GPT-5 Faces Ridicule as Altman Plans 'Trillions' for AI Infrastructure
OpenAI’s latest large language model, GPT-5, has been met with significant public skepticism since its release, with many users branding it as “dumb,” “boring,” and even inferior to its predecessor. Amidst this wave of criticism, OpenAI CEO Sam Altman has offered a candid assessment of the situation, acknowledging a “bungled” rollout while simultaneously forecasting an unprecedented level of investment in the company’s AI infrastructure – an expenditure he estimates will reach trillions of dollars.
Speaking at a recent dinner with journalists and OpenAI executives in San Francisco, Altman admitted to missteps in the GPT-5 launch. Despite the public backlash, he pointed to strong internal metrics as a sign of underlying success and future potential. “We totally screwed up some things on the rollout,” Altman conceded, yet quickly added, “our API traffic doubled in 48 hours and is growing. We’re out of GPUs. ChatGPT has been hitting a new high of users every day. A lot of users really do love the model switcher.” He emphasized that the experience served as a crucial lesson in upgrading a product for hundreds of millions of people simultaneously.
Altman also weighed in on the broader discourse surrounding the artificial intelligence industry, aligning himself with critics who describe it as a “bubble” reminiscent of the early internet boom. “Are we in a phase where investors as a whole are overexcited about AI?” Altman questioned, then affirmed, “My opinion is yes.” This sentiment resonates with a growing number of commentators who have speculated about a potential collapse in AI enthusiasm, concerns amplified by recent market fluctuations, such as a challenging stock day for datacenter and AI infrastructure startup Coreweave. Many observers note that, thus far, AI has largely functioned as a financial sinkhole, with companies pouring vast sums into the sector with the hope of future profitability.
Drawing parallels to historical economic phenomena, Altman explained his view on bubbles: “When bubbles happen, smart people get overexcited about a kernel of truth.” He elaborated, “If you look at most of the bubbles in history, like the tech bubble, there was a real thing. Tech was really important. The internet was a really big deal. People got overexcited.”
Despite his acknowledgment of an overheated market, Altman remains resolute in OpenAI’s commitment to monumental infrastructure investment. “You should expect OpenAI to spend trillions of dollars on data center construction in the not very distant future,” he told reporters, signaling a scale of capital deployment that dwarfs most corporate spending plans.
This truly staggering financial commitment prompts a crucial, yet often unasked, question: What is the ultimate societal cost-benefit of such an endeavor? Few public discussions with Altman seem to delve into whether a comprehensive societal analysis has been conducted to justify the immense resources being funneled into the AI industry. Is it truly worthwhile to invest trillions of dollars into developing sophisticated AI models, particularly when their current output, as some users contend, amounts to “mildly amusing chatbots” that provide accurate information only intermittently?
Alternative uses for such colossal sums naturally come to mind. Could these trillions be more effectively deployed to address pressing global issues like poverty alleviation or the improvement of educational systems? Furthermore, the fundamental utility of AI models, particularly chatbots, remains a subject of debate: Are they a societal necessity, or merely a “nice to have” convenience? How much more beneficial are they compared to existing search engines, and can’t we simply rely on the latter? These questions extend to the potential negative externalities associated with widespread AI adoption, including a massive energy footprint, alleged reductions in users’ mental capacities, and a documented surge in academic cheating. Do these drawbacks truly outweigh the perceived benefits, such as a slightly more convenient way to access online information? While these inquiries may seem self-evident, it remains unclear whether they are being adequately addressed in the high-level conversations shaping the future of artificial intelligence.