How to Be An AI Leader (According to OpenAI)

0:00Today on the AI Daily Brief, how to lead in AI according to OpenAI. And before that, in the headlines, it looks like Apple might finally be making a big AI play. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.

0:24All right, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, KPMG, Fanta, Superintelligent, and Robots and Pencils. To get an ad-free version of the show, go to patreon.com slash AIDailyBrief. Ad-free starts at just$3 a month. And if you're interested in sponsoring the show, shoot us a note at sponsors at AIDailyBrief.ai. Welcome back to the AIDailyBrief Headlines Edition, all the daily AI news you need in around five minutes. Could it actually be happening? Are we actually going to get a play from Apple on AI? The Cupertino giant appears to have made a call and will partner with Google on their next AI offering.

1:02Bloomberg's Apple insider Mark Gurman reports that the project will be an AI-powered search engine known internally as World Knowledge Answers. Sources said that the feature will be integrated into Siri and that Apple are also considering building it into Safari and their home screen search widget Spotlight. The feature is expected to be ready by the spring as part of a long overdue Siri overhaul. Sources said that the idea is to replicate the performance of Google's AI overviews or perplexity. The underlying technology enabling the new Siri could come in part from Google, Apple's longtime partner in internet search.

1:33The company's reached a formal agreement this week for Apple to evaluate and test a Google-developed AI model to help power the voice assistant. The reporting also spoke to a much broader Siri revision that's in the works. The new Siri is intended to be able to tap into personal data and on-screen content to better fulfill queries. This was, of course, the original idea of so-called Apple intelligence to have the context of your personal experience expressed through mobile that would make AI better for you. Now, of course, none of that vision has come to fruition, but the vision was something that many people were interested in, theoretically at least.

2:07The reporting also suggests that Siri will become a computer use agent, able to navigate an Apple device according to voice instructions. Craig Federighi, Apple's head of software engineering and the executive in charge of the Siri project, said in a race in all hands, the work we've done on this end-to-end revamp of Siri has given us the results we've needed. This has put us in a position to not just deliver what we announced, but to deliver a much bigger upgrade than we envisioned. The project has drawn in talent from multiple teams, including the AI division and the Apple services unit helmed by executive Eddie Q.

2:36You might recognize that name as the person who has reportedly been pushing for acquisitions as a path for Apple to get back in the AI competition, although he has not been successful in convincing his colleagues around that. Siri's new architecture is reportedly split into three components. A planner, a search system, and a summarizer to put it all together into an output. Sources said that Apple is leaning towards using a custom-built version of Google's Gemini model as the summarizer. It would need to be modified to run on Apple's proprietary hardware, and the two companies are now collaborating on fine-tuning and testing.

3:05Apple is also considering using Gemini for the planner, although Anthropics Cloud and Apple's own in-house models are still being considered as well. The reporting also said that Gemini hasn't been ruled out to drive the search system. In other words, in six months' time, we could have a version of Siri built entirely on Google's tech. Reportedly, Anthropic's cloud actually outperformed Google in the Siri bake-off, but Anthropic demanded too much money for the use of their model. Sources said that they asked for more than$1.5 billion a year, while Google was up to much more favorable terms. Now, one unspoken part of the reporting that I think is pretty interesting is the fact that OpenAI is nowhere in these conversations, despite the fact that ChatGPT was the first third-party AI app that Apple started pushing on the iPhone in that presentation about a year ago.

3:46I don't know what happened with that relationship, whether the fallout is from a business or a technological perspective, but it is notable that OpenAI is nowhere to be seen here. Still, I think at this point for most people, just having Siri that actually works like anything, even close to how you would expect in 2025, will be a big win. Speaking of OpenAI, there has continued to be incredible demand for shares in that company. They have boosted their secondary share sale up to$10 billion, an increase from the$6 billion reported last month. Current and former employees who have held shares for more than two years will have until the end of the month to decide whether they'll access that liquidity.

4:20The round is expected to close in October. Now, the round is noteworthy because it's the first to test OpenAI at a$500 billion valuation. Their last fundraising was of course conducted at the start of the year at$300 billion. Since then, the company has doubled its revenue and basically everything else, including users. So some big markups sort of make sense, although the numbers are eye-popping to some. Still, when push comes to shove, we continue to see the trend of demand for AI startup investments vastly outstripping supply as the number of shares that the market wants to consume just seems to keep going up.

4:51Another update on the AI fundraising front, Mistral is set to finalize a 2 billion euro investment that would value the company at roughly$14 billion. The rumor from last month was that they were seeking a billion dollars in fundraising at a$10 billion valuation, so quite a lot of additional funding and a meaningful bump in the valuation seems to have come through. This will be the first round of Mistral fundraising since June of last year when they were valued at 5.8 billion euros. Now, not only does this mean their valuation has doubled, but more notably, it'll be the first time that they've had access to such a significant war chest.

5:22They've raised around a billion euros in total in the two and a half years since they were founded, so this would be double that all in one fell swoop. It's still not super clear to me exactly where Mistral is going to fit in this ecosystem, but it's clear that there's a lot of folks who think that they have a place somewhere. Quick update on talent drama, XAI's CFO has left after just three months on the job. The Wall Street Journal reports that CFO Mike Liberatore left the company around the end of July after commencing in April. Liberatore oversaw XAI's debt and equity raise in June, which brought in$10 billion in fresh capital for the company.

5:54SpaceX contributed almost half of the equity raise, which suggested to some that demand from outside investors was comparatively sparse, at least relative to all these other rounds that we're seeing. Now, while it's dangerous to draw too many conclusions from a single departure, this is the latest in a recent trend. At the beginning of August, XAI's general counsel, Robert Keel, announced he was leaving after a little more than a year on the job. In his farewell post, he claimed he was leaving to spend more time with his two toddlers, something that any toddler parent will understand, but also noted that when it came to Elon Musk, quote, there's daylight between our worldviews.

6:24Raihu Rao, a senior lawyer who dealt with commercial affairs at the company, left around the same time. And then on August 13th, co-founder Igor Babushkin announced that he was leaving to start his own venture firm. There was, of course, also ex-CEO Linda Yaccarino announcing her departure in July a few months after the social media platform's merger with XAI. Occam's razor continues to be that people in high-impact jobs move around a lot more than people where the stakes are lower. But there's enough happening there that for anyone who's really interested in how much the competition between all these labs is going to increase the availability of cool products for all of us, let's just say it's worth paying at least a little bit of attention to.

6:57Lastly, a sort of messy one, Scale has sued rival data labeling startup Mercore, accusing the firm of corporate espionage. In a lawsuit filed on Wednesday, Scale claimed that their former head of engagement management, Eugene Ling, downloaded more than 100 customer strategy documents while in communication with Mercore. He's accused of meeting with Mercore CEO Brendan Foody to discuss business strategy and product while still in his role at Scale. The lawsuit claims that Ling was later hired in order to build a relationship with one of Scale's largest customers, and that the bulk of the documents relate to that customer.

7:27Bling is allegedly still assigned to this customer's account while working at Mercor. Surya Midha, a Mercor co-founder, responded to the lawsuit, stating, While Mercor has hired many people who departed Scale, we have no interest in any of Scale's trade secrets, and in fact, are intentionally running our business in a different way. Eugene informed us that he had old documents and a personal Google Drive, which we have never accessed and are now investigating. We reached out to Scale six days ago, offering to have Eugene destroy the files or reach a different resolution, and we are now awaiting their response.

7:54Now, of course, the situation is made even messier due to Meta's acquihire deal with Scale. It has been well reported that Scale lost multiple major clients in the wake of that deal, as rival AI labs didn't necessarily want to continue doing business with them after Meta became a major owner. Then, of course, there's the fact that earlier this week, it was reported that Meta themselves had also moved away from using Scale's data labeling services, adding rival providers, including Mercore. So I guess maybe the takeaway that is if you are looking for signs that AI is slowing down when it comes to things like big company competition, talent wars, fundraising, or basically any other vector, the answer is most distinctly no.

8:29However, that's going to do it for today's AI Daily Brief Headlines edition. Next up, the main episode. What if AI wasn't just a buzzword, but a business imperative? On You Can With AI, we take you inside the boardrooms and strategy sessions of the world's most forward-thinking enterprises. Hosted by me, Nathaniel Whittemore, and powered by KPMG, this seven-part series delivers real-world insights from leaders who are scaling AI with purpose. From aligning culture and leadership to building trust, data readiness, and deploying AI agents. Whether you're a C-suite executive, strategist, or innovator, this podcast is your front-row seat to the future of enterprise AI.

9:05So go check it out at www.kpmg.us slash AI podcasts, or search You Can With AI on Spotify, Apple Podcasts, or wherever you get your podcasts. As a founder, you're moving fast towards product market fit, your next round, or your first big enterprise deal. But with AI accelerating how quickly startups build and ship, security expectations are higher earlier than ever. Getting security and compliance right can unlock growth or stall it if you wait too long. With deep integrations and automated workflows built for fast-moving teams, Vanta gets you audit-ready fast and keeps you secure with continuous monitoring as your models, infra, and customers evolve.

9:44Fast-growing customers like Langchain, Writer, and Cursor trusted Vanta to build a scalable foundation from the start. And look, as someone who lives in the world of enterprise procurement, I love how Vanta makes it easy to get compliance right. The last thing you need when you're trying to win that big deal is to have it scuttled by something that Vanta has solved for over 10 ,000 companies. Go to vanta.com slash NLW to save$1 ,000 today through the Vanta for Startups program and join over 10 ,000 ambitious companies already scaling with Vanta. That's V-A-N-T-A dot com slash NLW to save$1 ,000 for a limited time.

10:41your agent readiness audits, we help you double-click on your most important use cases with what we call our use case planning reports. These reports are going to help you understand what sort of technical preparation you need to do to be ready for a use case, what challenges you might face in implementation, and whether you should be thinking about building, buying, partnering, or some combination. After that, you can even get a spec document in what we call our technical blueprint that gives either your developers or the developers of the partner you work with what they need to build exactly the agent that you're looking for.

11:10If you want to learn more about Superintelligence Agent Planning Suite, we built a custom GPT to answer your questions. Just go to bit.ly slash super super agent. That's bit.ly slash super super agent, all one word. And if you have any questions, the agent can even help you book an appointment with our team. Small, nimble teams beat bloated consulting every time. Robots and Pencils partners with organizations on intelligent cloud-native systems powered by AI. They cover human needs, design AI solutions, and cut through complexity to deliver meaningful impact without the layers of bureaucracy. As an AWS-certified partner, Robots & Pencils combines the reach of a large firm with the focus of a trusted partner.

11:52With teams across the US, Canada, Europe, and Latin America, clients gain local expertise and global scale. As AI evolves, they ensure you keep peace with change. And that means faster results, measurable outcomes, and a partnership built to last. The right partner makes progress inevitable. Partner with Robots and Pencils at robotsandpencils.com. Welcome back to the AI Daily Brief. Today we are talking about how to be an AI leader, at least according to a new guide from OpenAI. OpenAI just recently published a report called Staying Ahead in the Age of AI, a Leadership Guide. And it's part of a pattern where they have been releasing a set of resources at a fairly steady clip to help organizations and enterprises think about their AI strategy and implementation, and hopefully get some tips that help them along the way.

12:41None of these are meant to be super comprehensive or replacements for big, deep strategy that's run internally, nor is it meant to replace the work that consultants and partners and strategists might do. Instead, these documents tend to be tips and examples that OpenAI has seen in their own work with their customers and clients, captured and reshared for a broader audience. Now, one quick note, sometimes you will hear me say that if you're not watching an episode, you really should head over to YouTube or Spotify to watch it as well. This is kind of the opposite. This is one of the more visually uninteresting topics we've had because we're just examining a report, but I do think there's a lot to discuss here and I'm excited to dig into it.

13:15Now, first of all, from a context or setup perspective, this is a subtle or not really so subtle reminder that you are really on one if you think things are slowing down right now. I don't know when they started working on this. It was probably before the release of GPT-5 and all of the hullabaloo around that. It was probably before the MIT study came out. It was probably before the AI bubble conversation started happening in earnest. But it definitely smacks you over the head with the idea that if you are moving slowly because of any of those things, you are not going to make it. Indeed, the first thing that OpenAI does is drop a bunch of stats to point out how absolutely not slowing down this thing is.

13:52The first full sentence in the report is, artificial intelligence is accelerating on every front. And they give four statistics to validate that point. The first thing they argue is that AI capabilities have grown 5.6x when it comes to frontier models since 2022. That's based on a study from Epic AI called the pace of large-scale model releases is accelerating. Next, they talk about cost. One of the things that we've been talking about on the show recently is why we should include cost and efficiency of intelligence in part of our calculations for how advanced models are, given that we're increasingly on the frontier of wanting to use a lot more tokens to unlock new use cases because the cost of intelligence is coming down.

14:31They point to a 280x decrease in the cost of running a GBT 3.5 class model in just 18 months. Now obviously that number is a very big impressive number. I think the reality is that the vast majority of use cases that are in production today require at the very least 4 and really more like a 4.5 style capability. But those numbers are frankly similarly precipitous when it comes to how fast they are decreasing. Lastly, they point to the fact that not only is AI being adopted faster than desktop internet was, it's being adopted four times faster. Now those are the big banner statistics, but buried just underneath is one that I think might be even more significant at least for enterprises who are reading this report.

15:12They reference a BCG study called Where's the Value in AI that found that organizations that they designated as AI early adopters were growing revenue 1.5 times faster than their peers. So that's the setup. Basically, they are saying, like I say at every keynote presentation I do, the speed of disruption is faster than you think, and the magnitude of disruption is even bigger than you think. So that's the setup. But what do they actually cover in the next 12 or so pages of the report? They've divided things into five principles. Align, activate, amplify, accelerate, and govern. You got to have that alliteration in there.

15:45Would it even be a corporate report without alliteration. Now still, trying to zoom out, I think that the subtext here is that this is not a guide that's for some super advanced organization. This is not your 301 or 401 graduate level course on how to be an AI leader inside your AI-enabled or AI-native company. This is instead the basics. And frankly, I think there's a subtext here that's arguing that right now at this stage, You don't necessarily have to be a super advanced AI user to be, relatively speaking, ahead of the pack when it comes to using AI across your organization. Effectively, OpenAI is arguing that if you take these basic steps, you will still be ahead.

16:24Now, I will say, and we're going to get into this quite a bit towards the end, the whole thing does a little feel like a 2024 pre-agentic, pre-reasoning model era kind of guide. Specifically, pre-agentic is the one that I think is notable, which as you will see does not mean that I think it's not useful. I think it is. But it's missing this whole big section of what it means to engage with AI right now that I do think limits its total applicability. However, like I said, we're going to come back to that. Overall, it attempts to be very practical giving specific advice and specific examples. So let's quickly go through these sections and see what they are recommending.

17:00The first section is about alignment and getting employees and managers on the same page when it comes to AI. Now, this is something you've probably heard me talk a lot about. One of the big challenges for enterprise AI adoption that has absolutely nothing to do with model capability is the often large gap between how employees and managers are thinking about AI strategy in every way from how well articulated that strategy is, to how pro-employee that strategy is, to whether employees are supported enough in implementation of that strategy, and so on and so forth. So I think broadly speaking, thinking in terms of alignment is absolutely a necessary step, can fully co-sign based on everything that we see at Super Intelligent and everything that I hear doing this podcast as well.

17:42The four alignment practices that they discuss are one, executive storytelling to set the vision, two, setting a company-wide AI adoption goal, three, leaders role modeling AI use, and four, functional leader sessions, which is basically leaders modeling AI use, but bringing it down to the line of business level and closer to the point of actual implementation at the employee level. Now, one thing I will note, and I will warn you if you do not want to be shilled, press that fast forward button maybe once or twice. A lot in here kind of presumes that the use cases are known going into the conversation.

18:14In other words, that managers who are setting the vision for AI know what they should be using AI for. In my experience and in our experience at Superintelligent, that actually tends to be a major barrier that slows things down right at the top. It's literally why we designed these agent opportunity mapping audits that we deploy to make it simpler to actually understand which use cases might be viable and useful for an organization based on its own particular unique characteristics. So as always, if you are interested in those audits, shoot me a note at nlw at bsuper.ai and I will connect you to the right people.

18:46Now, moving back to OpenAI's practices, in addition to just the best practices that they're sharing, they also in many cases share examples that they've seen. One that they point to, for example, is the CEO of Moderna suggesting that their employees should be using ChatGPT 20 times a day. As we'll discuss in just a minute, there is a big difference between saying that should be the norm and actually holding people accountable for that. But still, I do think that this idea of getting specific about AI adoption goals is going to be really valuable. Now, one more note in each of these sections, after the four best practices that they share, they give a set of starter questions and example actions.

19:19So in this case, one might be a question, are we transparently communicating our progress? And a suggested action of maintaining and openly reviewing a dashboard to clearly track AI progress. The next practice area for this leadership report they call Activate. They noticed that on the one hand, almost half of employees said that they lacked the training and support needed to confidently and successfully use generative AI, despite the fact that they rank training as the single most important factor for actually adopting AI and using it successfully. The four best practices then that OpenAI points to are launching a structured AI skills program, establishing an AI champions network, making experimentation routine, and as they put it, making it count, or maybe a more simple way to put it, linking AI usage to performance.

20:05Now, I want to talk specifically about these last two, routinizing experimentation and linking it to performance. Across a huge percentage of the conversations that we have, and certainly the feedback that I get when I'm at, for example, executive sessions, or doing keynotes with big groups of AI practitioners, is the griping and bellyaching about the time that it takes to get good at AI and how infrequently that time is actually reflected in management's approaches to AI adoption. In other words, it tends to be the norm that even if leaders inside a company have articulated some sort of AI adoption goal, they haven't created the actual time and space to allow the people to do the hard work of getting good at using these tools.

20:43Employees are then just expected to find time around the edges to additionally add on this new layer of work, which is learning the AI tools, on top of doing everything else that they do. OpenAI's suggestion is to give employees regular time to explore AI tools. And I wanted to double click on this because this is an easy one to forget, but one of the most controllable things that leaders can do inside an organization. OpenAI goes on to suggest, try dedicating the first Friday of each month for teams to workshop how AI could improve their work. And even something as simple as this would make a meaningful difference in many of the organizations that we run into.

21:16Now, the second thing that I wanted to point to is this idea of directly linking AI engagement to performance evaluations. This is a major trend that we're seeing. For the last couple of years, in many organizations that really wanted to be AI forward, they spent a bunch of activation energy trying to get people to adopt AI because they argued it was going to be helpful for them. Basically, the most aggressive AI organizations were the ones who were regularly and specifically encouraging AI use. Over the last six months or so, we're seeing a much more dramatic shift to just mandating use. At this point, there are increasingly use cases where you are just frankly behind and not taking advantage of the current tools of the moment if you are not using AI.

21:54And increasingly, we're seeing companies who are not just thinking about that as a nice productivity enhancement, but as something that's actually a problem if you are not using it. Now, of course, that's the negative side of linking AI engagement to performance evaluations. There's also the positive upside where this could be a way to lever up and move your career growth faster. But the point that I wanted to add to the broader conversation is that moving towards more specific mandated usage is absolutely a trend. The third practice area for OpenAI's leadership guide is Amplify, by which they basically mean sharing information across the organization.

22:25They write the fastest way to scale AI impact is to stop solving the same problems in silos. Amplifying progress means turning scattered into shared knowledge, documenting successful prompts, workflows, and use cases so other teams can reuse, improve, and build on them. The funny thing about this is that this is actually exactly what a previous version of super intelligent was trying to be before we moved off of it to focus on agents. But I wholeheartedly agree that trying to surface new use cases, new best practices, and not force everyone into this weird role of having to be a use case discoverer is a key for successful enterprise AI adoption.

22:58So what does that look like in practice? Well, OpenAI suggests it's things like launching a centralized AI knowledge hub, insistently sharing success stories, building active internal communities, and reinforcing wins at the team level. Again, to me, this is an area that is completely inside your control and is just good AI leadership hygiene at this point. Next up, OpenAI's idea of acceleration is basically all about removing friction and increasing the speed of decision making. Now, there is a subtle thing here that I think is worth calling out. We're still often at a stage where companies are looking to do things that make them feel like they're making progress without really having to figure out how to do things differently.