The AI Daily Brief: Artificial Intelligence News and Analysis - Meet the AI Avatars Running for Elected Office

Episode Date: June 14, 2024

In Wyoming, an AI avatar is running for mayor. In the UK, one is running for Parliament. Are impartial, fact-based AI politicians the wave of the future? Also, OpenAI is now on a $3.4B annualized revenue run rate. ** Join Superintelligent at https://besuper.ai/ -- Practical, useful, hands on AI education through tutorials and step-by-step how-tos. Use code podcast for 50% off your first month! ** ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://aidailybrief.beehiiv.com/ Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@AIDailyBrief Join the community: bit.ly/aibreakdown

Today on the AI Daily Brief, OpenAI is now making more than $3 billion on an annualized basis. Before that in the headlines, meet some AI candidates running for public office. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show notes. Welcome back to the AI Daily Brief Headlines edition, all the daily AI news you need in around five minutes. If you had any hope that the world was going to get less weird rather than weirder, I hate to be the bearer of bad news, but we have not one, but two stories about AI candidates running for political office.

The first is a mayoral candidate in Wyoming, an AI persona named Vic with an upbeat masculine identity and the ability to read documents at Lightning Speed who is running for the mayor of Cheyenne. Dick was generated by the son of a local newspaper columnist and argues that the candidate will be better at listening to its constituents. Said creator Victor Miller, I'm unhappy with how public servants treat the public for one way or the other. How to fix that? I don't know. I'm going to try AI. Victor points out that elected officials make decisions for a whole variety of reasons, many of which have nothing to do with what their

constituents want. The opportunity for Vic, which stands for virtual integrated citizen, is to take in reams and reams of information, supporting documents, different arguments, and make decisions more objectively based on that. In terms of interaction, creator Victor Miller wears a speaker around his neck, and voters can ask Vic campaign questions, who then responds in a quote, cowboy-like masculine tone and persona. County officials are currently investigating whether Vic can legally appear on the ballot, but it seems like they're actually trying. Meanwhile, this is not the only AI-powered candidate running for office right now. Over in the UK, an AI candidate named AI Steve is running for Parliament. AI Steve is an avatar of Stephen Endicott, based in Brighton and running for

Parliament as an independent. Endicott will function as an in-person representative of AI Steve, attending meetings in parliamentary sessions on behalf of the AI avatar, but it will not be Endicott making decisions. writes Wired, AI Steve will transcribe and analyze conversations it has with voters and put issues of policy forward to validators or regular people who can indicate whether they care about an issue or want to see a certain policy enacted. For example, the AI Steve team says that they want to reach out to commuters at the Brighton train stop, asking them to fill out short policy surveys. Wired writes while Endicott says that he expects his own opinions or policy preferences may differ from those of AI Steve at some point, he says he is committed to voting in line with the

constituent preferences as expressed through AI Steve. Says Indicott, surely in a democracy, it's what your constituents want. I know that it sounds so obvious that a politician should be told what to do by his constituents, and if he doesn't like it, tough luck. Get out of the job. Lest you think that this is some crazy out their opinion, Joe Rogan talked about this on a show earlier this year as well, saying that he would actually be probably in favor of an AI president for these exact reasons, that if something really could take in all the information about a different situation and make an informed objective decision based on that, he'd be for it. Point being, these things may seem like novelties now,

but they may get a lot more normal, a lot faster than you would think. Moving back to real-world leaders, Pope Francis is about to become the first Pope to attend the G7 summit, and you better believe his topic is AI. The Pope you might remember got unintentionally involved with AI last year when an image of Pope Francis wearing a white puffer jacket went viral. CNN business writes, however, that the 87-year-old Pontiff is determined to use the soft power of his office at the G7 to try to ensure that the development of AI serves humanity and does not turn into a 21st century Frankenstein's monster. The Pope has previously highlighted threats posed by AI-controlled weapons systems, the dangers of a surveillance society, problems of interference and elections, and general worries about inequality.

Write CNN, the Pope and the Vatican have been pushing for an ethical framework to underpin the development and use of AI. The Vatican is seeking buy-in from big tech companies and governments. It's not exactly clear ultimately what specific steps, if any, the Pope will call for, but the fact that he has decided to get involved at the G7 level certainly suggests how important this issue is to the Vatican. One more story that might turn into a longer conversation at some point on this show is a discussion happening on Twitter slash X, as well as in the pages of fortune,

around whether Amazon dropped the ball on using Alexa to be the leader of generative AI. reporter Sharon Goldman wrote a thread summing up her piece. She writes, I spent the last few weeks speaking to over a dozen former employees at Amazon's Alexa organization, who say the company dropped the ball on releasing a generative AI-powered Alexa nine months after a splashy demo. At a packed event at Amazon's lavish second headquarters in the Washington, D.C. suburbs, Alexa was demonstrated to a room full of reporters and cheering employees. But after the event, there was radio silence.

The reason, according to interviews, is an organization beset by structural dysfunction and technological challenges that have repeatedly delayed shipment of the new generative AI-powered Alexa. Overall, the former employees paint a picture of a company desperately behind its big tech rivals, Google, Microsoft, and meta, in the race to launch AI chatbots and agents and floundering in its efforts to catch up. The September 2023 demo, the former employees emphasized was just that, a demo. The new Alexa is still not ready for primetime rollout. Research scientists who worked on the LLM said Amazon does not have enough data or access to the specialized computer chips needed to run LLMs to compete with rival efforts at companies like OpenAI.

Amazon has also former employees, say, repeatedly deprioritized the new Alexa in favor of building generative AI for AWS. And while Amazon is invested $4 billion in Anthropic, it has been unable to capitalize on that relationship to build a better Alexa. Mihail Eric also wrote a long thread about this that got much deeper into the details as someone who joined Alexa AI as a research scientist back in 2019. It's a fascinating conversation about one of the big battles. But of course, the question is, With something like 500 million Alexa devices sold at this point, can Amazon still use that incredible starting point to get itself back in the race? Something to continue to explore, but for now,

that is going to do it for the AI Daily Brief Headlines edition. Next up, the main episode. Today's episode is brought to you by Fractional. Fractional AI is my go-to AI Dev Shop. When we wanted to build an AI product feature for our company Superintelligent, we hired Fractional because they're some of the best and fastest AI engineers on the planet. The feature they built turned out great. It's already been released, and I'm about to hire them for another project, so I highly recommend them to anyone looking to build AI product features and workflow automations. The Fractional team is a group of senior engineers in San Francisco working on some of the most exciting projects in Applied AI. They work with everyone from startups all the way through the Fortune 500. To request a free consultation, head to Fractional.A.I.

If you want help identifying and building AI projects for your business, then I highly recommend that you go check them out. Hit pause on the show, open a web browser, go to Fractional.A.I, and get your free consult. Today's episode is brought to you by Super Intelligent. Regular listeners know that Super is our platform for helping people learn how to actually use AI tools. These are not long, laborious courses. These are fun, fast tutorials that get you actually using the world's most interesting and useful AI tools within minutes. If you want to build a web application with no code, we've got tutorials for that. If you want your presentations to look better than ever and take you less time than ever. We've got tutorials for that. If you want help brainstorming, writing social media

copy, and just generally working smarter, faster, and better, we've got tutorials for that. We've worked really hard to make it so that there is no better place on the internet to learn how to actually put AI to work for you, and I'd love for you to check it out. Go to besuper.a.i and use code podcast for 50% off your first month. Once again, that's besupor.aI. Welcome back to the AI Daily Brief. Today we are discussing a variety of stories around OpenAI, kicking off with reporting from the information that the company has seen significant revenue growth. According to Sam Altman's discussions with staff, as reported by the information, OpenAI has more than doubled its annualized revenue all the way up to $3.4 billion over the past six months. Previous estimates at various points we've had were the company was

on a billion dollar revenue run rate last summer. By the end of the year, it was closer to $1.6 billion. And in In addition to the overall number, we also got some more information about what it's composed of. The vast majority of OpenAI's revenue, 3.2 billion or so, comes from direct subscriptions to ChatGBT, BT, as well as from API fees. This is important because many had assumed that a big chunk, if not the lion's share, of OpenAI's revenue would come through sales of their software through Microsoft's Azure Cloud customers, but it appears that that revenue stream is only about $200 million, again, on an annualized basis, which Allman said was about 20% of the revenue Microsoft was generating

from that business. Keep in mind this is just a report and we don't have exact information. But if this report is correct, that would mean that OpenAI is making about three times as much from the sales of their software directly as Microsoft is making on OpenAI software through Azure. That would certainly be really interesting, if true. It could suggest a lot about the preference of enterprises in terms of working directly with the company rather than going through a third party like Microsoft.

It could also suggest for why Microsoft might be interested in diversifying away from just OpenAI. One thing we didn't get with this report is any amount of information about how this revenue compares to what OpenAI is spending. In other words, there's no claim of profitability or anything like that. So it's hard to get a sense of the state of the business overall. One thing that is interesting is that this certainly makes the valuation of OpenAI look a lot more reasonable. OpenAI's last valuation was around $86 billion in a tender offer of employee shares. At that valuation and at this level of revenue, OpenAI is being valued at about 25 times forward revenue, which is not only on the lower end of private AI startup financing, but it is only just a little bit above how Nvidia's stock is being valued in the public markets. OpenAI itself didn't comment on this except to say that the financial details were inaccurate, but it didn't elaborate.

Still, this does appear to put Open AI way ahead of rivals like Anthropic. Last fall, Anthropic told investors that it was generating revenue at around $100 million annualized run rate and projecting that it would reach around. $850 million in 2024. People took to Twitter slash X to speculate around which of OpenAI's revenue streams were most profitable. For example, investor Nat Friedman guessed that 50% of the revenue was coming from consumer, 35% was coming from API, including the Microsoft royalty, and about 15% from Enterprise. Arun Rao, however, had different calculations. He pointed out that 200 million consumers with 2% paid, would be 4 million paid monthly active users at $20 a month, or $80 million of revenue per month or 960 million per year, which would be about 28% of that $3.4 billion.

By way of comparison, he said that YouTube premium has a 2% conversion rate and a, quote, more widely used and mature product. However, I wouldn't be surprised if the number of people who have chat GPT accounts is actually much larger than that 2%. In fact, I think the fact that this is a less mature market is more likely to have a higher percentage of these early adopters paying for a subscription. By any measure, the revenue growth is certainly impressive, and the question is, what's likely to happen next? More specifically, how might this new Apple deal impact OpenAI's

revenue? According to the latest reports from Bloomberg's Mark German, the deal is not a cash type of financial deal. The headline reads Apple to pay OpenAI for ChatGPT through distribution, not cash. German writes, when Apple chief executive officer Tim Cook and his top deputies this week unveiled a landmark arrangement with OpenAI to integrate ChatGPT into the iPhone, iPad, and Mac, they were mum on financial terms. Left unanswered was which company is paying the other as part of tight collaboration that has potentially lasting monetary benefits for both. But according to people briefed on the matter, the partnership isn't expected to generate meaningful revenue for either party, at least at the outset. According to German sources, quote, Apple believes pushing OpenAI's brand

and technology to hundreds of millions of its devices is of equal or greater value than monetary payments. I think anyone with a pulse would agree that OpenAI getting exposure to so many new users is of huge monetary value even if Apple isn't paying directly. That said, some have pointed out as we talked about Sam Lesson on yesterday's show, that the deal in some ways does bury the chat GPT brand under Apple's, and so might not be as brand accretive as it might seem, but even more directly, the partnership could just become really expensive. If hundreds of millions of Apple users start to use ChatGPT regularly, well, then OpenAI's expenses are, of course, going to rise. The question is how much conversion to paid users of

chat GPT will offset any additional costs on the computing side. It also seems likely that OpenAI's time as the exclusive option on iOS is not going to last very long. Sources say that an agreement with Google Gemini should be in place later this year, and that Apple has also held talks with Anthropic as another potential chatbot partner. Some analysts speculate that, in fact, the goal for Apple here is to actually make money from these deals by cutting rev share agreements, where if Apple users sign up for premium versions of its AI partners' services, Apple actually gets a cut of that. This would potentially counterbalance how much it makes from Google in Google's search deal,

as users potentially shift away from traditional search engines over to chatbots. Anyways, it's a great reminder that even a very simple seeming deal can actually be immensely complex with a lot of unanticipated consequences. One of the things that people have wondered, as the Apple OpenAI deal became public, is what Microsoft and specifically its CEO, Satya Nadella, thinks of it. The Wall Street Journal is out with a new article arguing that OpenAI is just Nadella's first step. Indeed, they say that he is building an AI empire. The WSJ writes,

Nadella is not content to simply rely on OpenAI to dominate this new era. In recent months, he's been spreading his bets, turning the world's biggest company into the world's most aggressive amasser of AI talent, tools, and technology. They point to investments in companies like Mistral as well as G42, as well as the semi-acquisition of Mustafa Sullyman and his team from inflection to begin building, quote, what amounts to an in-house open AI competitor inside Microsoft. Most of the article is a reflection on how Nadella came up, how he reorganized Microsoft and how Microsoft's culture has changed over the last decade, but there is some interesting insight around the, quote, new balance of power between Sam Altman from Open AI and Mustafa Sullyman,

the former co-founder of Deep Mind and Inflection, who's now leading Microsoft's internal AI efforts. The journal writes Altman's status as the most important figure determining Microsoft's AI strategy is becoming less certain thanks to the arrival of Suleiman and his team. Microsoft Insiders say the internal politics and the balance of power between a longtime rivals Suleiman and Altman have been confusing. Anyways, it's a fascinating piece around one of the most important companies in this evolving AI space. A couple more quick hits on OpenAI in the news. Elon Musk has recently dropped his breach of contract lawsuit against the company.

The Verge writes Musk's decision to withdraw the lawsuit comes just one day before a scheduled hearing when the judge would have reviewed OpenAI's request to dismiss the case. As once again, the Wall Street Journal points out, the filing, which didn't include any explanation, doesn't necessarily mean the fight between Musk and Altman is over. Musk hasn't posted about his reasoning for withdrawing the suit or what it means for his posture about OpenAI going forward. Now, of course, it doesn't seem like Elon has changed his opinion on OpenAI, given that he

threatened to ban all Apple devices from Tesla and SpaceX, based on the OpenAI Apple Partnership. Something that is being discussed quite a bit on Twitter slash X right now is recent comments from OpenAI CTO Mira Muradi, who said that the AI models that OpenAI has behind the scenes in their labs are not particularly more advanced than those which are publicly available. Inside the labs, we have these capable models, and, you know, they're not that far ahead from what the public has access to for free. Now, the big question here is whether this is actually true. I think by and large, there is much skepticism that this is not true, although as people point out to the extent that it is, it has big implications for whether we're actually reaching a capacity plateau with current methodologies. Meanwhile, in the search for ever more advanced models, OpenAI has had to look outward for more

access to computing power. The company has recently announced that they and Microsoft are partnering with Oracle to get even more computing capacity to run ChatGBTGBT. Up until now, OpenAI has fully relied on Microsoft for its computing needs, but as the Verge puts it, as this Oracle deal makes clear, OpenAI needs more compute than Microsoft alone can give it if it wants to keep up with demand and prevent future ChatGPT outages. Whatever the stickiness of the deal behind the scenes, the market loved it with Oracle's stock up 13% based on that new AI demand in the announcement of the OpenAI deal.

So there you have it, friends. Lots of interesting things going on in the world of OpenAI. For now, though, that is going to do it for the AI Daily Brief. Appreciate you listening or watching as always. And until next time, peace.