It’s been less than two years since OpenAI changed the world by launching ChatGPT, but there are already signs that the technology could be hitting a ceiling.
OpenAI’s latest model, Orion, is designed to replace GPT and go a significant step further, but the model has not met the company’s performance goals. While it’s an improvement over OpenAI’s GPT models, it’s not the leap the company had hoped for, and evidence is now mounting that artificial general intelligence (AGI) may be further along than technologists like OpenAI- CEO Sam Altman had hoped.
Start your morning smarter! Wake up with Breakfast news in your inbox every market day. Register for free »
After all, OpenAI is not the only AI start-up facing such challenges. According to Bloomberg, the latest version of Alphabet‘s Gemini is falling short of internal expectations, and Anthropic, seen as the AI startup most challenging OpenAI, is behind on the release of its updated Claude chatbot model called 3.5 Opus.
Image source: Getty Images.
The biggest reason why these models seem to be reaching a ceiling is that they have difficulty finding new sources of substantial training data, as previous models have exhausted sources like Wikipedia, social media, and news sites. Margaret Mitchell, the chief ethics scientist at AI startup Hugging Face, told Bloomberg about the technology challenges: “The AGI bubble is bursting a little bit.”
In other words, until the problem of securing reliable training datasets is addressed, the expected performance of advanced AI models will likely disappear, at least in the short term.
It’s unclear how big this slowdown is at this point, but at a time when other industry insiders have been calling an AI bubble, the news could lead to inflated stock valuations in the tech sector.
As the law of diminishing returns appears to be hitting major language models (LLMs), the AI sector could take a hit Nvidia (NASDAQ: NVDA) seems to be at greatest risk here.
After all, Nvidia’s graphics processing units (GPUs) are used to train AI models like ChatGPT, and demand for those components has skyrocketed since ChatGPT’s launch. Cloud infrastructure companies, autonomous vehicle companies such as Teslaand AI startups have been stocking up on Nvidia chips in anticipation of an AI boom.
However, there is still no ‘killer app’ in generative AI, and the criticism of the technology seems to be that it is impressive and capable, but the use cases are not entirely clear, especially if it is still prone to errors.
Some Wall Street analysts have expressed skepticism about the billions spent by companies like Alphabet Microsoft The capital expenditures will pay off as end-user spending on generative AI still appears disappointing.
David Cahn, general partner at venture capital firm Sequoia Capital, reported the inflating AI bubble in June and said revenue expectations from the AI expansion are on track to reach $600 billion by the end of this year. while OpenAI, the industry leader, is on track to reach just $3.7 billion in revenue this year. It targets $11.6 billion by 2025.
Nvidia shares have shrugged off the news so far, indicating investors don’t see it as a threat, and AI stock is now trading near record highs. The slowdown in profits in LLMs does not mean this is the end of technological progress. OpenAI’s Orion model is currently undergoing post-training, a routine procedure designed to tweak and adjust the model’s tone before it is released to the public, which is expected to happen early next year.
Besides LLMs, there are other ways to advance AI, although this has been the preferred approach of major tech companies in the two years since ChatGPT was launched.
Nvidia will report third-quarter results on Wednesday, and we could hear some questions about the challenges with scaling AI models and the potential implications for Nvidia.
Analysts expect Nvidia to report another huge set of results, with the consensus calling for revenue growth of 82% to $32.9 billion. For now, the company’s rising growth doesn’t appear to be in trouble, but if AGI is further away than investors expected, the stock will likely feel the impact at some point.
If our analyst team has a stock tip, it could be worth listening to. After all, Stock Advisors the total average return is 875% – a market-shattering outperformance compared to 172% for the S&P 500.*
They just revealed what they think are the 10 best stocks for investors to buy now…
View the 10 stocks »
*Stock Advisor returns November 11, 2024
Suzanne Frey, a director at Alphabet, is a member of The Motley Fool’s board of directors. Jeremy Bowman has no position in any of the stocks mentioned. The Motley Fool holds positions in and recommends Alphabet, Microsoft, Nvidia and Tesla. The Motley Fool recommends the following options: long January 2026 $395 calls to Microsoft and short January 2026 $405 calls to Microsoft. The Motley Fool has a disclosure policy.
OpenAI has a warning for Nvidia. Is the AI bubble bursting? was originally published by The Motley Fool
