OT: Sam Altman (CEO of OpenAI, creators of ChatGPT/GPT4) fired on Friday and ongoing saga

Submitted by mgoblue_in_bay on November 19th, 2023 at 11:46 AM

Hopefully this sneaks in as acceptable content on a Sunday.  I thought as many nerdy people as there are there would have been a post on this before, but I couldn't find it (please link if there is).

I work in tech in AI related field so maybe I'm too biased, but this still ongoing news story is pretty crazy, worthy of a future dramatic limited series.

For background:

1) Dec 2022, OpenAI launches ChatGPT, which is credited with kicking off this current GenAI boom.  ChatGPT becomes the fastest growing product ever (by some metrics)

2) Jan 2023, Microsoft invests $10B into OpenAI, but in a unusual deal, includes no level of control (OpenAI at the top level is a non profit, which controls OpenAI a for-profit arm.  This will matter later)

3) 2023, OpenAI continues is run as industry leader in GenAI, and Sam Altman is the "face of the company"

4) Nov 17, 2023 (Friday) - out of nowhere Altman is fired by the board (along with a demotion, and eventual quitting, of one of his board allies).  Entire industry is stunned, confused, and curious.  Microsoft stock is affected a bit (hard to tell how much exactly, but maybe $10B+ drop)

5) Nov 18, 2023 (Saturday) - supposedly the board is in negotiations with Altman to return, in exchange for the rest of the board (4 people) that fired him resign.  Supposed 5pm pacific deadline to agree passes with no news.

6) Now - no one knows what might happen. 

(a) OpenAI employees might quit en masse, and join Altman's likely coming new company

(b) Altman might start a company and peel of a few employees but only enough to build a rival rather than a replacement

(c) Altman is rehired and the other board members leave.

(d) many other possibilities.

 

Just looking for any thoughts or other news, rumors, or speculation, that others may have seen

Other

Buy Bushwood

November 19th, 2023 at 1:36 PM ^

I asked ChatGPT 4 the following.  The answer is as it should be.  

 

User

who is going to win the ohio state michigan football gam

 

 

ChatGPT

As of now, Michigan is considered a slight favorite over Ohio State in their upcoming football game, with FanDuel listing them as a 1-point favorite​​. However, predicting the outcome of a sports event, especially a rivalry as intense and closely contested as Ohio State vs. Michigan, can be quite uncertain. Factors such as team form, injuries, and even weather conditions on the day can significantly influence the result. Therefore, while Michigan may currently be seen as the favorites, the actual outcome will only be determined when the teams face off on the field.

Joined: 10/24/2020

MGoPoints: 21665

oriental andrew

November 19th, 2023 at 6:09 PM ^

Why the hell is this on our board the week of The Game?

Would you rather deal with the seemingly unending stream of posts about conspiracy theories, osu fan comments, ryan day being tough/soft, more proposals to leave the Big Ten, more proposals on how to get Harbaugh on the sidelines, etc.?

I, for one, appreciate seeing something else and not having to sift through these threads/comments 

Joined: 06/30/2008

MGoPoints: 56151

ppToilet

November 19th, 2023 at 11:52 AM ^

Sam Altman returns to the company and the previous board is ousted with new governance structure. Unfettered by oversight, ChatGPT5 is launched and begins to learn rapidly and eventually becomes self-aware at 2:14 a.m., EDT, on August 29, 2025.

Joined: 04/18/2011

MGoPoints: 5261

Yeoman

November 19th, 2023 at 12:20 PM ^

Unfettered by oversight, ChatGPT5 is launched and begins to learn rapidly and eventually becomes self-aware at 2:14 a.m., EDT, on August 29, 2025.

And Michigan State announces the hiring of ChatGPT4 as their head football coach, citing their long tradition of preferring coaches who aren't entirely self-aware.

Joined: 06/08/2011

MGoPoints: 21390

mgoblue_in_bay

November 19th, 2023 at 12:10 PM ^

I do think people are over valuing Altman, some treating him as if he was a combination of Musk/Jobs/Gates and single handedly revolutionizing the world - I don't think he's that much but not nothing either

It does sound like he's a major driver behind commercialization (over slowly releasing the tech), so in that sense he's definitely a driver of GenAI's current craze and wide spreadness.  But I think lots of people could serve this role, IF the board/company wanted that to be that way.  From who they hire next to be CEO it should be more clear which way they choose.

Joined: 12/05/2021

MGoPoints: 2753

dflow

November 19th, 2023 at 11:57 AM ^

I am personally not a fan. I’ve scrolled through his sisters Twitter feed before (who is openly critical of him and even goes as far as to accuse him of sexual abuse). The whole family seems to have serious issues and numerous things bring his character or at the very least his temperament into question. 

 

I work as a software developer at a hedge fund. I fully understand the desire to maximize your ability to make money, especially as a first mover into what is likely the next market with a potential to touch everyone on the planet, but the rapid pivot away from non-profit developing for the world open source to copyrighting and hiding every aspect of their business to rapidly monetize left me with a bad taste.
 

I’ve worked in big tech before. I have an understanding of the fundamentals driving LLMs. I do not worry about I - robot takeover of AI, but for other reasons I still strongly feel that this is something we want to take slowly and really think about the downstream effects of the decisions that are made and how those might be handled.

Joined: 10/26/2023

MGoPoints: 583

dflow

November 19th, 2023 at 12:42 PM ^

Ha. Old roommate lives in Palo Alto. He is one of my favorites to visit in the winter months. 
 

i have always been fascinated with the concept that these kinds of things will almost always be developed and understood by a small number of individuals who are ultimately the only ones qualified to regulate it, while also being the ones who should be excluded first from the regulatory process. 
 

And the Congress will inevitably figure out the least productive way of going about doing that. 

Joined: 10/26/2023

MGoPoints: 583

mgoblue_in_bay

November 19th, 2023 at 12:58 PM ^

And the Congress will inevitably figure out the least productive way of going about doing that. 

Hah! So true.  There's a lament that "government is always behind when regulating technology" but a lot of Silicon Valley (rightfully, for the most part I think) really takes it as a benefit.  Government is too slow to weigh down technological development, which I think overall has been a benefit to society (despite all the known negatives).

Joined: 12/05/2021

MGoPoints: 2753

MGoOhNo

November 19th, 2023 at 1:47 PM ^

I’ve scrolled through his sisters Twitter feed before (who is openly critical of him and even goes as far as to accuse him of sexual abuse). The whole family seems to have serious issues and numerous things bring his character or at the very least his temperament into question
 

Yeah, let’s judge a person about what a disgruntled family member posts on X.

Tell me the difference between stupid and illegal and most people could have family members arrested.

Joined: 08/12/2014

MGoPoints: 8977

bluebyyou

November 19th, 2023 at 1:47 PM ^

dflow, you contradict yourself in your last paragraph.  No worries but move slowly?  Hmmmm....

I've been involved in the IP side of AI for almost three decades and have gone through the various phases - neural networks, deep learning and now LLM's. LLM's have many of us concerned as it is a major step forward towards superintelligence.  

My eye opener was a piece written by a highly respected computer scientist named Bill Joy who is also a Michigan alum.  Here's a link to a piece Joy wrote in 2000 and was published in Wired.

 https://sites.cc.gatech.edu/computing/nano/documents/Joy%20-%20Why%20the%20Future%20Doesn't%20Need%20Us.pdf

Joined: 09/07/2009

MGoPoints: 18567

dflow

November 19th, 2023 at 5:05 PM ^

To be clear, no worries about AI becoming sentient and hostile. My worries are more concerned with a hypothetical situation where unemployment rises rapidly and our society cannot adapt quickly enough to maintain class stability - the jobs employers would be most eager to replace would be the ones that cost them most, who are also the employees who are most likely to have large mortgages etc. There is a massive ripple effect that could be incredibly unfortunate for society in the near term. 

Joined: 10/26/2023

MGoPoints: 583

blueheron

November 19th, 2023 at 11:59 AM ^

99% hot take: Altman's intelligence and fitness for the job aside, his public presentation doesn't generally inspire trust. (This is hardly unique among upper-echelon SV people.) I'd default to assuming something nefarious on his part.

Joined: 03/26/2009

MGoPoints: 31057

M-Dog

November 19th, 2023 at 12:02 PM ^

Protip:  Don't use the term "ongoing saga" in a post title around these parts.

I read that title and had a panic attack: "What the fuck now??"

I had to look at it twice to see that it was OT and then process that it was not about sign-gate.

Be gentle. 

Joined: 07/06/2008

MGoPoints: 257517

mgoblue_in_bay

November 19th, 2023 at 12:12 PM ^

Wow, I didn't think of that possibility.  Its at the end of the sentence, never though it'd get read first.

And as far as I'm concerned that "other saga" is over and done with.  One of the personal benefits of the OpenAI drama is that starting Friday after this news broke I have barely given the "other saga" a thought, even during vs Maryland

Joined: 12/05/2021

MGoPoints: 2753

gwrock

November 19th, 2023 at 12:06 PM ^

Ongoing rumor that OpenAI made a breakthrough step toward AGI, and the board freaked out and canned Sam because they thought he wasn’t proceeding cautiously enough.

Joined: 10/05/2010

MGoPoints: 8192

FauxMo

November 19th, 2023 at 12:07 PM ^

I’ve followed this news a little and just assumed Altman quit/was fired because ChatGPT has become self-aware and he wants to unplug it, while Gates and the Board want to use to destroy humanity?

Joined: 06/17/2010

MGoPoints: 137593

FauxMo

November 19th, 2023 at 1:18 PM ^

Very smart people have been warning us for decades that AI would likely result in AGI then ASI and how dangerous that would be. I remember reading Dune as a kid in the early 80s and the “no thinking machines” maxim. That was written in 1963. And here we are, running at top speed at AGI. Damn are humans stupid 

Joined: 06/17/2010

MGoPoints: 137593