OpenAI Battles Safety Concerns and High-Profile Exits | Pivot

Pivot with Kara Swisher and Scott Galloway

Ask about Pivot with Kara Swisher and Scott Galloway

Episode Description

OpenAI is getting messy again! Kara Swisher and Scott Galloway discuss the recent high-profile exits, and former employees voicing concerns about AI safety.

Read the full transcriptHide the full transcript
  1. things are getting messy at open AI again oh man is this the most like telen Nolla of a Company the company had other
  2. high-profile departure again this week with the resignation of yan Leica uh the head of super alignment which is the
  3. team focused on AI safety that's the words we're going to be all super aligned uh Leica explained his departure
  4. in a series of social media posts saying in part that open AI safety culture and
  5. processes have taken a backseat to shiny products and there's been a bunch of shiny products they showed off last week
  6. uh open it's not the things are not unrelated um open AI has dissolved that super alignment team the company told
  7. Bloomberg the group will be integrated across research efforts to help achieve safety goals uh Sam Alman put out a
  8. statement and also did open a co-founder Greg Brockman uh sharing their view of the future they said the company has
  9. quote raised awareness of the risks and opportunities of AGI so the world can better prepare for it whether they're
  10. preparing for it or not is a big question there's been at least 11 high-profile exits in the last few
  11. months um you know this is an issue again of speed versus safety they have been rolling out the products because
  12. they're deathly terrified of getting rolled over by the big companies I can feel it I can feel such a Netscape
  13. moment for them um you know uh we'll see uh we'll see what'll happen here but it's definitely a company still shaking
  14. off or dealing with these issues that they've had these two different types of people who are um involved in this
  15. company which is some that are think this is a a risk to humanity others who are like calm the [ __ ] down let's make
  16. some stuff and we'll figure it out later um uh one of the things that got a lot of reporting was open aa's offboarding
  17. agreements that have non-disclosure and non-disparagement Provisions not uncommon but theirs were particularly
  18. stringent if a departy employee violated these Provisions they were in danger of
  19. losing all their vested Equity according to Vox Sam mman confirmed in a tweet there was a provision about potential
  20. Equity cancellation for departing employees but it was never en forced the company is currently changing that
  21. language uh it sounds like they're just like just tough customers on that thing um it is further than other people do
  22. it's usually more Talent friendly in general in Silicon Valley Scott what are your thoughts on all this when Ilia was
  23. part of the board yeah when when he was part of the board that fired Sam Alman if you're going to stab the prince you
  24. better kill him when he came back Ilia became the information age equivalent of prosan he was dead man walking he just
  25. wasn't going to survive oh no feelings for firing me come on water under the bridge that just wasn't going to happen
  26. and this is similar to meta the fastest way to get a severance check is to go to
  27. work for the trust and safety team because every once in a while in response to real heat they'll pretend to
  28. give a good godamn and they'll create a trust and safety team I doubt Mark listens to them or cares about them and
  29. then under the cover of dark fires most of them I like the fact that open AI is becoming more like what they really are
  30. and that is they are a for-profit company and they're not pretending tough mothers and they're not pretending to be
  31. any anything I'd rather them be like New Yorkers that's that's one of the reasons
  32. I love New York versus doing business in California is they don't pretend to be something they aren't
  33. and this is a four profit companies and open AI are going to be so good at making profit they shouldn't be trusted
  34. to do anything else and the fact and first off and also well except they were founded with a slightly different idea
  35. but go ahead yeah and then when they took 11 billion those people wanted their money back so they should have
  36. never taken that money but the one thing the one compensatory thing here is that
  37. any group of people that decides to call themselves super alignment should be fired that is I thought you'd like that
  38. is those people should endure a certain amount of ridicule and pain what are your thoughts well you know AI is going
  39. to kill us Scott I mean it's a really interesting thing because these people are you know there's a group over in
  40. anthropic that are much more concerned in that regard and I think they have a right to be absolutely I think listen
  41. I've always been a safety let why are you not paying attention to any bit of safety from the very get-go and so I
  42. would naturally be affiliated with the super alignment people at the same time for them to think this is anything other
  43. than a for-profit institution is kind of well then go on a board go on a you know
  44. go to Stanford and and become a high-profile naysayer of these things or write a book like burnbook right um
  45. because they're inside these companies I think that because of the amount of money here there's just no way people
  46. aren't going to be aligned around the money making and you know some people are like oh you like Sam mman I said I
  47. said yes but he's a tough [ __ ] I was like are you kidding he's so aggressive he's so much like all the
  48. people he you know he's just like he's not like Elon because elon's a toxic piece of [ __ ] sometimes and most of the
  49. time uh I don't consider Sam like that but he is interested in not having this company die you know he is interested in
  50. making it the most dominant I have no question he's hyper aggressive and just as feral as the rest of them and so um
  51. you know they're going to have these things because what they want to do is pretend that they care about this safety
  52. stuff which they do peripherally you know more than other people I guess they bring it up more but they don't they
  53. just don't care about that issue and they're not they're not they're not going to be around to see the machines
  54. eating Us Alive or something like you know what I mean like it's it's not their job it's it's well but it this is
  55. the problem its birth was like it was like two it was like a hippie parent and a a non- hippie parent right and they're
  56. fighting forever this group these groups of people are fighting forever and so they are never going to be this whole
  57. company isn't going to be in alignment it started out out of alignment and they're never going to I know that I'm
  58. using a term about the idea of making sure it's safe but they will never this company will never it might be what
  59. kills them right it might be what kills them because they're going to get unusual it's like Google saying don't be
  60. evil that was this one big mistake like why did they do that right because they're evil not evil that's too far but
  61. you know what I mean like that that that whole cosplaying about being heroic has
  62. always been a problem for that's not these companies always figure out they're like okay unless we're re Ben
  63. and Jerry's let's be honest folks we all want our own big home we all want to take care of our kids we all want a
  64. broader selection set of mates than we deserve so let's let's stop let's stop pretending where it's dangerous is that
  65. we keep thinking that Sam Alman is actually a better generation of leaders we don't need to be as worried about AI
  66. because Sam's in charge and he'll speak in hush tones and say how concerned he is and what it does is it dampens the
  67. urgency in the need to elect people who can craft legislation to regulate these guys that's right it's not Sam's job
  68. that's exactly right his Sam is doing his job he's firing people who get in the way of him making being the no if
  69. Sam were to lose to to gemini or to llama or whatever or xai they would say yeah but he was more ethical and more
  70. concerned and we love him for that he's not going to get a statue for that let's
  71. just say there have to be ethical things if you're working for this kind like it's working for Palante here you're
  72. going to make defense department stuff don't work there like that you know or you know the Google employees that's
  73. different because Google sort of gave them an in to to complain because they were like we're better than this but
  74. they're really not better than this right so I think when you you sort of you know you you per have a performative
  75. nature of being being heroic you're going to be slapped later cuz you're going to let people down period and it's
  76. going to be very clear and again the reason my first line in my book was and so it was capitalism after all sticks
  77. that's what it is that's what's happening here and I think you all should go off and form a group of people
  78. that's that scares the [ __ ] B Jesus out of of the potential and you go up to Congress and you march in those offices
  79. and you explain to them why they need to make legislation that's what that's what
  80. you need to do I think when people all backed Sam a lot of people are like oh they love them I'm like no they love the
  81. money and they want to make money here and they you know and and they want to make they want to be not just make money
  82. they want to be at the coolest company making the [ __ ] they want to be at the winning company that makes them Rich
  83. yeah and also is the coolest company right that's more than that more than that anyway we'll see what happens I
  84. thought Yan's um I thought his his series of tweets was interesting but it doesn't really I'm sorry Yan do
  85. something else

Brief Summary

OpenAI faces turmoil as Yan LeCun resigns, raising concerns about safety culture amid rapid product development.

Key points

  • High-Profile Departures at OpenAI. OpenAI faces turmoil with multiple high-profile exits, raising concerns about its safety culture.
  • Conflict of Speed vs. Safety. The tension between rapid product rollout and safety concerns is palpable at OpenAI.
  • Sam Altman's Leadership Style. Sam Altman's aggressive approach to leadership raises questions about ethics in AI development.

Related episodes