OpenAI CEO Sam Altman Predicts AI’s ‘Make or Break’ Moment is Almost Here

Watch the Video

Key Points In This Video

• Apple is the one Magnificent 7 company not actively involved in the race to build massive models. Instead its relying on partnerships with companies like OpenAI. 
• At 24/7 Wall St., we believe the release of reasoning models – roughly corresponding to the release of o3 in December last year … Is the next step-change in AI model capabilities. 
• However, a recent paper from Apple researchers casts doubt on the overall capabilities of reasoning models.
• The paper – named ‘Illusions of Thinking’ – detailed how reasoning models solve a series of puzzles. The paper’s findings were essentially that reasoning models capabilities are more limited than believed – they’re just advanced pattern matching. 
• Our take: we don’t find the paper especially compelling. We won’t get into too many details, but AI researchers we follow have poked some pretty sizable holes into how the study was conducted and its overall conclusions. 
• Here’s what we think is important to watch if you’re concerend about Apple’s recent paper. One of the biggest criticisms of AI has been: Where are the novel discoveries?
• That is to say, while AI has contributed to some major discoveries like protein-folding, there have been no major discoveries directly made by leading AI models.
• Sam Altman, CEO of OpenAI, recently published a blog that’s worth reading. It’s named ‘The Gentle Singularity.’ 
• In it, he lays out a timeline we can come back to and hold him accountable to. 
  • 2025: Rise of Agents 
  • 2026: Novel insights from AI 
  • 2027: Robots that can do tasks in the real world 
• So, essentially he’s dropping the gauntlet that GPT5 will be powerful enough for these novel discoveries… 
• If that’s the case, the upside for AI essentially becomes infinite; it is the driver for economic progress for the remainder of our lives. 
• If it’s wrong… Total spend on AI is likely too high at current rates. 

The AI Investor Podcast Is Your Source What Could Become Technology’s Biggest Trend Ever

This discussion comes from 24/7 Wall St.’s AI Investor Podcast.

The podcast covers the biggest developments in AI and how you can build a portfolio filled with the highest-upside stocks in the space. You can listen to our most recent episode on YouTube or leading podcast apps like Apple Podcasts or Spotify.

Transcript:

[00:00:05] Austin Smith: And now, you know, a, a counterpoint to that is I wanna talk about this Apple Illusions of thinking paper, because Apple is one of the largest tech companies that, you know what I mean? Like, like Google (NASDAQ: GOOG) and Microsoft was really hot on AI. But in their most recent, uh, worldwide developer conference, there was way less AI, way less Apple intelligence than in the prior release.

[00:00:24] Austin Smith: So on one hand, we’re seeing Apple sort of standalone here, and some of the, the steam is coming out of their own AI enthusiasm and they’re releasing this illusions of thinking paper that sort of puts a little bit of cold water on the AI narrative. So, but, but everything else we’re seeing is an increase in demand and compute and inference from everybody else in Oracle (NYSE: ORCL) and Microsoft and Google.

[00:00:45] Austin Smith: So. What is going on with Apple? Are they, you know, is this a cultural change? Is this, is this something? Are they seeing something other people don’t? Why does Apple sort of stand alone here in this super major AI space?

[00:00:58] Eric Bleeker: Yeah. Apple’s such an interesting company, right? Because everyone else is spending tens of billions of dollars on CapEx and they’ve kind of not played that game.

[00:01:07] Eric Bleeker: They can do that in part because they have unique distribution right at the start of smartphones. You have said their greatest weakness is. Something like search all these Google services, they, they don’t control that and they’re beholding, but the end outcome was they end up charging Google 20 or $30 billion per year in pure profits to become the default search engine.

[00:01:29] Eric Bleeker: So them being that initial point of distribution as the operating system for all consumer devices allows them to play a little bit of a different game. But it’s also kind of in their best interests. To kind of talk down some of the areas that they’re not competing in, which is a bit what this paper did.

[00:01:50] Eric Bleeker: It comes from, um, apple researchers and basically we’ve talked about along this show, our belief is that in December of last year, the birth of new powerful reasoning models, which really began when oh three was released on December 20th, was the next step change in AI demand. And what Apple’s researchers came out and said is that they believe that these reasoning models that are coming out through a series of puzzles that they tried having them solve, that they put in the paper, basically had just a lot less capability than people believed.

[00:02:28] Eric Bleeker: Essentially. They’re just much better at pattern matching, uh, than people give them credit for. So this Austin, this got a lot of attention, right? Because it’s potentially saying it’s, it’s another. A potential piece of evidence that we’re in a bubble, that AI is just going to hit a wall, where when you ask it to take that next step, the next step being discovering novel discoveries.

[00:02:55] Eric Bleeker: Mm-hmm. You know, asking AI to start doing things like finding cures for diseases. It’s not going to be able to make that jump. And if that is the case, that would be tremendously bad for that future of outlook and all the spending that’s happening right now. Right. So my take on this, and again, I am not an AI researcher.

[00:03:16] Eric Bleeker: My main job is to basically find who I believe are the smartest minds and, and people that have great track records on this. I. And evaluate who I think is the most accurate. And, and I don’t find the paper especially compelling. I, I won’t get into too many details, but a lot of people have poked some pretty sizable holes into how this study was conducted.

[00:03:39] Eric Bleeker: But I, I do wanna point out Austin, you know, of the other big storylines that’s, you know, maybe on the fringes that if, if you’re investing in ai you don’t see, was Sam Altman, who’s the CEO of Open AI. He released a blog this year, uh, or. This week, I should say, um, just about his thoughts on AI. It, it’s named the gentle Singularity.

[00:04:01] Eric Bleeker: You could search that, you could search his blog and you could find it, but in layman’s terms, he’s, he’s making a lot of predictions you can hold him accountable to and in his belief. 2025. The story is the rise of agents. That’s something we agree with. We’ve been saying since last year, you know, mid last year, we’ve always said we believe the narrative this year, especially in the back half, would be the rise of agents.

[00:04:27] Eric Bleeker: He believes 2026 is going to be the birth of novel insights from AI, and 2027 is robotics that can change the real world. So you know this, this lines up Austin, right? With a lot of storylines we’re talking about. But he is essentially dropping the gauntlet that, as I mentioned earlier, OpenAI is waiting for Stargate to become operational, uh, to basically train their next version, uh, GPT five, to have the capability to do that.

[00:04:59] Eric Bleeker: And once that’s happened, he believes those novel discoveries will start becoming a storyline next year. So that becomes one of the things that we will be actively tracking on this podcast that. If you start seeing novel discoveries from AI, we are on pace. And frankly, if AI starts making meaningful novel discoveries, that is the strongest validation yet that it becomes the main economic driver for the rest of our lives.

[00:05:30] Eric Bleeker: Mm-hmm. And if that is the case and, and you’re not actively investing in this, good luck to you because that, that will be a little bit of a rubber meets the road moment. We’re gonna be watching out for.

[00:05:42] Austin Smith: Right. And you know, to, to anybody who’s listening to that and saying, oh, well maybe I’ll wait until AI has proven in the, the novel discovery framework before I become an investor.

[00:05:51] Austin Smith: I think that would be a mistake as well, because that’s not to say that the current versions of AI and LLMs don’t have incredible economic forces and tailwinds behind them. You talked about medicine discovery, uh, famously alpha fold, you know, detected protein foldings and predicted 3D structures of proteins, which is a very complicated, uh, thing to do.

[00:06:10] Austin Smith: And it didn’t necessarily uncover new breakthroughs per se, but it completely supercharged the discovery of these, uh, folding structures that would’ve been discovered just over a much longer timeframe. So it pulled forward all of those discoveries faster. Now, that might not necessarily have been novel.

[00:06:26] Austin Smith: It was applying techniques that, as I understand it, certainly not an AI researcher, but was applying techniques that. We already had and knew, but just completely supercharged them and allowed us to make discoveries earlier. You know, when I read the, the, or tried to read, I should say, the illusions of Thinking paper by Apple, it, it sounded a little bit to me like a case.

[00:06:45] Austin Smith: Maybe it’s just a case of someone coming from sour grapes a little bit like I. You know, Apple’s always been a bit of a taste maker company and designing the thing you, you didn’t know you needed and then you can’t live without. And they’ve been uniquely good at that, perhaps the best of all time. Um, but technologically speaking, a reasoning model just gives up a lot of that to, to the model.

[00:07:02] Austin Smith: And it might be just culturally antithetical to, to what Apple is built on as a company. So, you know, I’m, I you gotta pay attention to everything that happens in this space. Apple’s incredibly gifted at what they do. Um, but this, this didn’t change my thinking about the potential of, of what we’re seeing.

[00:07:17] Eric Bleeker: And I do wanna double click on one point you made there that you know, you can always say, I will just wait for this. Next, threat or overhang to resolve before I invest. That’s why it’s so important to go through the cycles that as new trends happen that you’re able to see in real time.

[00:07:35] Eric Bleeker: There is always something new that says, there’s a reason I shouldn’t invest in this today. So if you’re waiting for one threat to go. There will be another. Nothing is an absolute slam dunk. Last week, Austin, we talked about the best time on a risk adjusted basis to buy Apple was six years after mobile started.

[00:07:52] Eric Bleeker: Mm-hmm. Um, the trough of disillusionment for the internet itself came eight years after it started. There’s always going to be some new threat that, you know, unless you followed one of these cycles, to truly appreciate that you’re never going to make money investing in a perfectly rosy picture. It’s just really hard to understand that that’s part of why we wanted to start this podcast was to give that record of investing throughout the a i Age.

[00:08:22] Eric Bleeker: ’cause at some point in retrospect, it’ll be, I believe it’ll be of course. Why didn’t you invest in that AI was obvious, right? And it’s not obvious. And we’ve broken down all the areas where we had to be incredibly. Counter to the dominant narratives, you know, you’ll be able to go back around When we did our deep seek episodes, where we called it before it happened, where we broke it down after, and we have been knocking down a lot of these, you know, bear cases.

[00:08:53] Eric Bleeker: It’s, it’s not because we are blind cheerleaders for AI. Far from it, you know, we are people who want to make money and the only thing that makes you money in the long run is what happens in reality. But you know, again, we, we have believed that this story’s happening and, and we have gone through many of these threats in the front half of the year to have a show today that is.

[00:09:19] Eric Bleeker: Largely just focused on acceleration and bullishness, which couldn’t be any different from what we were seeing five, six months ago.​

The post OpenAI CEO Sam Altman Predicts AI’s ‘Make or Break’ Moment is Almost Here appeared first on 24/7 Wall St..