9 AI Predictions from Jensen Huang

0:00Today on the AI Daily Brief, Jensen Huang's nine predictions for the future of AI. Before then, in the headlines, GPT-5 appears to be coming in August. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI.

0:35Welcome back to the AI Daily Brief Headlines Edition, all the daily AI news you need in around five minutes. We kick off today with sort of a good news, bad news thing. Everyone has been very excited about GPT-5 for a very long time. And heading into the summer, it seemed like the release of GPT-5 was imminent. There were rumors of Microsoft preparing additional server capacity early in the year. And more recently, people started spotting live testing of something new on ChatGPT. So the bad news is, ChatGPT still is not here, and we have not gotten it yet. But the good news is that this could be the final month of waiting, with The Verge's Tom Warren reporting that his understanding is that the release date is scheduled for early August.

1:15Remember, last week Sam Altman tweeted that they were releasing GPT-5 soon, and he's even now discussing performance of the model in the open now. In an appearance on the Theo Vaughn podcast this week, he said, This morning I was testing our new model. I got emailed a question that I didn't quite understand, and I put it in the model, this is GPT-5, and it answered perfectly. And I really kind of sat back in my chair and I was just like, oh man, here's the moment. I felt useless relative to the AI and this thing that I felt like I should have been able to do and I couldn't. It was really hard, but the AI did it just like that.

1:44It was a weird feeling. So I guess the news is that we will have to wait a few more weeks to have that same weird feeling that AI has taken another step forward. But in the meantime, it does seem like OpenAI is attempting to ship their OpenWeights model by the end of July. There is a lot of discussion around what GPT-5 is going to be like. We've discussed how big a change for many it will be to have a model router that immediately gets people to the most useful model. Gphoto writes, It's going to be really funny seeing average people get dropped into approximately two years of AI advancements.

2:16This came after Lisan Al-Gaib said, By the way, I don't want a model router. I want to be able to select the model I use. Rune responded, You'll still be able to select. This is a product to make sure that doctors aren't stuck on 4.0 Mini. Next up, welcome GitHub to the Vibe Coding Party. GitHub has launched their Vibe Coding tool, GitHub Spark. The feature allows GitHub Copilot to function in a similar way to Replit or Lovable, allowing users to develop apps using natural language prompting. Now, interestingly, Microsoft has used Anthropics Cloud Sonnet 4 as the model driving the feature, rather than any of Frenemy OpenAI's models.

2:49The feature also kind of steps on the toes of the recent integration of Replit into the Microsoft Azure ecosystem, although you know that when you're dealing with Microsoft, nothing's going to be truly exclusive. From the early reviews, the feature seems perfectly capable. Native integration with the Microsoft ecosystem is obviously a big plus, allowing fast deployment if you want to publish your idea. But ultimately, this is definitely more about a feature catch-up than any sort of big advance. Indeed, mostly the release is a big deal because of GitHub's reach. While we've talked endlessly about the incredible growth of Vibe Coding, realistically, all the tools combined still have user numbers in the low tens of millions at most.

3:25A big part of what makes it so incredible is just how much those users are able to do. GitHub gets those numbers up to nine figures, meaning that with their distribution, they have the potential to get hundreds of millions of people actually vibe coding. Eric T put it simply, I'm sure GitHub Spark will be good eventually, but what's really blowing me away is how quickly the entire AI app builder space was made into a commoditized and expected feature. We're basically back to square one where the ecosystem around the AI app is what drives people to one platform over another. Can't rule out Microsoft in that regard.

3:58Speaking of Microsoft, CEO Satya Nadella is trying to steady morale after a string of deep layoffs. In a memo sent to All Employees yesterday, he wrote, I want to speak to what's been weighing heavily on me and what I know many of you are thinking about, the recent job eliminations. After mentioning that these decisions are among the most difficult he has had to make, Nadella continued, I also want to acknowledge the uncertainty and seeming incongruence of the times we're in. By every objective measure, Microsoft is thriving. Our market performance, strategic positioning, and growth all point up and to the right.

4:25We're investing more in CapEx than ever before. Our overall headcount is relatively unchanged, and some of the talent and expertise in our industry and at Microsoft is being recognized and rewarded at levels never before seen. And yet, at the same time, we've undergone layoffs. He concluded, This is the enigma of success in an industry that has no franchise value. Progress isn't linear. It's dynamic, sometimes dissonant, and always demanding. but it's also a new opportunity for us to shape, lead through, and have greater impact than ever before. And after addressing the specter of layoffs, Nadella tried to promote the company's three main business priorities, security, quality, and the AI transformation.

5:00He urged employees to keep a growth mindset, but didn't make any promises of job security for the 200 ,000 workers that remain. Instead, he wrote, it might feel messy at times, but transformation always is. Teams are reorganizing, scopes are expanding, new opportunities are everywhere. And this gets to something that we'll speak to a little bit later in the main part of the episode. We're talking about NVIDIA CEO Jensen Huang's predictions for the future, and a lot of it's very optimistic about the jobs that AI creates. But even as it does so, the process of getting there is going to be a messy one.

5:30It's going to be painful, confusing, with lots of very real people's very real lives affected in the meantime. We're going to need to be able to think in terms of both the future and the now simultaneously and figure out how to navigate the through line. Speaking of which, Scuttlebutt continues around the Google-Windsurf deal. You'll recall that earlier this month, we had one of the more dramatic acquisitions in recent venture history when Google arranged an acquihire deal with Windsurf after their OpenAI deal fell through. Google hired the CEO and select members of the engineering team, leaving several hundred employees in limbo.

6:01By Monday, Cognition Labs had snapped up the remains, neatly tying up the narrative. Software engineer Prem Kuhner has now revealed what the hectic weekend looked like from the inside, posting, I've joined Cognition to continue to work on the future of software engineering. I was employee number two at Windsurf and have worked on AI code for years. There's never been a more exciting time and place for it now than at Cognition. I had a place at Google DeepMind as part of the deal. I won't go into detail for legal reasons, but I'd like to be transparent about my personal situation. I was given an offer that would explode same day.

6:32I had to forfeit all of my vested shares earned over my 3.5 plus years at Windsurf. I was ultimately given a payout of only 1 % of what my shares would have been worth at the time of the deal. In going to Cognition, I've chosen a different direction. For someone who loves software engineering, Cognition feels like home. It reminds me of the energy of the earliest days of Windsurf, where we wrote excessive amounts of code and had excessive amounts of fun. Really excited to see how we can take the best of Devin and Windsurf to make the world's best IDE and coding agents. Now, I'm not sure how much this elucidates versus further confuses.

7:02A lot of people are confused about how the terms would have worked that would have only given him a 1 % payout. But one thing that is resonant and is very upsetting to people is this idea of exploding offers. This is something that appears with much more frequency than it should. And the issue, of course, is not just that it screws the people who are involved in the immediate term, it's that it creates a totally different type of incentive set for everyone who's even considering joining a startup. Ishan Goswami writes, if this keeps happening, then no smart person would join as a founding engineer.

7:31Overall, not negative for the startup ecosystem. Chao Wu writes, the startup ecosystem should absolutely condemn these types of deals. Bren Chay Smith writes, if aqua hiring becomes the norm, early stage VC will consolidate and at the hands of a small few, a fraction of capital will be allocated to a fraction of ventures at pre-seed, and smart people will stop working at startups. The windsurf deal is truly awful for the ecosystem. And if you need a signal that things are actually changing, look no further than herald of the startup apocalypse, Roy Lee. Roy is of course the CEO at Cluley, who is setting off all sorts of fires with his extremely different, aggressive, and even some would say nihilistic approach to winning the startup game.

8:09Even if you don't agree with Roy, he is 100 % a bellwether of where young founders are thinking. And that's why I took notice when the other day he wrote, a startup truth I disagree with. Don't pay high cash compensation. We now pay$250k to$350k base for designers and$300k to a million base for software engineers. He then gets into why, but this is the sort of thing that's going to be required if everyone assumes that they're just not going to get a big payday when the founders get acquired. Anyways, super interesting stuff, beyond the scope, certainly for the headlines. And with that, let's wrap and move to the main episode.

8:44This episode is brought to you by Blitzy, the enterprise autonomous software development platform with infinite code context. Blitzy uses thousands of specialized AI agents that think for hours to understand enterprise-scale code bases with millions of lines of code. Enterprise engineering leaders start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzy platform provides a plan, then generates and pre-compiles code for each task. Blitzy delivers 80 % plus of the development work autonomously while providing a guide for the final 20 % of human development work required to complete the sprint.

9:15Public companies are achieving a 5x engineering velocity increase when incorporating Blitzy as their pre-IDE development tool, pairing it with their coding co-pilot of choice to bring an AI native SDLC into their org. Blitzy is providing a limited time, 30-day free proof of concept for qualifying enterprises. The team will provide a 5x velocity increase on a real development project in your org. Visit blitzy.com and press book demo to learn how Blitzy transforms your SDLC from AI-assisted to AI-native. That's blitzy.com. As a founder, you're moving fast towards product market fit, your next round, or your first big enterprise deal.

9:52But 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. Fast-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.

10:26The 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. Today's episode is brought to you by superintelligence, specifically agent readiness audits. Everyone is trying to figure out what agent use cases are going to be most impactful for their business, and the agent readiness audit is the fastest and best way to do that.

11:02We use voice agents to interview your leadership and team and process all of that information to provide an agent readiness score, a set of insights around that score, and a set of highly actionable recommendations on both organizational gaps and high-value agent use cases that you should pursue. Once you've figured out the right use cases, you can use our marketplace to find the right vendors and partners. and what it all adds up to is a faster, better agent strategy. Check it out at bsuper.ai or email agents at bsuper.ai to learn more. Over the last couple of years, NVIDIA CEO Jensen Huang has become an extremely notable figure.

11:38In the business and startup world, he's basically a one-name founder. But in the halls of power in Washington and Beijing, he has taken on an even more important significance. NVIDIA's AI chips are not only at the very center of the AI revolution, They are also at the center of the most important geostrategic relationship and conflict of our time, which is, of course, the constantly shifting relationship between the United States and China. Jensen is one of the few CEOs who actually bridges that gap, with this week being a perfect example. At the beginning of the week, he attended meetings with Chinese leaders, and then immediately flew back to Washington to play a central role in the Trump administration's week-long event about winning the AI race.

12:15Now, there was lots of Jensen speaking this week. There's, frankly, at this point, always a lot of Jensen speaking. But one event was a conversation with the folks from the All In podcast. They spoke with a handful of leaders, including AMD CEO Lisa Su, as well as several figures from the administration. The conversation with Jensen was definitely the most interesting, however, because whereas everyone else was focused on the topic du jour, which was what is needed in the next few years to cement the U.S. as the undisputed leader in AI, Jensen instead zoomed way out to the long-term trends. Prediction one.

12:46Jensen said AI will create more millionaires in five years than the internet did in 20. So obviously one of the big topics of conversation was Mark Zuckerberg's recent poaching spree. Co-host and former Meta employee Chamath Palahapitiya remarked that these huge contracts were happening exclusively at the model research layer, but don't seem to be happening at the hardware layer. Now, a couple things to note about this. First of all, it seems pretty clear to me that the reason that Zuckerberg is paying so much for these people is not just their outsized talent. It's the fact that they basically have IP in their heads.

13:15When you reframe it to think about basically taking all of OpenAI's IP, then paying$3 billion or whatever it ends up being for access to$300 billion worth of IP or whatever OpenAI's most recent valuation is, looks a little less crazy. Still, Jensen, however, wanted you to know that his people are doing just fine. He said, I've created more billionaires on my management team than any CEO in the world. Don't feel sad for anybody at my layer. My layer's doing just fine. More broadly, though, that's where he got into this wave of wealth creation, which he says is going to be more rapid than any other event we've ever seen.

13:47And building on that point about wealth creation, prediction number two, Jensen believes we're starting to transition to an era where elite human labor will be valued in a similar way to premium capital goods. The big idea, he said, is that the impact of 150 or so AI researchers can probably create, with enough funding behind them, create an open AI. Chamath was shocked by that number, but Jensen reinforced that DeepSeek is 150 people, Moonshot, the creators of Kimmy, which has gotten a ton of buzz, are 150 people. Basically, that the actual number of researchers creating this technology is vanishingly small.

14:18And frankly, Jensen pointed out that until Zuckerberg realized it earlier this year, no one had really done the math on how much these people are worth. Jensen commented,

14:33Prediction number three is basically about jobs. And rather than worrying about AI disrupting jobs, Jensen is more concerned about not being able to create jobs fast enough to keep up. He commented that 100 % of his workers use AI and the company is busier than ever. Not only are there zero layoffs, they can't keep up with their own possibilities. Now, obviously, NVIDIA is a bit of a special case. But I think that if every CEO emulated the way that Jensen was thinking about AI, it would lead to some better results. You've heard me squawk on endlessly here about efficiency AI versus opportunity AI, and it's very clear that Jensen is living in the opportunity side.

15:08He said, we have so many ideas that we want to go and pursue. AI makes it possible for us to go and pursue those ideas now that we're not doing the mundane stuff. This idea of not doing the mundane stuff is still a type of framing we use, but I really think it's even more fundamental than that. Especially as agent capabilities increase and we get better at managing teams of agents that can work together to do increasingly more complex work, the benefit won't just be that we have more time to do higher order work. it'll be that we have incredible amounts of intelligence that we can deploy on our own behalf to do that work.

15:40In other words, it's not just what we can do with the freed up time based on AI taking on the boring stuff, it's what we can do at that time with armies of AI and agents backing us up. A fourth big idea or prediction about AI from Jensen was around how AI changes capabilities, specifically with programming. Jensen called AI the greatest technology equalizer of all time. Now, of course, the internet was the big equalizer when it came to geography. A company selling books out of Washington state could have the same distribution power as a company with hundreds of locations in major cities across the US.

16:10Jensen is articulating that AI could have the same effect for skills. He said, everybody's a programmer now. You used to have to know C, then C++, and then Python. Even if you don't know how to program an AI, just go up to the AI and ask, how do I program an AI? It's a great equalizer. Everybody is going to be augmented by it. Now, one interesting example of this was with the Norway Sovereign Wealth Fund we talked about recently, where half of their team was now coding thanks to AI tools. Prediction five, taking it a step further, it's not just that everyone is a programmer now. Jensen continued, everybody's an artist now.

16:42Everybody's an author now. Now, this obviously is deserving of some nuance. As AI normalizes and equalizes certain types of productive outputs across lots of different skill sets, we will reset the gradients of what high skill means, even if that high skill takes into account how well you use AI. But the point that Jensen is making is that everyone's productive capability is going way, way up. That on average, we will be able to produce much, much more each individually than anyone could have produced in the past. Jensen reinforced the point that the flip side of this is that everything is changing.

17:15He said everybody's jobs will be different. Many jobs will be obsolete, but many jobs will be created. Now, as we know, none of this, of course, speaks to the difficulty of transition. One of the things that I harp on, and I will continue to, is that it is completely coherent and non-mutually exclusive to be extremely excited and optimistic about the future of a world with greater capability, greater capacity, new jobs, more outputs, and yet still be concerned about the process of creative destruction and upheaval that will absolutely happen and impact individual people in the transition period.

17:48Ultimately, that's not what Jensen was talking about, but I never want to brush past that and live entirely in the land of future utopia. Moving back to the technical a little bit, a sixth prediction is Jensen's concept of twin factories. Now, Jensen first unveiled this concept at NVIDIA's GTC event in March, but the idea was still fairly narrow. He talked about the need to create two factories for manufacturing work, a physical one that does the actual production, and then a digital twin that does all the prototyping and simulation of physical processes. In March, Jensen's big idea was that you can now use AI simulation to train robots, test production lines, and troubleshoot problems.

18:25All the tasks that could previously shut down a factory at great expense. Now Jensen's expanding that concept to everything, stating, everything that moves will be autonomous. Every company in the future will have two factories. There's the machine factory that creates a product, for example. Then there's the AI factory to create the AI for the cars. The future of industry is really two factories. And it's not just industry either. Jensen suggested that even things like air traffic control might one day be a human workforce overseeing a giant AI. He commented, In the future, every industrial company is going to be an AI company, or you're not going to be an industrial company.

19:00Prediction number seven was for all of the folks on Wall Street who are concerned about overspending on CapEx. And I guess it follows from the fact that Jensen is predicting that every company will be an AI company. But he suggests that right now, despite how much money has been spent, were actually only a few hundred billion dollars into what will be a multi-trillion dollar build-out. Now, holding aside limits around energy and infrastructure capacity, which of course is a lot of what the White House's AI action plan was all about, fundamentally Jensen's concept is that the demand for AI is orders of magnitude larger than most people assume.

19:32There are still many people who are thinking about AI as an additional computer program that folds into the current ecosystem. Jensen, however, views this as a fundamental revolution in tech, and really, in many ways the first one we've seen in decades. The big pull quote here, he said, we are reinventing computing for the first time in 60 years. This is a very big deal. Prediction aid is around the same themes, and basically it's that we are about to see a huge infrastructure gold rush. Jensen said, in Arizona and Texas, we're probably going to produce half a trillion dollars worth of AI supercomputers, and that's probably going to drive a few trillion dollars worth of AI industry.

20:06And that's only over the next several years. And the point that he was making is not just that this is good for investors or for the AI industry, but that this fundamentally changes the shape of the US economy and the economic strategy of the nation. Speaking to economic theory and protectionism, Jensen commented, people always degrade down to tennis shoes. We don't have to go there. We'll just manufacture chips and supercomputers. Basically, what Jensen is saying is that if the spectrum of options is throw up walls on the one end and outcompete on the other, he is very firmly in the outcompete column.

20:36Number nine, pulling it all together, Jensen made the case for why the American tech stack needs to be at the center of this. He remarked, could you imagine if DeepSeek came out and it only worked on the non-American tech stack? Could you imagine if the same was true for Quinn or Kimmy? These are the top three open models in the world today, downloaded hundreds of millions of times. So the fact of the matter is, the American tech stack being the world's standard is vital to the future of winning the AI race. Any computing platform wins because of developers. Half of the world's AI developers are in China.

21:06Now, in some ways, this was one of the more clear articulations that we've heard from Jensen on why he's wanted to push NVIDIA chips back into China. Outside, of course, obviously the money and his fiduciary responsibility to his shareholders. But the point that he was making, or at least the subtext of it, was that NVIDIA got to the top of the stack in AI chip making by being early, but they retained their lead and expanded their lead through the CUDA programming platform. AI models developed to run on NVIDIA systems run into compatibility issues on other systems. There's workarounds, but CUDA has been an incredible moat for NVIDIA.