planing for quite some time Mark Jacob

play18:15

Kevin S Matt and Josh will all report to

play18:18

me me being Sam I have over the past

play18:20

year or so spent most of my time in the

play18:21

non-technical parts of our organization

play18:24

I'm now looking forward to spending most

play18:25

of my time on the technical and product

play18:26

parts of the company very interesting

play18:28

that has been stuck doing non Tech stuff

play18:30

and this shift is so he can go back in

play18:32

the weeds again leadership changes are a

play18:34

natural part of companies especially

play18:36

companies that grow so quickly and are

play18:38

so demanding I obviously won't pretend

play18:39

it's natural for this one to be so

play18:41

abrupt but we're not a normal company I

play18:43

think the reasons Mira explained to me

play18:44

there's never a good time anything not

play18:46

abrupt would have leaked she wanted to

play18:48

do this while open AI was in an upswing

play18:49

all of these things made sense for it to

play18:51

be now we can both talk about this more

play18:53

tomorrow during all hands thank you all

play18:55

for your hard work and dedication Sam

play18:59

kind of nuts that they're doing a full

play19:01

restructuring internally and definitely

play19:03

lends Credence to the leak info we just

play19:07

read from Reuters because it seems like

play19:10

they're ready to fundamentally change

play19:12

how open works and go all in on Startup

play19:15

mode very very interesting I have

play19:21

concerns

play19:22

but yeah we'll see where this one ends

play19:25

up I'm using CLA right now anyways but

play19:27

I'm curious what y'all think do you like

play19:29

the open AI drama are you tired of me

play19:30

covering this and until next time peace

play19:32

nerds

Rate This

5.0 / 5 (0 votes)

Related Tags
AI GovernanceOpenAI ShiftNonprofit ChangesTech LeadershipAI SafetySam AltmanFor-profit AIMicrosoft InvestmentsAI DramaStartup Mode