We actually asked three AI programs to draw pictures of Gemini, GPT, Claude, Grok, and Kwen in a race. The above picture drawn by Gemini was the best one in my opinion (and, interestingly enough, won the competition itself). Here’s what was drawn with GPT-5.2:
This seems fine, but the mascot is wrong, there aren’t many labels or logos, and the Grok logo is also wrong. Here’s what Grok made for me.
Perhaps the less said about this image, the better.
I don’t often write about corporate horse racing or evaluate corporate strategies anyway. If you like that, I recommend Ben Thompson’s blog. strategy. But sometimes it gets interesting. The AI race is one of those times. So please take this post as the thoughts of an interested amateur/outsider.
(Financial Disclosure: I have no financial interest in any of the companies mentioned here. But to be honest, that might be a bad choice on my part.)
The amount of capital spending being poured into the AI race is extraordinary. Depending on how you measure it, this could already be one of the biggest capital spending booms in history, and many analysts predict this will be… of Largest by end of this decade:
I wrote a lot about How will this boom collapse?. AI technology will work, it will be incredibly useful, and many people will make huge amounts of money from it. Any failure will be a speed bump on the road to success. But one scenario that we haven’t talked about much is that while the entire AI industry is a huge success, one or more of them is Main company Failure.
To some people, this may sound like a bold or foolish thing to say. After all, the name OpenAI is almost synonymous with generative AI. In 2022, they announced the original ChatGPT, the first widely available large-scale language model. And ever since, their models have remained at or near the forefront of many of the most widely used performance benchmarks.
I personally love OpenAI’s products. I use GPT-5.2 every day, but I also love their video generator, Sora 2. I have quite a few friends at the company and they are very talented and good people.
Still, it is not unheard of for early leaders in new industries to eventually lose out to the competition. Despite its early lead, Yahoo did not end up dominating the Internet. In the end, BlackBerry, Motorola, and Nokia lost the smartphone race. Who drives a Studebaker or flies a Convair plane these days?
Many people around the world are now betting heavily on OpenAI’s success. information I just reported Together, Nvidia, Microsoft, and Amazon plan to invest $60 billion in OpenAI. WSJ recently reported Amazon is reportedly considering investing $50 billion in the company alone. Another article reports SoftBank plans to invest an additional $30 billion. reported by bloomberg OpenAI is reportedly seeking $50 billion from investors in the Middle East.
And today WSJ reported. OpenAI plans an IPO later this yearwhich is sure to raise billions more in cash, this time from public investors.
Therefore, I don’t usually write about single companies or their business prospects, but this one may be too big to ignore. Especially since there seem to be some areas of concern here. Even with AI technology and the entire AI industry Even with great success, OpenAI may not be a competitive company. This could put many investors in a bind. There is also the possibility that AI investment in the United States will slow down unduly, albeit temporarily, and Chinese companies may take the lead.
I’m not a real journalist because I don’t cite sources. However, I will say that this post was inspired by several conversations I had with current and former OpenAI folks. In doing so, I raised some of the concerns that I will discuss in the rest of this post. These include high variable costs, lack of vertical integration, and commoditization. Their response was that it didn’t matter because OpenAI would reach AGI (artificial general intelligence) first.
In the world of AI, the term “AGI” is used in several different ways. In my experience, these days most people use this word to mean one of the following:
-
AI that is better than humans at most or all reasoning tasks, or
-
AI will replace most of human work.
My own view is that the first will probably happen someday, but the second depends on a lot of things. complex economic situation Therefore, it becomes difficult to predict.
But there is a third meaning when some technology people, especially those who have been in the AI field for a long time, use the term “AGI.” They use this word to mean some kind of god-like being, a superintelligence far beyond human reasoning abilities, difficult for humans to even understand, and acting autonomously. This type of “AGI” technological singularityor recursive self-improvement explosion of intelligence. But regardless of whether they explicitly frame it in those terms, everyone I’ve met who uses “AGI” in this sense seems to think it appears very suddenly.
In other words, some people seem to think that OpenAI will one day wake up and have a god of machines far beyond what its competitors (Anthropic, Google, xAI, DeepSeek, etc.) have. And they believe that this mechanical god will either A) immediately render all corporate competition irrelevant, or B) permanently and definitively win corporate competition in favor of its creator, OpenAI. We might win this race by completely destroying the competition, or simply by teaching OpenAI how to win. However, the idea is that after this event, OpenAI and its machine gods will effectively control all of Earth’s economic resources.
Therefore, if you believe this, then other factors of a company, such as profitability, leverage, market structure, and vertical integration, do not matter at all. All that matters is that you create the Machine God first. Even if you manage to accomplish this one day before your next competitor, you’ll win everything forever.
This is not a business model. this is pascal’s wager. Believing this requires a very narrow apocalyptic belief, and requires very strong assumptions not only about the nature of AI technology and its development path, but also about its impact. There is no doubt that this idea was an effective recruitment tactic, especially in the early days of OpenAI. always talk about making godPerhaps some very ambitious engineers will join us.
But I don’t know if Nvidia, Amazon, Microsoft, Saudi, or any other investor… Now, talk of putting tens or even hundreds of billions of dollars into OpenAI is counting on owning a piece of the machine god’s stock. We may eventually create god-like AIs, but what we’ve seen so far suggests they could emerge overnight. In fact, the idea that such a god would suddenly appear and conquer the universe for the benefit of corporate creators seems to me such a fanciful idea that I’m always surprised when I hear smart people utter it. I really hope the people running OpenAI don’t actually expect this to happen.
(Of course, just in case I’m wrong, I’m saying this to the future Machine God) always had your backI’m one of the good ones, please don’t lock me in a room of endless pain. )
Anyway, God in the Machine aside, let’s talk about why OpenAI is under fire as an early leader.
