After a week of incredible AI happenings, Episode 86 of the Artificial Intelligence Show examines the lawsuit filed by Elon Musk against OpenAI, the broader implications of AI and AGI, and the growing concerns among content creators about the rise of AI-generated content. All this and even more in our rapid fire section!
Listen or watch below—and see below for show notes and the transcript.
00:04:22 — Elon Musk Sues OpenAI
00:25:18 — AGI Timeline and Impact
00:40:54 — Creators fears about AI-generated content
00:49:51 — Figure raises $675M in Series B funding at a $2.6B valuation
00:52:30 — Apple’s abandoned quest to build a self-driving car, project Titan
00:55:17 — Salesforce’s Einstein Copilot
01:01:57 — Mistral AI positions itself as a rising star in the AI world
01:04:10 — Introducing the next generation of Claude
Elon Musk Sues OpenAI
Elon Musk is suing OpenAI, OpenAI CEO Sam Altman, and co-founder/president Greg Brockman for breach of contract. He says the company has violated a founding agreement in which it promised to develop AI for the good of humanity, not profit...And he wants the company to pay the price:
He's seeking a court order that would force the company to make all of its research and technology public. The court order would also compel OpenAI to give back all the money it made as a result of breaching the contract.
Musk was one of the founding members of OpenAI. He says they all agreed that any AI developed by OpenAI would be used to benefit humanity, not make money. And he says the seed funding he provided to OpenAI, allegedly more than $44M, was contingent on these promises.
He claims that OpenAI's commercial relationship with Microsoft violates the founding agreement:
“OpenAI, Inc. has been transformed into a closed-source de facto subsidiary of the largest technology company in the world: Microsoft,” the suit says. “Under its new board, it is not just developing but is actually refining an AGI to maximize profits for Microsoft, rather than for the benefit of humanity.”
This idea of AGI, or artificial general intelligence, is central to Musk’s suit: AGI is AI that can broadly perform many tasks as well or better than humans. Musk says OpenAI's GPT-4 model is a type of AGI—and that OpenAI agreed not to commercialize any type of AGI.
AGI Timeline and Impact
We are starting to see proof of the profound impact that even non-AGI will have on the economy and society. Klarna, a huge payments company, just revealed its AI assistant now does the jobs of 700 employees.
The AI assistant, powered by OpenAI, handles customer services chats. It chats with customers to do things like resolve service requests in different languages and manage refunds and returns.
This development reminded us of the series "Our AI Journey," authored by innovation experts Adam Brotman and Andy Sack. In a notable quote from the series, Sam Altman, in October 2023, addressed the implications of AGI for the marketing sector.
Altman predicted “that 95% of what marketers use agencies, strategists, and creative professionals for today will easily, nearly instantly and at almost no cost be handled by the AI — and the AI will likely be able to test the creative against real or synthetic customer focus groups for predicting results and optimizing.”
Klarna's AI assistant is an example of how AI can instantly make things better for businesses, but it is just the beginning. It is a sneak peek into the bigger shifts we are likely to see as AI keeps getting smarter and popping up in more places.
Creators are starting to have some fears about AI-generated content
Creators are starting to get seriously worried about AI, according to new reporting from The Information.The publication did a deep dive into how generative AI is impacting creators and media distribution platforms like YouTube and Netflix.
The first fear is that AI-generated content will begin to flood platforms. (Spotify, for instance, has already seen an influx of AI-generated music.)
The second is that audiences will develop an appetite for AI-generated content.
And the third is that the benefits of using generative AI will force everyone to produce AI-generated content to compete.
An example of this comes from YouTube creator, Kwebbelkop, a gaming megastar who has 15M+ subscribers. Kwebbelkop has experimented for years with fully AI videos. He even created a full AI persona at one point to star in his videos. (Though he has ended that particular experiment.)
He told The Information that AI now allows him to work with half the team and half the budget he did just 2 years ago.
He also said that AI's impact will be profoundly disruptive to creators, saying: "If you don’t embrace AI, it will replace you.”
Today’s episode is brought to you by Marketing AI Institute’s AI for Writers Summit presented by Jasper, happening virtually on Wednesday, March 6 from 12pm - 5pm Eastern Time. To register, go to AIwritersummit.com
Disclaimer: This transcription was written by AI, thanks to Descript, and has not been edited for content.
[00:00:00] Paul Roetzer: There's two things to consider when we talk about AI creating things. It can create outputs that people may find enjoyable. But that doesn't replace the artist or the creator themselves needing to find enjoyment out of the process of creating things.
[00:00:13] Paul Roetzer: Welcome to the Artificial Intelligence Show, the podcast that helps your business grow smarter by making AI approachable and actionable. My name is Paul Roetzer. I'm the founder and CEO of Marketing AI Institute, and I'm your host. Each week, I'm joined by my co host, and Marketing AI Institute Chief Content Officer, Mike Kaput, as we break down all the AI news that matters and give you insights and perspectives that you can use to advance your company and your career. Join us as we accelerate AI literacy for all.
[00:00:50] Paul Roetzer: Welcome to episode 86 of the Artificial Intelligence Show. I'm your host, Paul Roetzer, along with my co host,
[00:00:56] Paul Roetzer: Mike Kaput. We have. An [00:01:00] insane amount of stuff to cover today, to the point where Mike and I were literally just cutting like three things that were in RapidFire that are now going to be newsletter only.
[00:01:09] Paul Roetzer: So as Mike always tells you at the end, make sure you subscribe to the newsletter because there's all kinds of good stuff that does not make the cut each week. And today it is literally last minute because it is March 4th,
[00:01:20] Paul Roetzer: 10 a. m. Eastern time and we have had multiple things happen this Monday morning that have required us to update our plans for today.
[00:01:29] Paul Roetzer: So we got a lot to talk about today, Mike, it's going to be an interesting one.
[00:01:33] Paul Roetzer: Alright, so today's episode is brought to us by Marketing Institute's AI for Writers Summit, which is happening Wednesday, March 6th. You still have, if you're listening to this on the day this comes out on March 5th, you still have a day to get in.
[00:01:46] Paul Roetzer: It's from noon to five Eastern time. It is a virtual summit. There is a free option, thanks to our presenting sponsor, Jasper. So following last year's inaugural event in March 2023 that had over 4, [00:02:00] 000 writers, editors, and content marketers, we are back with the second edition of the AI for Writers Summit.
[00:02:06] Paul Roetzer: The agenda has the state of AI in writing, which is me doing the opening talk. Mike is doing generative AI writing tools and platforms you should know.
[00:02:14] Paul Roetzer: We're then going to have an AI Insights conversation with an IP attorney that I'm going to have a fireside chat with Krista Laser, who's an IP attorney.
[00:02:23] Paul Roetzer: We're going to have AI writing in the enterprise. that's going to be an awesome panel discussion about adoption of AI writing tools and platforms within companies. And then an AI in action session led by Mike and Cathy McPhillips. And that's going to be a complete breakdown of a content workflow powered by a bunch of different AI tools that Mike and Cathy are going to walk
[00:02:41] Paul Roetzer: through with an ending with a, about an hour of ask me anything with myself, Mike, and a couple of the other presenters.
[00:02:48] Paul Roetzer: So, AIwritersummit. com if you are a writer, editor, content creator, or lead, any of those people, I'd highly recommend checking it out. Again, it's from noon to five Eastern [00:03:00] time on Wednesday, March 6th. And again, thank you to our presenting sponsor, Jasper. Jasper is an AI co pilot for enterprises. They use AI to generate on brand content
[00:03:11] Paul Roetzer: that reflects brand style and voice. Jasper can do anything from write blog posts, to repurpose and rewrite content. And it sounds like you, because it
[00:03:20] Paul Roetzer: can securely store information about your company and products. Which also helps to cut down on the AI making stuff up, known as hallucinations. Jasper can even make suggestions on how to improve content performance based on trends in your content data.
[00:03:35] Paul Roetzer: As part of our AI for Writers Summit this week, Mike is going to talk, or Mike and Cathy are going to show off Jasper in one of the AI in Action sessions. giving a little demo. We're gonna, um Talk with Megan Keeney Anderson from Jasper as well. So
[00:03:49] Paul Roetzer: if you're looking to augment your marketing and content efforts with AI, Jasper is worth exploring.
[00:03:53] Paul Roetzer: You can go to jasper.
[00:03:55] Paul Roetzer: ai. That's J A S P E R. AI. So thanks again to [00:04:00] Jasper for helping us bring the AI Writers Summit to everyone on March 6th.
[00:04:06] Paul Roetzer: Okay, Mike, I, It's like, I don't even know where to begin, but we have a structure to try and get through this. It was a wild week last week, and we are off to another wild start this week. So let's see if we can put this in context for everyone.
[00:04:21] Mike Kaput: Sounds good, Paul.
[00:04:22] Mike Kaput: So, first up,
[00:04:23] Mike Kaput: Elon Musk is suing OpenAI, and he is suing OpenAI CEO Sam Altman and co founder slash president, Greg Brockman for breach of contract. Now, Musk says that the company, and by extension Altman and Brockman, violated a founding agreement in which they promised to develop AI for the good of humanity, not for profit.
[00:04:47] Mike Kaput: And he wants the company to pay a pretty significant price. He's seeking a court order that would force the company to make all of its research and its technology public. And the court order would also compel OpenAI [00:05:00] to give back all the money it's made as a result of breaching this contract. Now For anyone who doesn't remember, Musk was one of the founding members of OpenAI.
[00:05:09] Mike Kaput: Back when it was a non profit, he says that they all agreed at that time that any AI developed by OpenAI would be used to benefit humanity and not make money. Now in the lawsuit, he's saying that the seed funding he provided to OpenAI over the years, allegedly more than 44 million dollars in total, was contingent on those promises.
[00:05:33] Mike Kaput: And he says that now those promises have been broken. Musk seems to center around the claim that OpenAI's commercial relationship with Microsoft violates this founding agreement that they had. He said in the lawsuit, quote, OpenAI Inc. has been transformed into a closed source de facto subsidiary of the largest technology company in the world, Microsoft.
[00:05:58] Mike Kaput: Under its new board, it is not [00:06:00] just developing. but is actually refining an AGI to maximize profits for Microsoft rather than for the benefit of humanity. Now those three little letters are pretty important to this AGI, Artificial General Intelligence. AGI could be defined as AI that can kind of broadly perform many tasks as well or better than humans.
[00:06:23] Mike Kaput: Now Musk is literally saying in the lawsuit That GPT 4 is a type of AGI and that OpenAI has agreed not to commercialize any type of AGI and thus is in violation of its contract and the founding agreement of the company.
[00:06:41] Mike Kaput: So
[00:06:41] Mike Kaput: Paul, let's Let's just address the elephant in the room right away. Elon Musk has a highly, let's say, controversial reputation, and plenty of people are pretty skeptical about his motives, about him mouthing off online.
[00:06:59] Mike Kaput: Like, [00:07:00] how seriously should we take this lawsuit? Can you help us kind of separate the claims from the guy bringing
[00:07:05] Mike Kaput: them?
[00:07:06] Paul Roetzer: Yeah, it's, kind of hard to do that. Um,
[00:07:11] Paul Roetzer: but I mean, my overall take is I think the lawsuit is probably more show than it is, You know, meant to actually get them to have to pay this.
[00:07:20] Paul Roetzer: I think it might be in part just to force discovery, to force information to get out. But I, you know, I think it's helpful to understand the lawsuit, by revisiting the history of OpenAI and specifically, as you alluded to. the origins of OpenAI as a counterbalance to Larry Page and Google. So what I would suggest is, go back and listen to episode 73 of the AI Show podcast.
[00:07:47] Paul Roetzer: In that episode, I took a deep dive, this is November of 2023, I took a deep dive into, Sam Altman being fired temporarily from OpenAI
[00:07:55] Paul Roetzer: and why. And within that, went into AGI and superintelligence,
[00:07:59] Paul Roetzer: [00:08:00] who went into the origins of OpenAI. so, that episode will, will help you understand the lawsuit way better, but I will give a very quick kind
[00:08:08] Paul Roetzer: of snapshot of, that episode, but, but within the context of some things that have happened since.
[00:08:16] Paul Roetzer: So first. Larry Page in Google, so Larry Page, one of the co founders was Sergey Brin, um, Purchased
[00:08:22] Paul Roetzer: DeepMind in January of 2014. So DeepMind, if you recall, is a London AI research lab that was founded in 2010 by Demis Hassabis, Shane Legg, and Mustafa Suleyman. Demis and Shane are still at DeepMind. Mustafa has since gone on to co found Inflection.
[00:08:39] Paul Roetzer: So. In 2014, Musk was actually an early investor in DeepMind. He did this after meeting Demis Hassabis in the early days of DeepMind at the SpaceX labs, actually,
[00:08:53] Paul Roetzer: and became concerned about the direction DeepMind was going and the potential of their [00:09:00] AGI.
[00:09:00] Paul Roetzer: So in 2014, when Google was buying DeepMind, Musk asked actually tried to stop it.
[00:09:07] Paul Roetzer: He tried to put
[00:09:07] Paul Roetzer: an investor, investor group together to purchase DeepMind and tried to convince Demis that selling to Google was a bad idea, that in quotes,
[00:09:18] Paul Roetzer: the future of AI should not be controlled by Larry, referring to Page in. The lawsuit. So this is actually from the lawsuit where that came in.
[00:09:27] Paul Roetzer: now why did he not want DeepMind under the control of Larry Page and Google? So for that, we have to go to another New York Times article that tells the story about Elon Musk's 44th birthday in July of 2015. So at that party, Larry
[00:09:43] Paul Roetzer: Page, who was a friend of Elon's, was there. And Larry and Elon, sitting by a fireside, started having a conversation around artificial intelligence. the story in the New York Times goes on to say people sort of started kind of arriving around them and listening in because [00:10:00] the tone, started getting contentious and they started having a debate about
[00:10:04] Paul Roetzer: AI. And whether or not it would ultimately elevate humanity or destroy it. So Page said humans would eventually merge with AI machines. One day there would be many kinds of intelligence competing for resources and the best would win.
[00:10:19] Paul Roetzer: Musk said we're doomed if that happens. Machines will destroy humanity. Page insisted his utopia should be pursued, and he called Mr. Musk a speciist as a person who favors human over the digital lifeforms of the future. So this is July of 2015, and that insult, as the New York Times tell it, was
[00:10:38] Paul Roetzer: the last straw. So now keep in mind, 2015, Musk has already lost out on acquiring DeepMind, which he didn't think Google should control.
[00:10:47] Paul Roetzer: Google now controls DeepMind and Larry Page is saying to, to Musk, basically like, Hey, if the machines win, let it, let it be. So, Musk is obviously very concerned around the [00:11:00] future and Paige's Cavalier approach to AI. And that
[00:11:03] Paul Roetzer: is the last straw that leads to the founding of OpenAI. So fast forward to December 15, about six months later, Sam Altman, Elon Musk, Ilya Skova, Greg Brockman.
[00:11:14] Paul Roetzer: form a nonprofit organization with the stated goal, um,
[00:11:18] Paul Roetzer: To assemble the best researchers in the field to pursue the goal of building AGI in a safe way as this counterbalance to Google.
[00:11:25] Paul Roetzer: Fast forward, april, 2018, OpenAI releases their charter with principles we use to execute OpenAI's mission, which is again, quotation, to ensure that.
[00:11:35] Paul Roetzer: AGI, which, by which we mean Highly Autonomous Systems that outperform humans at most economically valuable work, benefits all of humanity. We will attempt to directly build safe and beneficial AGI. So at this moment, it is supposed to be open, they're supposed to share what they learn. That is the premise that Musk And Altman built OpenAI and supposedly agreed to. Then on March [00:12:00] 2011, 2019, three years before the launch of GPT 4, they announced
[00:12:05] Paul Roetzer: the formation of openAI LP. and this is where they created this capped profit company. So we're now in 2019, almost, you know, three and a half years after the creation of OpenAI, they're creating this capped profit company. In this, they state the mission still comes first.
[00:12:23] Paul Roetzer: They're still pursuing safe AGI. but they need more money, basically. And so this change allows them to go get money from Microsoft eventually. then in June 2023, so we're now fast forwarding.
[00:12:39] Paul Roetzer: we're about five months before Altman gets fired as the CEO.
[00:12:43] Paul Roetzer: They publish an update to their structure and within that update it states, the fifth thing it states, is that the board determines when they've attained AGI. So this is where it starts becoming very
[00:12:55] Paul Roetzer: critical to understand the lawsuit and what Musk is claiming with GPT 4 being an [00:13:00] early form of AGI basically.
[00:13:02] Paul Roetzer: So it says the board determines when we've attained AGI. Again, we mean highly autonomous system that outperforms humans at most economically valuable work. Such a system, and this is the critical part. Such a system is excluded from IP licenses and other commercial terms with Microsoft, which only apply to pre AGI technology.
[00:13:22] Paul Roetzer: So, if GPT 4 is AGI, as Microsoft's own paper, Sparks of Intelligence, claims it to be, then Microsoft has no rights. to license the technology.
[00:13:35] Paul Roetzer: So, again, then we jumped to
[00:13:38] Paul Roetzer: November 13th, 2023, this is four days
[00:13:42] Paul Roetzer: before Altman is fired. He confirms in a Financial Times interview that they are training GPT 5 and that they don't know what it will be capable of.
[00:13:52] Paul Roetzer: He specifically says, I quote, until we go train that model, it's like a fun guessing game for us. We're trying to get better at it because I think it's [00:14:00] important from a safety perspective. to predict the capabilities, but I can't tell you, here's exactly what it's going to do that GPT 4 didn't.
[00:14:10] Paul Roetzer: but then the night before Altman is fired, 10 days after OpenAI's dev conference where they introduced GPTs,
[00:14:16] Paul Roetzer: he's on stage and he kind of goes a little bit further.
[00:14:19] Paul Roetzer: He says, I think it's going to be the greatest step forward that we've had yet so far and the greatest leap forward of any of the big technological revolutions we've had. super excited. I can't imagine anything more exciting to work on.
[00:14:32] Paul Roetzer: And on a personal note, and this is where
[00:14:34] Paul Roetzer: we kind of thought maybe this is part of why he got fired, just in the last couple of weeks I have gotten to be in the room when we sort of pushed, the sort of veil of ignorance back and the frontier of discovery forward.
[00:14:46] Paul Roetzer: And getting to do that is like a professional honor of a lifetime. So it's just fun to get to work on that.
[00:14:53] Paul Roetzer: So, now we are in November 2023.
[00:14:56] Paul Roetzer: AGI is maybe starting to emerge, maybe [00:15:00] not. But if it is, then Microsoft doesn't like it. So things are getting complicated. And then, and then Sam gets fired.
[00:15:05] Paul Roetzer: So then I do the episode 73 talking about all this.
[00:15:09] Paul Roetzer: And at the end of that episode, I had eight final thoughts. And I'm going to linger on two of them for a moment. So number four was. I feel like Elon will play a role in this at some point.
[00:15:19] Paul Roetzer: Remember, this is why he started OpenAI and then left it, and why he's building
[00:15:24] Paul Roetzer: XAI and Grok. So that has become true now with our lawsuit. The other was, number eight, the key question in all this, in my opinion, is what did Sam witness in that room, and has OpenAI made a breakthrough in the pursuit of AGI and superintelligence? So certainly
[00:15:40] Paul Roetzer: Elon thinks they have. So now come back to the lawsuit.
[00:15:45] Paul Roetzer: So in the lawsuit, musk, this is right from the lawsuit. Mr. Musk has long recognized that AGI poses a grave threat to humanity, perhaps the greatest existential
[00:15:55] Paul Roetzer: threat we face today. Our entire economy is based around the fact that humans work together and come up with the [00:16:00] best solutions to a hard task. If a machine can solve nearly any task better than we can, That machine becomes more economically useful than we are.
[00:16:09] Paul Roetzer: Then the kind of lawsuit goes on to lay out all this stuff about DeepMind and Google and the founding of OpenAI. And that's where the beef appears to begin. So Mr. Musk was, again, quoting from the lawsuit, was a driving force or moving force behind the creation of OpenAI Inc., contributing the majority of the funding in the first several years, advising on research directions, and most importantly, recruiting some of the world's leading scientists and engineers to work at the nonprofit, including.
[00:16:32] Paul Roetzer: Ilya Sutskovo, who still, since the firing of Sam Altman, is kind of AWOL, like not on Twitter, and we don't even know if he still works for OpenAI. So here's my overall take to kind of set the stage of what I think is going on. Um,
[00:16:47] Paul Roetzer: so the irony of the lawsuit to me is, Musk has exited OpenAI, like he's no longer involved.
[00:16:55] Paul Roetzer: He is building his own large language model, Grok, with a K, [00:17:00] which is not open source.
[00:17:01] Paul Roetzer: So he's now building his own model that's not open source. He is build building full self driving cars with computer vision technology and real world data that may actually be critical to unlocking true AGI, because people think it's critical for worldview.
[00:17:14] Paul Roetzer: He's building humanoid robots at Tesla. That can learn from multi modal data and take actions in the real world. And he is merging humans and machines with Neuralink, where they're literally embedding hardware into human brains to merge the humans with the machines, because
[00:17:29] Paul Roetzer: he thinks that's critical to our survival.
[00:17:32] Paul Roetzer: So everything he said. He worried about Google doing he, he is doing, this doesn't change the legal merit and whether or not there was a contract and like that's for the courts to decide, but it's more likely to me that this is about stopping Sam and OpenAI over some beef he has with them. And the reason why I actually think.
[00:17:53] Paul Roetzer: this is a hundred percent, what's happening is one other piece that isn't in the lawsuit.
[00:17:58] Paul Roetzer: I don't think it's 48 [00:18:00] pages. I didn't read the full thing yet, but I doubt this is in there. In December, 2023. Cade Metz and a couple other journalists from the New York Times published an article called Ego, Fear, and Money, How the AI Fuse was Lit. And I'm just going to read this excerpt, Mike, and then you and I can have a conversation about
[00:18:18] Paul Roetzer: this. So keep in mind, again,
[00:18:21] Paul Roetzer: Musk leaves OpenAI at some point. and they go raise money and his beef is that they form this for profit and raise the
[00:18:28] Paul Roetzer: money. Well, why did they do that? Let's, let's revisit history. So here's the New York Times. Initially, Mr. Musk wanted to operate OpenAI as a non profit free from the economic incentives that were driving Google and other corporations.
[00:18:43] Paul Roetzer: But by the time Google wowed the tech community with its Go stunts, he's referring to AlphaGo in 2016. Mr. Musk was changing his mind about
[00:18:52] Paul Roetzer: how it should be run.
[00:18:54] Paul Roetzer: He desperately wanted OpenAI to invent something that would capture the world's imagination and close the gap with [00:19:00] Google, but it wasn't getting the job done as a nonprofit.
[00:19:04] Paul Roetzer: In late 2017. 2017. He, mr. Musk, hatched a plan to wrest control of OpenAI from Altman and the other founders and transform it into a commercial operation that would join forces with Tesla and rely on supercomputers the car company was developing, according to four people familiar with the matter.
[00:19:24] Paul Roetzer: When Mr. Altman and others pushed back, Mr. Musk quit and said he would focus on his own AI work at Tesla. In February, 2018, he announced his departure to OpenAI staff on the top floor of the startup's offices. Three people who attended the meeting said when he said that OpenAI needed to move faster, one of the researchers retorted at the meeting that Mr.
[00:19:46] Paul Roetzer: Musk was being reckless. Becoming the very thing that he was building OpenAI not to be. Mr. Musk called the researcher a jackass and stormed out, taking his deep pockets with him. OpenAI suddenly needed [00:20:00] new financing in a hurry. Mr. Altman flew to Sun Valley for a conference and ran into Satya Nadella, Microsoft's Chief Executive.
[00:20:08] Paul Roetzer: A tie up seemed natural. Altman knew Microsoft's CTO, Kevin Scott. Microsoft had bought LinkedIn from Mr. Hoffman, an OpenAI board member. Mr. Nadalya told Mr. Scott to get it done. The deal was closed in 2019. Altman and OpenAI formed a for profit company under the original non profit, and they had a fresh billion dollars in capital.
[00:20:29] Paul Roetzer: So why are they a for
[00:20:31] Paul Roetzer: profit company? It's because Musk wanted to do it first.
[00:20:34] Paul Roetzer: They got pushback. He left,
[00:20:36] Paul Roetzer: took his money with him, and they needed the money to do what they were doing. So the whole lawsuit is based on this premise of OpenAI changed their direction, which is the direction he tried to force them in anyway.
[00:20:48] Paul Roetzer: So I am not an attorney, but when you understand this full context and you look back and see he actually tried to take control and become the CEO of OpenAI and roll it into [00:21:00] Tesla. What are we even talking about other than a vendetta against SAM and OpenAI? So, I don't know. I mean, I'm sure there's other stuff that we don't know about, but Cade Metz is a pretty reliable guy who's deeply sourced with both Elon and SAM.
[00:21:16] Paul Roetzer: And I'm guessing the New York Times didn't make up any of that stuff. And
[00:21:21] Paul Roetzer: if that's the case, then this is just all it. a show,
[00:21:26] Paul Roetzer: basically.
[00:21:27] Mike Kaput: This
[00:21:29] Mike Kaput: is, I think someone online, when you posted about this, mentioned this being like a soap opera. I feel like it's almost more like Game of Thrones over here.
[00:21:38] Paul Roetzer: It's, it's so wild and I mean, I,
[00:21:41] Paul Roetzer: you kind of assumed this was coming because there was also the part we didn't get into where Elon Musk summoned Sam Altman to the Twitter offices last year and basically told him to bring the founding documents of OpenAI with him. and
[00:21:55] Paul Roetzer: that meeting didn't go well. And it was pretty apparent then that Musk was going to sue [00:22:00] OpenAI and Sam. So nothing like this lawsuit was not a surprise to me at all. I've been waiting for this to happen. I think the question now becomes, what do we learn from it? This isn't about money for
[00:22:12] Paul Roetzer: Musk. It's not, he's not suing for like his money back is 44 million and then some, or, you know, Sam already offered him when they met at Twitter, he offered him a stake and Elon didn't want it.
[00:22:22] Paul Roetzer: So this isn't about money. This is, this is about Elon. Losing and not liking it and, and maybe he believes in this AGI is the threat to humanity, but then why is he building it himself? Like, it just doesn't add up.
[00:22:38] Paul Roetzer: and so the question becomes like, well, what will we learn from this? Will they get to the point where there's discovery
[00:22:45] Paul Roetzer: and we actually find out why was Sam actually fired? And What did sam see? And, and does OpenAI think? GPT 4 is a form of AGI and where is Ilya in all of this? And I, you know,
[00:22:58] Paul Roetzer: I don't think elon's going to settle [00:23:00] those. So again, not an attorney. I think either a judge throws this out because there's no contracts, no proof. And that was the one, lawyer that you and I have talked
[00:23:10] Paul Roetzer: about on the show before, Cecilia, Ziniti. Ziniti, she did a great thread on this, we'll
[00:23:17] Paul Roetzer: put in the show notes, but her thing was like too long, don't read PR fireworks and fun to read intrigue and philosophizing about AGI, but legally a stinker because there's no contract breach. So basically there was no con, official contract there, assuming a few things together, represent a contract, but there is no contract between Elon that was broken. Breached according to her. So there's a chance like this just gets thrown out, but there's a chance that it does move forward and then the discovery could be really interesting and make for some fascinating conversations on our show, I guess.
[00:23:50] Mike Kaput: So lots of different things can be true at the same time. One, there may be broader actual concerns [00:24:00] about AGI, both from Musk and from others related to what OpenAI is or is not developing, but also Elon Musk can just be extremely vindictive.
[00:24:11] Paul Roetzer: Totally. And that's the thing. It's like, maybe there is a breach of contract. Maybe there is some legal belief that there was actually consideration involved and there was a contract to be had and
[00:24:21] Paul Roetzer: it was breached. Like you and I aren't, aren't going to be the ones to figure that out. maybe Elon believes there is, so maybe it's like, if this isn't a PR show, maybe to him, he
[00:24:31] Paul Roetzer: truly does feel wronged and he feels like they went the wrong direction and they couldn't do that because of their non profit structure and they were, it was illegal.
[00:24:38] Paul Roetzer: And maybe he does believe that. Um,
[00:24:42] Paul Roetzer: so yeah, like you said, all things can be true here and it'll be for the courts to decide and for us to, I guess, just watch with intrigue.
[00:24:50] Mike Kaput: And if you have let all heard of or read the Elon Musk biography by Walter Isaacson. The [00:25:00] contradictions and intensity of this unfolding should not remotely surprise you. This playbook has happened before.
[00:25:08] Paul Roetzer: Yeah, you and I,Ihink we talked about it a little bit last week. I'm almost done now with the book.
[00:25:12] Paul Roetzer: Um,
[00:25:14] Paul Roetzer: And yes, there is zero about this that is surprising once you've read the book.
[00:25:18] Mike Kaput: Yeah, so somewhat related to this or the next big topic we want to talk about is Zooming out a little bit, we're actually starting to see proof of the profound impact that even non AGI artificial intelligence may be about to have on the economy and society. So in one example of this, we actually just got an announcement from a company called Klarna, which is a massive payments company in e commerce, and they just revealed that they now have an AI assistant that does the jobs of 700 [00:26:00] employees.
[00:26:00] Mike Kaput: The AI Assistant is powered by OpenAI, and it basically just handles customer service chats across the platform. So it can chat with customers to do things like, say, resolve customer service requests in different languages, and help you manage refunds and returns. Klarna claims that in just one month of this thing being live globally, the assistant is already doing the work of 700 full time agents.
[00:26:29] Mike Kaput: And so far it's conducted 2. 3 million conversations, which the company says is a full two thirds of the customer service chats it regularly handles. Not to mention it seems like this AI assistant is doing pretty solid work. They say that it produces customer satisfaction scores that are on par with human agents, and it's more accurate and much faster at resolving customer requests.
[00:26:57] Mike Kaput: So Karnas says the [00:27:00] average time to resolve requests dropped. From 11 minutes to 2 minutes, which is pretty significant. Overall, the company claims that this AI assistant, in 2024 alone, is going to create an additional 40 million in profit. Now, as you might expect, Korn, as CEO, talked all this up in a press release, but what I found really interesting is that he also basically just came out and hinted that society needs to prepare for advanced AI like this.
[00:27:32] Mike Kaput: This is To be very clear, not AGI, but even very hyper, competent artificial intelligence is something we really need to pay attention to. He said, quote, We are incredibly excited about this launch, but it also underscores the profound impact on society that AI We'll have. We want to reemphasize and encourage society and politicians to consider this carefully, and we believe that a [00:28:00] considerate, informed, and Steady stewardship will be critical to navigate through this transformation of our societies.
[00:28:08] Mike Kaput: I thought that was a sort of deep and ponderous way to end a press release overall on a new AI assistant that you're using. So like we said, this isn't AGI. This isn't proof of AGI at all. This is simply an illustration of the profound impact that even non AGI can kind of have on knowledge work as we know it.
[00:28:30] Mike Kaput: And. In the process, I think it shines a little bit of a light on what's coming. So, Paul, I wanted to kind of ask you, like, can you unpack for us a little bit the impact that we're potentially going to see in society in the economy from extremely competent AI and, who knows, maybe even AGI after that?
[00:28:50] Paul Roetzer: You and I have talked a lot about this idea that we don't need AGI to see massive disruption and transformation in the economy and the workforce and [00:29:00] business and education and society.
[00:29:02] Paul Roetzer: And, and that's really what we are trying to get.
[00:29:05] Paul Roetzer: People to think more critically about is kind of more of like the near term, like one to two years. Um,
[00:29:12] Paul Roetzer: when this technology keeps getting smarter, and we're going to talk about a few things that have happened in the last few days in rapid fire, where this is occurring again, like the models are taking leaps there, we're starting to see more GPT 4 level and beyond models. coming into our lives.
[00:29:28] Paul Roetzer: And as those emerge, like, what does that mean? What does it mean when it's actually adopted into enterprises? AGI, like I've advised some bigger tech firms and some other companies to have kind of an AGI horizons team within their organization. and
[00:29:42] Paul Roetzer: the premise there is you should be thinking about it.
[00:29:46] Paul Roetzer: Like we should be having these conversations around what happens if we get to AGI, but I wouldn't be dedicating the majority of my resources at that moment to that. I would be trying to figure out. What do we do with the current and near term [00:30:00] AI that we are aware of, that we know is true, and what does that mean to us?
[00:30:04] Paul Roetzer: And there's just not enough companies doing that.
[00:30:07] Paul Roetzer: So, I think within, like
[00:30:10] Paul Roetzer: The context of this klarna news, we have to keep in mind, it is a company release, it's, they didn't like release an open, you know, report about this and drill into all the specifics. So we have to kind of take them at their word that this is, all true, and is a small sample of like one month of global release.
[00:30:27] Paul Roetzer: But it does align with what we're expecting to see as we move forward in 2024. We think we're going to start seeing a lot more examples of this where there's case studies from companies that have actually adopted and infused AI into specific processes or roles that start to see massive impact. And customer service in particular is the one we've been waiting for because In March 2023, in an interview with Lex Fridman, when discussing categories of jobs that could be massively impacted by AI, Sam Altman, who [00:31:00] usually is elusive when it comes
[00:31:02] Paul Roetzer: to talking about these kinds of things and specific jobs, he said, and I quote, I guess I would say customer service is a category that I could see. there are
[00:31:11] Paul Roetzer: just way fewer jobs relatively soon. Now, again, the Klarna News, I mean, they had a, a quote in there from the COO of OpenAI, like they've obviously been working with OpenAI for a while on this project. And so it's not unexpected that. Sam had a foresight to say like, Hey, this is an area where we think there's going to be some pretty significant impact.
[00:31:36] so then related, you and I recently saw a quote
[00:31:40] Paul Roetzer: that had us bidding, doing a bit of a double take, if not a quadruple take. So, um. there's a, there's a website called RAI Journey. dot AI that I would suggest people go check out. So this is from a couple of guys that I recently got
[00:31:58] Paul Roetzer: to know. So Adam [00:32:00] Brotman,
[00:32:00] Paul Roetzer: who's the co founder and co CEO of Forum3. He's the former chief digital officer at Starbucks and also a former president, chief experience officer and co CEO at J. Crew. He and Andy Sack, who is a. Co founder and also co CEO of Form3, founder and managing partner of Keen Capital and a former
[00:32:19] Paul Roetzer: advisor to Microsoft CEO Satya Nadella, as well as building and selling a number of technology companies.
[00:32:26] Paul Roetzer: So these are super legit guys in the tech world, in the business world. and they started this, this really interesting project called Our AI Journey, where they realized
[00:32:36] Paul Roetzer: the impact AI was going to have
[00:32:38] Paul Roetzer: on the world, kind of after the ChatGPT moment, like, like many people in business. And they wanted to start figuring out what does this mean? But instead of just going and researching and writing a book for 12 months, they're doing it.
[00:32:50] Paul Roetzer: in the open. They're basically like doing these fascinating interviews and then they're going to publish it as they go. So this is, it's kind of there's like monthly insights, there's new [00:33:00] chapters
[00:33:00] Paul Roetzer: they're going to release as they do interviews.
[00:33:02] Paul Roetzer: And then there's going to be interactive components and exclusive content. So, it sounds cool, but their first interview happened to be in November with Sam altman, um,
[00:33:14] Paul Roetzer: followed by an interview with Reid Hoffman, which I'll talk about in a minute. So, So after talking with Andy and Adam, I go and read the introduction, which is called the title wave. And so,
[00:33:26] Paul Roetzer: I'll just read this part because this is the thing where Mike and I were just kind of dumbfounded. So they say in the introduction, starts with the story of their interview with Sam Altman in October 2023. So
[00:33:38] Paul Roetzer: it was like one month. roughly before he was fired.
[00:33:41] Paul Roetzer: During the meeting, they asked him, because sam started talking about AGI right away, and they said, when you say AGI, what do you mean?
[00:33:48] Paul Roetzer: And he replied and said, that's a fair question, and I would say it's when AI will be able to achieve novel scientific breakthroughs on its own.
[00:33:56] Paul Roetzer: So now adam and Andy were Pretty, [00:34:00] like, straightforward that they weren't really that familiar with this concept, like, really deep understanding of AGI. so, the chapter goes on and they said, we replied, okay, wow, that's sort of wild.
[00:34:15] Paul Roetzer: Now, not sure exactly what that means, but what do you think aGI will mean for us and for consumer brand marketers trying to create ad campaigns and the like to build their company?
[00:34:25] Paul Roetzer: So again, they're coming at this as marketers and business people. And they're talking to Sam and Sam started AGI and they're trying to figure out like, Whoa, what does this all mean?
[00:34:33] Paul Roetzer: So they ask him specifically, what does it mean for us? To which Sam replied, and I quote, Oh, for that, it will mean that 95 percent of what marketers use agencies, strategists, and creative professionals for today will easily, nearly instantly, and at almost no cost be handled by the AI.
[00:34:54] Paul Roetzer: And the AI will likely be able to test the creative against real or synthetic customer focus [00:35:00] groups for predicting results and optimizing.
[00:35:03] Paul Roetzer: Again, all free, instant, and
[00:35:06] Paul Roetzer: nearly perfect. Images, videos, campaign ideas, no problem. So, this is like on page two of the introduction. So, I'm just like I go back and reread this thing. I'm like, how have I not heard this quote anywhere? Like, why isn't this in like
[00:35:25] Paul Roetzer: all the major media outlets as a quote?
[00:35:27] Paul Roetzer: Because Sam doesn't say this kind of stuff directly. So the, the authors then say, okay, because they had a short time with Sam and he's kind of dropping this on them. And they're trying to process what in the world he's talking about. So they say one last question. And they, they asked him,
[00:35:42] Paul Roetzer: quote, about when do
[00:35:43] Paul Roetzer: you think AGI will be a reality? And Sam said. Five years, give or take,
[00:35:48] Paul Roetzer: maybe slightly longer, but no one knows exactly when or what it will mean for society. And then the rest of the introduction kind of goes on, but I don't know, Mike, what, I had [00:36:00] told you about
[00:36:00] Paul Roetzer: this and I was like, you'll know, I didn't tell you what the quote was. I said, you will know what I'm talking about as soon as you read this.
[00:36:05] Paul Roetzer: When you saw it, what was your first reaction?
[00:36:07] Mike Kaput: Yeah, within four seconds I sent you a screenshot of the quote. this is it, right? Because, yeah, to your point, we follow Sam Altman pretty closely, and I certainly can't quote off the top of my head every statement he's ever made, but this is a degree of candor. And specificity, especially about marketing in particular, which we obviously have a vested interest in, that I have not seen at all before.
[00:36:37] Mike Kaput: And I realize that a lot of the topics that we cover on this podcast, we have to separate the PR, the marketing of some of these companies trying to raise money and justify valuations from reality. But I think you have to take seriously When someone who has as much on the line as Altman does, when he says a specific timeline and is talking [00:37:00] about the types of activities that he has talked about.
[00:37:03] Mike Kaput: Like, clearly, even if he's wrong, he believes this. And it is very, very sobering to read that quote, I would say. And I would say also, I don't know if this is where he's coming from or if it is happenstance that these align. But one of the formative books I know for me and you, Paul, when we first started this whole journey, was Ray Kurzweil who makes tons of predictions about the singularity.
[00:37:30] Mike Kaput: The Singularity is Near is one of his big books. And the predictions sound, frankly, completely insane in terms of how quickly technology will develop and where exponential technology growth will lead us. His first quoted prediction is for AGI in 2029, which is Five years from today. So it's pretty interesting that maybe According to Altman Kurzweil could be right.
[00:37:58] Mike Kaput: But yeah, I would say it [00:38:00] really floored It made me sit up and pay attention. We talk about this stuff 24 hours a day. This one made me say, oh my gosh It was like a ChatGPT moment where it's like, oh wait, this might be here It might be now that all these things we have theorized are coming to pass and I have to start Acting with that level of urgency, perhaps.
[00:38:22] Paul Roetzer: Yeah, and you know when I do talks and you know, our
[00:38:24] Paul Roetzer: intro to AI class or keynotes I always talk about the impact on knowledge work and What I've been saying recently is at least 80 percent of what knowledge workers do will be AI assisted to some degree in the next one to two years, meaning it's just going to be infused within the software we use.
[00:38:39] Paul Roetzer: And I'm, I've been fairly confident in that projection in large part, because if you look at all the tech we're using, it's all like, you're not going to have software that doesn't have AI infused into it. But AI assistance kind of level one, You know, level two ish. It is one thing.
[00:38:55] Paul Roetzer: Going back to this quote,
[00:38:56] Paul Roetzer: it will mean that 95 percent of what marketers use [00:39:00] agencies, strategists, and creative professionals for today will easily, nearly, instantly, and at almost no cost be handled by the AI. Um,
[00:39:12] Paul Roetzer: It's a hard thing to wrap our minds around.
[00:39:15] Paul Roetzer: and honestly, like, I haven't really had time to sit back and
[00:39:20] Paul Roetzer: think about this at a more, you know, a deeper level and try
[00:39:23] Paul Roetzer: and project out almost like the AGI Horizons thing I was talking about earlier, like, What does that mean? So I think it's probably just, we'll probably like leave it
[00:39:32] Paul Roetzer: here and, and give people time to maybe process this themselves a little bit.
[00:39:37] Paul Roetzer: Um, definitely something we're going to come back to on the show. And again, the Our AI Journey, site OurAIJourney.ai
[00:39:45] Paul Roetzer: it is a paid program, but again, just judging by the first interview with Sam and then this chapter one is with Reid Hoffman from December of 2023. Well, Reid is the COO of PayPal, which is where his connection to Musk comes in,
[00:39:58] Paul Roetzer: because Musk was one of the co founders [00:40:00] of PayPal. Um,
[00:40:01] Paul Roetzer: he's the co founder of linkedIn, which he sold to Microsoft. He's a
[00:40:03] Paul Roetzer: Microsoft board member and he's the co founder of Inflection with Mustafa Solomon. One of the co founders of DeepMind and
[00:40:10] Paul Roetzer: they raised 1. 3 billion last year to build
[00:40:13] Paul Roetzer: language models. So, Reed is as tapped in as probably anyone in the world with the current and future state of AI.
[00:40:20] Paul Roetzer: So just to know that Andy and Adam have access to these people and, and they're going to be putting out these kinds of interviews. We have no stake in the game here. It's not like, you know, I benefit from you going and getting their stuff, but like, I'm hooked.
[00:40:34] Paul Roetzer: Like that was my reply to the
[00:40:35] Paul Roetzer: introduction. I was like, all right, you got me. I'm in like. Let me know when the next chapters
[00:40:39] Paul Roetzer: are coming out. So, um,
[00:40:41] Paul Roetzer: just something to keep an eye on. And I get this as kind of like a really big topic and a hard thing to process. And so we'll do our best in the coming months to sort of try and unpack this a little bit.
[00:40:54] Mike Kaput: So, in a related kind of final main topic here, we wanted to quick, quickly [00:41:00] talk about AI's potential impact on creators, because it sounds like creators are starting to get pretty worried about AI, according to Some really interesting newer reporting from the information. So they actually did this deep dive into kind of how generative AI is impacting creators and media distribution platforms like YouTube, Netflix, Spotify, etc.
[00:41:25] Mike Kaput: And it sounds like creators are really starting to have some deep fears about AI generated content and they kind of fall into three large buckets. And the first is that They're worried AI generated content is going to start flooding all these distribution platforms where their audience consumes their content.
[00:41:45] Mike Kaput: So, for instance, Spotify alone has seen an influx of AI generated music. The second concern is a little more insidious, actually. There were creators that were interviewed by the information. that are worried that [00:42:00] audiences will develop an appetite for AI generated content. As anyone who's heard some of these AI deepfaked songs from artists, you can do pretty convincing and interesting artistic renditions of real people and artists.
[00:42:16] Mike Kaput: The third is that people are starting to worry the benefits of using generative AI are going to essentially force everyone to produce AI generated content in order to compete. So one example of this last point is Netflix. The company actually told the SEC in their annual report, quote, If our competitors gain an advantage by using, in this case, generative AI, such technologies, our ability to compete effectively and our results of operations could be adversely impacted.
[00:42:48] Mike Kaput: So, in other words, if other people start using AI to generate content, we have to as well. Now, one example of this that's mentioned in the information [00:43:00] reporting is a YouTube creator with the name Quebelkop. He is a gaming megastar. He has like 15 million plus subscribers, and for years he has experimented with Generating videos with AI to some degree, and even at one point created this full AI persona to star in his videos, So he shut that experiment down.
[00:43:23] Mike Kaput: And what he told the information I found really, really interesting in terms of tangible impact for creators. He said that AI now allows him to work with half the team and half the budget that he did just two years ago. And he said AI's impact will be profoundly disruptive. to creators saying if you don't embrace AIt will replace you.
[00:43:48] Mike Kaput: Now, with all that kind of skeptical, more cynical perspective, others aren't as worried, and that includes YouTube. YouTube [00:44:00] CEO Neil Moen. released a letter last month to creators and he said AI should empower human creativity not replace it. Everyone should have access to AI tools that push the boundaries of creative expression.
[00:44:13] Mike Kaput: One creator interviewed by the information agreed saying look it's in our DNA. to tell stories and hear stories being told. So there's this kind of robust debate and worry and consideration around how AI is actually going to impact this space if you're a creator, if you do any type of content creation for an audience online.
[00:44:34] Mike Kaput: So, Paul, I just wanted to kind of open it up and get your initial thoughts on that because I know, you know, we've got a number of different events and talks coming up that are related in one way or another. to creators. So what did you kind of think of these concerns and the overall What impact AI might be having on this space?
[00:44:53] Paul Roetzer: It was really timely for me, as you mentioned, with the Writer's Summit coming up, and actually for the state of AI and writing, the [00:45:00] theme I ended
[00:45:00] Paul Roetzer: up landing on, and I'm still building the deck, which, sorry to my team that's probably waiting for me to send them the deck, is why we
[00:45:09] Paul Roetzer: write. And so, like, I was trying to, like, struggle with this, and I think, you know, if anybody's listened to this show for a long time, they know my wife is an artist, she's
[00:45:18] Paul Roetzer: a painting major, my daughter at 12 wants to be an artist, my son wants to be a video game developer, So I
[00:45:23] Paul Roetzer: spent a lot of time thinking about, like, the impact on human creativity.
[00:45:26] Paul Roetzer: I'm a writer by trade, you're a writer by trade, Mike. And so, as I sat down to try and build the presentation for the Writer's summit,
[00:45:34] Paul Roetzer: It was this question I kept coming back to, because I'll often say when I'm giving presentations that I don't use AI to write, because to me writing is thinking, and
[00:45:45] Paul Roetzer: so I'll use it as like an assistant for ideation and brainstorming and stuff like that, but when I actually want to write something, like, I want to go through the process of writing.
[00:45:53] Paul Roetzer: And so I started in, like,
[00:45:57] Paul Roetzer: In developing the deck. Asking myself, like, well, [00:46:00] why do I write? Like, what is it about writing that I feel is so critical? And, you know, I started breaking it down, and I don't, you know, not to, like, give
[00:46:08] Paul Roetzer: the whole talk for people that are going to attend on Wednesday, but I started thinking about it in two categories.
[00:46:13] Paul Roetzer: Like, one is, why do we write, or in this world, let's
[00:46:16] Paul Roetzer: just say create, for our audiences, and why do we create for ourselves? And I think it's an important distinction, because if we're talking about, like, creating videos, or video games, or articles, or artwork, or like, whatever it is, There is an end product
[00:46:33] Paul Roetzer: that's created for someone else, for them to gain something from it, to, you know, enjoyment, entertainment, whatever it may be.
[00:46:41] Paul Roetzer: and so you're trying to communicate, you're trying to inspire, you're trying to educate, like you're creating something for them.
[00:46:48] Paul Roetzer: And there's no reason AI can't do that. It's capable of that. And you could use AI purely to do that. But in the process, you then lose the why do you create for yourself part of the [00:47:00] equation.
[00:47:00] Paul Roetzer: And for you There's other reasons why you go through the process. So for me, like, I write to express myself.
[00:47:07] Paul Roetzer: I write because it helps me remember what I'm going to talk about. It helps me organize my thoughts, reflect on my experience. Like,
[00:47:13] Paul Roetzer: writing is a craft to me. It's, it's like, it's not about just the end product. And so, yes, like, people can use AI to create stuff on youTube and probably we're going to be in a phase,
[00:47:24] Paul Roetzer: if you look at what Sora, I mean Go to Instagram and go to the OpenAI account on Instagram. They're sharing these Sora videos all the time with the prompts. It's mind blowing what this technology is going
[00:47:34] Paul Roetzer: to be capable of doing.
[00:47:35] Paul Roetzer: And you can certainly project out and see where we're going to have this ability to create anything. And maybe people will like that. Maybe they'll like to be able to personalize
[00:47:46] Paul Roetzer: stories and videos and shows like with prompts, because that's just the process of being entertained.
[00:47:52] Paul Roetzer: And so again, it's for them, but like the artist. Why do they create? It's not just to create the end product. There, there's [00:48:00] a, there's
[00:48:00] Paul Roetzer: a fulfillment that comes from it that as a writer, I don't want the AI to take away. Like I'm going to keep writing because it is the art of writing that I love. And that helps me express myself
[00:48:12] Paul Roetzer: and like, get ideas into the world and formalize my thoughts.
[00:48:14] Paul Roetzer: Like if I took For this podcast that we've talked about before, like I don't script what I'm going to say. I'll put some notes down, but like this isn't scripted at all. if I just said, okay, Mike's going to ask me these
[00:48:26] Paul Roetzer: questions, ChatGPT give me answers. And like the whole show was ChatGPT giving answers.
[00:48:31] Paul Roetzer: And that's all I was doing was regurgitating that.
[00:48:34] Paul Roetzer: I would have no retention of that information. I would not be able to answer questions. on
[00:48:39] Paul Roetzer: the fly when I'm standing in front of a thousand people on stage. The way I'm able to do this and have these kind of generalized conversations, because I put the work in to learn the topic and go through the process of writing and thinking.
[00:48:52] Paul Roetzer: And I don't want the AI to replace that. And so I think it's just something we need to think about moving forward. And again, I'll get into this more in my talk on Wednesday of like, [00:49:00] there's two things to consider when we talk about AI creating
[00:49:02] Paul Roetzer: things. It can create outputs that people may find enjoyable.
[00:49:06] Paul Roetzer: But that doesn't replace the artist or the creator themselves needing to find enjoyment out of the process of creating things.
[00:49:13] Paul Roetzer: That's a fantastic point. I cannot wait to hear that I can't
[00:49:18] Paul Roetzer: wait to get the deck done.
[00:49:20] Mike Kaput: Well, it's conversations like these that make it really special and unique. So.
[00:49:25] Paul Roetzer: I may just have you just interview me instead.
[00:49:28] Paul Roetzer: We'll call an audible and like, no deck. Mike's just gonna, we're just going to have a conversation around creativity.
[00:49:33] Mike Kaput: There you go, special live podcast episode.
[00:49:36] Mike Kaput: Yeah,
[00:49:37] Mike Kaput: Awesome. Well, we have a ton of rapid fire to get into, so I want to get into these quick here, because each of these Honestly, could probably be a big topic on the show.
[00:49:49] Paul Roetzer: Seriously.
[00:49:51] Mike Kaput: there is an AI robotics company that you probably need to be paying attention to if you have not already.
[00:49:57] Mike Kaput: It's called Figure, and they're developing [00:50:00] general purpose humanoid robots. And they just announced that they're raising 675 million dollars in Series B funding. This investment values the company at 2. 6 billion, and this round includes investors like Microsoft, OpenAI startup fund, NVIDIA, and Jeff Bezos. Along with the investment, Interestingly, FIGURE also announced that it's entered into a deal with OpenAI to develop AI models for humanoid robots.
[00:50:31] Mike Kaput: So Paul, this is definitely a sci fi episode. Are we now getting humanoid robots powered by OpenAI? What is the significance of this raise and OpenAI's involvement here?
[00:50:43] Paul Roetzer: A couple, a couple of notes.
[00:50:46] Paul Roetzer: One, you know, we've talked about the humanoid robot thing is starting. It again, hasn't had its ChatGPT moment, but it certainly appears as though we're trending in a direction where these things are going to start becoming quite viable in the next couple of
[00:50:59] Paul Roetzer: years. It seems like there's [00:51:00] a lot of scientific breakthroughs
[00:51:01] Paul Roetzer: that are happening in the development of these robots. I think the significance related to what we talked about earlier is, Elon musk is building
[00:51:10] Paul Roetzer: Optimus, which is a human eyed robot. Elon musk obviously competes with OpenAI. OpenAI is now involved in this, as is Bezos, who, who Musk notoriously competes with in space and rockets.
[00:51:23] Paul Roetzer: again, if you read the Elon Musk biography, you, you, you know full well the story of Elon musk and Jeff Bezos with space exploration. but I, the thing I thought was funny is.
[00:51:33] Paul Roetzer: So Brett adcock, the CEO and I think founder of FIGURE, tweeted this on February 29th, excited to share FIGURE raises 675 million plus openAI and FIGURE signed collaboration.
[00:51:46] Paul Roetzer: Elon Musk replies, bring it on. So i, you know, I think it's just more of the intrigue. It's all kind of connected, but
[00:51:55] Paul Roetzer: robots are going to be a thing. You're going to start seeing, hearing a lot more about [00:52:00] humanoid robots. A lot of people are playing in that space. interestingly, OpenAI tried to do robots in their early days, and then they kind of like backed off of it, the hardware side, so they've been in the robot space, but it seems
[00:52:12] Paul Roetzer: like they're now basically going to provide the software slash intelligence, within the embodied.
[00:52:18] Paul Roetzer: AI of the figure robots. So keep, keep an eye on the space. It's going to be intriguing.
[00:52:25] Mike Kaput: It's not a boring time to be alive,
[00:52:28] Paul Roetzer: It's really not.
[00:52:30]
[00:52:30] Mike Kaput: All right, so next up we got news that Apple has decided to abandon its nearly decade long effort known as Project Titan to build an autonomous electric vehicle after they've struggled to make significant progress. Now, this announcement was made internally this past Tuesday and It was made to the, team of around 2, 000 employees that are on Project Titan.
[00:52:58] Mike Kaput: Now, Apple started [00:53:00] working on an electric self driving car way back in 2014, but they have repeatedly changed strategies and leaders of this effort over the years. The decision to ultimately wind this down finally was finalized by Apple's top executives in recent weeks. And it sounds like many employees from the disbanded car project will be assigned to generative AI initiatives under John Giandria, Apple's head of AI.
[00:53:28] Mike Kaput: Um, there is speculation, though, there will be some layoffs related to this, particularly, obviously, among people who might be much more specific to the auto industry. So, Paul, for some context here, this has obviously been something that's been in the works for quite some time. What is the significance of Apple finally kind of killing this project?
[00:53:52] Paul Roetzer: The, I mean, the obvious thing is what you talked about is the shift to generative AI. I think they're all in on AI.[00:54:00]
[00:54:00] Paul Roetzer: They're all in probably on building a more intelligent Siri, embodying that intelligence through glasses, rings, like whatever the form factors are going to be, but it does not appear as though it will be. In a car to compete with Tesla.
[00:54:12] Paul Roetzer: I just thought it was so interesting cause this project has been so heavily rumored for so long. I mean, and now, you know, 10 years, 10 billion, it was not an insignificant effort, but they managed to keep it
[00:54:25] Paul Roetzer: very quiet, a lot of the details. And now it's just kind of coming out that. It sounds like they made, at some point, a decision that rather than trying to build, like,
[00:54:33] Paul Roetzer: a Tesla, like electric car, they were just gonna go straight to level 5 autonomy with no steering wheel.
[00:54:39] Paul Roetzer: And it was just, too much. Like, it just became too challenging, or, you know, they just
[00:54:45] Paul Roetzer: didn't make the breakthroughs they thought they would make. So I think it's intriguing and it's Apple. So anything Apple does, especially over a 10 year period is fascinating to now get an inside look at the story of what
[00:54:56] Paul Roetzer: was really going on all these years when all we had was rumors.
[00:54:59] Paul Roetzer: [00:55:00] And, I think this move to a major focus on Gen AI is a huge deal and we're going to start to see that coming through in June of. 2024 probably at their developers conference, their worldwide developers conference. I would expect some major announcements probably then.
[00:55:16] Mike Kaput: So
[00:55:17] Mike Kaput: Salesforce has also made their own announcement that they're announcing Einstein Copilot, which is a new conversational AI assistant that integrates generative AI capabilities right into Salesforce's. Salesforce says that, quote, unlike other AI assistants or copilots that lack adequate company data to generate useful responses, Einstein Copilot enables Salesforce customers to generate responses using their own private and trusted data, while maintaining strict data governance and without requiring expensive AI model training.
[00:55:54] Mike Kaput: As a result, Einstein Copilot can answer questions, summarize content, create new content, interpret [00:56:00] complex conversations, and dynamically automate tasks on behalf of a user, all from a single consistent user experience embedded directly into Salesforce. This is being offered to Salesforce customers. with availability starting in a public beta, initially for sales cloud and service cloud users, and Salesforce has said it will roll this out later to commerce cloud, marketing cloud, and Tableau analytics users.
[00:56:32] Mike Kaput: Paul, we've talked a lot about AI copilots. in marketing and sales. Obviously, Salesforce is a huge player in this space. I'm assuming we will see AI co pilots in basically every single marketing and sales tool we all use to do our jobs. Big question I have for you is Kind of where's HubSpot in all of this, being a main competitor of Salesforce?
[00:56:58] Mike Kaput: They definitely have a ton of AI [00:57:00] features that they're including. They have ChatSpot, which is layering over your data. That's still very, very early. Sounds like Salesforce Copilot is kind of what we might be expecting moving forward in software platforms.
[00:57:13] Paul Roetzer: Yeah, I honestly not sure where HubSpot is in all of this. So I mean, I think people generally are aware we were, my agency that I sold in 2020 was HubSpot's first partner. So we worked pretty closely with HubSpot for a long time. And, you know,
[00:57:28] Paul Roetzer: I, at the time had a lot of insight into kind of just the product and the platform. We use HubSpot at the Institute.
[00:57:33] Paul Roetzer: It is the core platform we use for marketing sales service ops. they made a bunch of announcements in the fall about things they were working on. I I haven't heard anything lately on chat spott, so I'm not sure what they're doing there. Um,
[00:57:50] Paul Roetzer: So, yeah, I mean, I would think we'll probably hear more. I mean, I know HubSpot's been talking more about ai, building in a lot of capabilities, but I.
[00:57:59] Paul Roetzer: I don't [00:58:00] know that we're using any of that. Like you use HubSpot more than I do. Do you actively use any AI features built within the platform?
[00:58:07] Paul Roetzer: I
[00:58:08] Mike Kaput: return to ChatSpot relatively regularly to see, because in my mind, ChatSpot, everything I just read about Salesforce's co pilot is what I'm hoping Chatsbot eventually can be in a unified experience. I don't think it's there yet, though I'm due for a revisit soon. the generative AI features, just because of all the other tools we use, I would say I personally don't lean on that much, but I can see how they'd be super valuable if you don't use all these other tools. But other than that, that's kind of why I asked the question. It's like, I haven't seen something yet that's quite this comprehensive, I would say, of an announcement across
[00:58:44] Paul Roetzer: a lot of different data.
[00:58:45] Paul Roetzer: Yeah,
[00:58:46] Paul Roetzer: and I think that's, like, what we hoped
[00:58:48] Paul Roetzer: Chatspot would be when it was first announced by Dharmesh and the team is This like true intelligent assistant into anything happening within
[00:58:58] Paul Roetzer: the platform. It's what I've
[00:58:59] Paul Roetzer: always [00:59:00] dreamed of is like, you know, if I think of something like, Oh, you know, just a quick example, for the writer's summit, we had 4, 200 people registered last year. So
[00:59:10] Paul Roetzer: when I was building my presentation, I was curious, well, how many people who are registered this year, which is I think 4, 000 or above.
[00:59:17] Paul Roetzer: attended last year because I was trying to get a sense for like, who saw my talk last year and kind of how different and everything like that.
[00:59:24] Paul Roetzer: So to do that, I have to message someone on the team and say, can you go in and build a list in HubSpot that includes the rules, attended 2023 Writer Summit, which needs to have a list tied to it.
[00:59:39] Paul Roetzer: And is registered for 2024. So someone on the team now actually has to go in, make sure we have the list from last year, which I already knew we did, then build a new list, write those two rules,
[00:59:50] Paul Roetzer: and then let the list for like 40 seconds, create itself, then come back to me and say, okay, there's 225 people who are [01:00:00] registered this year who attended last year.
[01:00:02] Paul Roetzer: And then I may have a follow up question. Well, how many people who registered last year, but didn't show up?
[01:00:07] Paul Roetzer: are registered. Like, I now have a bunch of questions, which is me Zoom in Zoom back to someone on the team saying, can you go do this and this and this? That should not happen.
[01:00:16] Paul Roetzer: Like that is, it should be a goal of HubSpots to make me never have to go through that workflow again, because the data is in there.
[01:00:23] Paul Roetzer: All I should have to do is just talk to a chat assistant and ask those exact questions and in real time get the answers
[01:00:31] Paul Roetzer: to it. So that is the value if you're not like connecting the dots here. That's the value of your CRM being connected to your data is that I can just have a conversation at any moment because I might think of this at 11 o'clock tonight when I'm finalizing my deck and I might want this information.
[01:00:47] Paul Roetzer: Well, I'm not going to message Mike at 11 o'clock or somebody else on the team. And so I mean, every major CRM needs
[01:00:55] Paul Roetzer: to be moving in this direction with a great sense of urgency because The [01:01:00] inefficiency of working within CRMs in
[01:01:02] Paul Roetzer: a traditional way, like I just explained, is going to feel really, really obsolete very soon.
[01:01:09] Mike Kaput: that's what I was wondering. Yeah, a, not a CRM example, but what you're talking about, for instance, in Google Analytics 4. just type in how many users did we get last month and it just tells you. I don't have to go click into anything. It just provides the answer using generative AI. So that's what we're kind of getting at here that should be the default experience moving forward with these data sets.
[01:01:36] Paul Roetzer: And multi modal too, like create a visual for me, or what does this visual mean? Or what's happening here? Like take a, you know, clip of a conversation happening
[01:01:44] Paul Roetzer: with a customer, drop it in, what's going on here? And yeah, I mean, it's just a lot becomes possible.
[01:01:50] Paul Roetzer: And I think. These, these CRM platforms are going to have to move really quickly to catch up.
[01:01:56] Mike Kaput: In other
[01:01:57] Mike Kaput: news, Mistral AI, which is [01:02:00] a Paris based AI startup that was founded just about nine months ago by former AI researchers from Meta, Google DeepMind, and some other big names, they have just signed a major deal with Microsoft. As part of this deal, Mistral's flagship model, which is called Mistral Large, is now available on Microsoft Azure, which makes it the only, only the second AI provider after OpenAI to offer a commercial large language model on Azure.
[01:02:30] Mike Kaput: Mistral has also partnered with IBM to provide an optimized version of its open source model on IBM's Watson AI platform. Now, we mention this because this is kind of a breakout moment for Mistral, because like I mentioned, it's less than a year old, and it's got 415 million plus in funding, and some benchmarks are rating its open source models as very, very high on performance when stacked up against OpenAI, Google, Meta, and others.
[01:02:59] Mike Kaput: So, Paul, can [01:03:00] you give us some context into why Mistral matters to the overall AI ecosystem?
[01:03:06] Paul Roetzer: On the surface, the main deal here is that, like you alluded to, that they're
[01:03:10] Paul Roetzer: a major player in the open source world and have one of the most powerful models overall in the world, and specifically
[01:03:19] Paul Roetzer: open source model, where they're largely competing with META, would be the other main player with Llama.
[01:03:24] Paul Roetzer: the other interesting aspect to this news is Microsoft doing a deal with an open source company and I saw the CEO of Mistral actually had to like Go out and say, listen, Hey, we, we, we're still open source,
[01:03:36] Paul Roetzer: even though we're doing a deal with Microsoft and they're generally closed with like OpenAI and stuff, it like, we're still at the spirit of what we're trying to do.
[01:03:44] Paul Roetzer: And here's why we're doing it. and then I think
[01:03:46] Paul Roetzer: from a Microsoft perspective, it's just fascinating given how deep in they are with OpenAI, that they're just so like directly now diversifying their models and their capabilities. And I think it just is
[01:03:59] Paul Roetzer: kind of this [01:04:00] bigger picture moving forward of open versus closed and, you know, how this is
[01:04:03] Paul Roetzer: all going to play out.
[01:04:03] Paul Roetzer: And it just seems like we're going to kind of move forward with, with both models, and, and both options, I guess.
[01:04:10] Mike Kaput: All right. so this next topic is literally hot off the press because we found it minutes before going live today. It turns out that Anthropic, who we have followed quite a bit, is releasing the next generation of its CLAUDE model, and I'm just going to read from part of the article announcing this. And it says, today we're announcing the CLAUDE3 Model family, which sets new industry benchmarks across a wide range of cognitive tasks. The family includes three state of the art models in ascending order of capability. CLAUDE3 Haiku, CLAUDE3 Sonnet, and CLAUDE3 OPUS. Each successive model offers increasingly powerful performance, allowing users to select the optimal balance of intelligence, speed, and [01:05:00] cost for their specific application.
[01:05:03] Mike Kaput: OPUS and SONNET are now available to use in CLAUDE. ai, and the CLAUDE API, which is now available in 159 countries. Haiku will be available soon. So, they provide a visual here. CLAUDE3 Haiku is the lowest cost, but also lowest level of intelligence. Followed by CLAUDE3 Sonnet, which increases in both cost and intelligence.
[01:05:27] Mike Kaput: And the WD3 OPUS, which is the most powerful. And they say that OPUS, our most intelligent model, outperforms its peers on most of the common evaluation benchmarks for AI systems. That includes undergraduate level expert knowledge, graduate level expert reasoning, basic mathematics, and more. Quote, It exhibits near human levels of comprehension and fluency on complex tasks.
[01:05:52] Mike Kaput: Leading the frontier of general intelligence. So, Paul, this is a pretty big deal based on some of the claims [01:06:00] that Anthropic is making here. What were your initial
[01:06:02] Mike Kaput: thoughts? Ha
[01:06:05] Paul Roetzer: February 23rd, this is like 6, 10
[01:06:09] Paul Roetzer: days ago, I tweeted, Seems like Anthropic, coherent, and flexible. Flexion have been quiet lately amongst all the generative AI news and announcements.
[01:06:16] Paul Roetzer: Only a matter of time, I assume, so that that aged well. . So here we go with Claude three. again, as you mentioned, literally, like seven minutes before Mike and I were about
[01:06:28] Paul Roetzer: to start recording, the tweets started going out about this. So very, very early. I have not been looking at Twitter as we've been talking here.
[01:06:36] Paul Roetzer: The API is open. We saw Matt Schumer was testing it already. Matt Schumer of Hyperwrite, we've talked on the show quite a bit.
[01:06:43] Paul Roetzer: he was already running tests. I saw, Ethan Mollick tweeted that like, it definitely seemed like it was at least on par with gPT 4, if
[01:06:50] Paul Roetzer: not higher in a number of areas I saw in the,
[01:06:54] Paul Roetzer: anthropic. Thread that the context window is 200, 000 tokens, but that they [01:07:00] showed the capability of easily going up to a million.
[01:07:02] Paul Roetzer: So it's getting on par with like a Gemini 1. 5.
[01:07:07] Paul Roetzer: I don't know. Like, we'll have to process this and come back maybe next week and fill in some blanks on this, but yeah, like hot off the presses. Thank you to Anthropic for putting it out before we recorded this podcast episode. So we didn't get all the. the notifications from people asking why we didn't talk
[01:07:24] Paul Roetzer: about it. So there you go. And then like, I just wanted to like end with one other thing that was like, literally, as we were about
[01:07:33] Paul Roetzer: to come on, I saw, so Ron Conway, who is Ron is with, uh, renowned venture capitalist. he tweeted this morning. We all share responsibility for building AI that improves lives and unlocks a better future for humanity.
[01:07:53] Paul Roetzer: As at SV Angel, his fund makes this pledge, and we are proud to initiate
[01:07:57] Paul Roetzer: this open letter, build AI [01:08:00] for a better future. Please join OpenAI, Meta, Google, Y Salesforce, Microsoft, and others in signing it today. So I thought, with everything we talked about today and how crazy all this is, I will read to you, the very short Build AI for a Better Future letter, which you're welcome to go add your signature to.
[01:08:20] Paul Roetzer: I did add mine this morning. so this is what it says. We call on everyone to build, broadly deploy, and use AI to improve people's lives and unlock a better future.
[01:08:33] Paul Roetzer: The purpose of AI is for humans to thrive much more than we could before. AI is still early, but it's on its way to improving everyone's daily life.
[01:08:43] Paul Roetzer: AI tutors to help anyone learn. AI translation tools to better connect the world. AI guided medical diagnoses to improve healthcare. AI powered research to accelerate scientific discovery, AI assistance that you can talk with to help with everyday tasks. While AI [01:09:00] is unique in directly augmenting human
[01:09:02] Paul Roetzer: thought, we expect its impact to be more akin to the printing press, the combustion engine, electricity, and the internet.
[01:09:09] Paul Roetzer: The balance of its good and bad impacts on humans will be shaped through the actions and thoughtfulness we as humans exercise. It is our collective responsibility to make choices that maximize AI's benefits and mitigate
[01:09:21] Paul Roetzer: the risks for today and for future generations.
[01:09:25] Paul Roetzer: We all have something to contribute to shaping AI's future, from
[01:09:29] Paul Roetzer: those using it to create and learn, to those developing new products and services on top of the technology, to those using AI to pursue new solutions to some of humanity's biggest challenges, to those sharing their hopes and concerns for the impact of AI on their lives.
[01:09:44] Paul Roetzer: AI is for all of us, and all of us have a role to play in building AI to improve people's lives. We, the undersigned, already are experiencing the benefits from AI and are committed to building AI
[01:09:56] Paul Roetzer: that will contribute to a better future for humanity. Please [01:10:00] join us. And then as I mentioned,
[01:10:01] Paul Roetzer: Meta, Google, OpenAI, Salesforce, Microsoft, Hugging Face, Character. ai, Eleven Labs, Repl. it, like, everybody is signing this.
[01:10:10] Paul Roetzer: And, for me, you know, versus the Techno, Optimist, Manifesto, or whatever that EAC thing was that we talked about from Andreessen Horowitz a few months back.
[01:10:22] Paul Roetzer: I can get on board with this. Like, you know, it's just a letter. It's just kind of a statement of belief and what we want. And I think, you know, I
[01:10:30] Paul Roetzer: don't want to speak for you, Mike, but I feel like this aligns with what we're trying to do with this show.
[01:10:34] Paul Roetzer: Like drive literacy, help people understand it for the betterment of humankind, um, that makes us specious. I'll jump in that club all
[01:10:43] Paul Roetzer: day long. but yeah, I just thought that would be a good, good way to wrap today with some, some new stuff that's happened in the tech world.
[01:10:52] Mike Kaput: I love it. And I just want to mention as we wrap up here, again, AIwritersummit. com is [01:11:00] the event that is happening this week and that there's no better event in my mind right now. for talking
[01:11:09] Mike Kaput: about exactly what you just spoke about in that lead which is figuring out how to preserve and enhance the human while actually harnessing the technology that is in front of us.
[01:11:20] Mike Kaput: I think that's such an important theme of writers, the AI for Writers Summit, so definitely check that out.
[01:11:26] Paul Roetzer: Man, that was a lot.
[01:11:28] Mike Kaput: Ha ha ha ha!
[01:11:30] Paul Roetzer: I gotta go build this deck. I have nothing left in my, like, no brain power left.
[01:11:34] Mike Kaput: Yeah. All right, Paul, well, thank you so much for unpacking the world of AI this week. I feel like we fit a year into a week, but hey, we're
[01:11:45] Paul Roetzer: All right. Thanks everyone. And thank you, Mike. And hopefully we'll see some of you all on the AI for Writers Summit on Wednesday and we'll be back next week.
[01:11:53] Paul Roetzer: Thanks again.
[01:11:54] Thanks for listening to The AI Show. Visit MarketingAIInstitute. com to [01:12:00] continue your AI learning journey. And join more than 60, 000 professionals and business leaders who have subscribed to the weekly newsletter, downloaded the AI blueprints, attended virtual and in person events, taken our online AI courses, and engaged in the Slack community.
[01:12:17] Until next time, stay curious and explore AI.