The more I watch OpenAI, the more convinced I become that Sam Altman’s ego was the beginning of the company’s decline.
OpenAI did not become huge because Altman was some once-in-a-generation operator. It became huge because ChatGPT was a once-in-a-generation product. There is a difference. The company stumbled into one of the most important consumer tech moments since the iPhone, rode the sheer shock value of that innovation, and then somehow convinced itself that the person sitting on top of the rocket must have designed the laws of physics.
OpenAI’s first real advantage was novelty. ChatGPT felt magical. That gave OpenAI a massive head start, but when the novelty vanished and the rest of the market caught up, the company failed to prove itself not just as an innovation lab with a celebrity CEO.
Altman seems to want OpenAI to become Apple: a closed, prestigious, centralized, gatekept ecosystem where everyone builds inside his cathedral. Apps inside ChatGPT. Agents inside ChatGPT. Hardware.
ChatGPT is popular, but OpenAI does not own the phone. It does not own the operating system. It does not own the enterprise workflow. It does not own the cloud layer the way Microsoft, Amazon, or Google do. It does not even have a product moat that feels as unbreakable as people thought it was two years ago. The underlying model quality gap keeps narrowing. Switching costs are low. Developers and businesses will use whatever works, whatever is cheaper, and whatever integrates better.
That is why Anthropic looks much better run right now.
Anthropic is not pretending Claude is some holy object that needs an Apple-style walled garden around it. Their strategy feels much more Microsoft-like: accept that the core product may not be permanently magical, then build the boring, useful, sticky layers around it. Claude Code, enterprise integrations, developer tools, workflows, partnerships, APIs, reliability, business adoption. Not as sexy. Much smarter.
Anthropic’s venture capital money is obviously being burned too. This whole industry is basically setting money on fire to buy GPUs. But Anthropic’s burn feels more strategically allocated. Compute, yes. But also marketing, sales and developer adoption. Enterprise positioning. Product polish. Peripherals that make the model useful in actual workflows. They are not just trying to win the “my chatbot is smarter than your chatbot” contest. They are trying to become infrastructure.
OpenAI, meanwhile, is gatekeeping and guard railing the shit out of their models and for some reason just restricting them as much as possible.
He went from being one of the most respected figures in AI to becoming the face of a company that increasingly looks like it is being run aground by ambition without operational coherence. OpenAI’s original image was almost wholesome: brilliant researchers building something open source. Now it feels like a capitalist machine run by someone who does not fully understand capitalism beyond fundraising and valuation theater. Altman religiously narrowing his vision towards his AGI mission believing VC money won't dry down. Amodei also talks a lot about AGI but he understands profit matters.
That is the irony. Altman was chosen and celebrated largely because he came from the venture/startup world. He knew how to talk to capital. He knew how to sell a vision. He knew how to make investors believe the future was being negotiated in whatever room he happened to be standing in.
But being good at venture mythology is not the same as being good at running a giant operating company.
A VC can be rewarded for telling a compelling story before the business fundamentals exist. A CEO eventually has to make the fundamentals exist. OpenAI had the best possible starting position: the brand, the users, the developer mindshare, the press, the money, the talent, the cultural moment. And yet instead of consolidating that lead into a focused, profitable, durable company, it seems to have chased grandeur.
Anthropic seems to understand something OpenAI forgot: the winner may not be the company with the loudest AGI rhetoric. It may be the company that makes AI useful, embedded, and rational.
[link] [comments]