There is growing concern in Silicon Valley that progress around artificial intelligence is slowly “running out of steam.” CNBC’s Deirdre Bosa explores whether progress in AI is slowing and what that means for the industry. The first signs of struggle are bubbling among the big names in technology. As Bosa points out, the first sign of a slowdown was the lack of progression between models. OpenAI has led the pack in AI advancements, with each of ChatGPT’s new models exponentially better than the last.
Unfortunately, the same cannot be said today. OpenAI’s highly anticipated model, Orion, was intended to be a revolutionary system bringing us closer to AGI. However, the initial vision is being revised downwards.
“Even though Orion’s performance eventually exceeded that of previous models, the increase in quality was much less compared to the jump between GPT-3 and GPT-4, the last two flagship models released by the company, according to some OpenAI employees who have used or tested Orion.
The same obstacles have occurred with other AI models, like Anthropic’s. Just like OpenAI, Anthropic is also witnessing a schedule slip for the release of its long-awaited Claude model called 3.5 Opus.
“The AGI bubble is bursting a bit. It’s become clear that ‘different training approaches’ might be needed to make AI models really perform well on a variety of tasks.”
Even some of the big tech giants aren’t excited about the advances AI is making. Speaking at the New York Times’ annual DealBook Summit at Jazz at Lincoln Center, Sundar Pichai said generative AI won’t change lives in 2025, at least not any more than it already has.
“I think progress is going to get more and more difficult. When I look at (2025), the lowest hanging fruit is gone. The slope is steeper… We will definitely need deeper breakthroughs to move to the next stage.”
However, if progress stagnates, it is because of the laws of scaling. Anthropic CEO Dario Amodei says scaling laws are a misnomer. He points out that while these models have historically guided AI development – suggesting that increased data and computing power leads to improved AI capabilities – it is not guaranteed that ‘they will continue indefinitely.
This is why AI companies are turning to synthetic AI, i.e. artificially generated data. The Information reports that Orion was produced based on synthetic data. However, the problem is that poor quality data in turn leads to poor quality performance. So as major tech companies prepare to unveil new models over the next eighteen months, their pace of progress – or lack thereof – has the potential to radically redefine competitive dynamics.
For this article, we selected AI stocks by looking through news articles, stock analyses, and press releases. These stocks are also popular among hedge funds.
Why are we interested in stocks that hedge funds are piling into? The reason is simple: our research has shown that we can outperform the market by imitating the stocks selected by the best hedge funds. Our quarterly newsletter strategy selects 14 small- and large-cap stocks each quarter and has returned 275% since May 2014, beating its benchmark by 150 percentage points (see more details here).
A technician showing a projection of the company’s geolocation software.
Number of hedge fund holders: 14
Fastly, Inc. (NYSE: FSLY) is a leading cloud platform provider offering cloud computing, image optimization, security, edge computing and streaming solutions. On December 16, the company announced the general availability of Fastly AI Accelerator, a semantic caching solution that enables developers to optimize their LLM generative AI applications. Delivering response times on average 9x faster, the AI Accelerator was initially released in beta with support for OpenAI’s ChatGPT and is now also available with Microsoft Azure AI Foundry. Developers often only need a single line of code to update their application to a new API endpoint. This reduces the need to repeatedly call the AI provider because Fastly’s Edge cloud platform caches and responds to repeated requests. As a result, developers benefit from more efficient AI performance, lower costs, and an overall better experience.
“AI helps developers create many new experiences, but too often at the expense of performance for end users. Too often, today’s AI platforms keep users waiting. With Fastly AI Accelerator , we’re already averaging 9x faster response times and we’ve only just started.1 We want everyone to join us in our quest to make AI faster and more efficient.”
Overall FSLY ranks 8th on our list of AI stocks that dominate Wall Street. While we recognize the potential of FSLY as an investment, our conviction lies in the belief that certain AI stocks hold more promise in terms of higher returns and in a shorter time frame. If you’re looking for an AI stock that’s more promising than FSLY but is trading at less than 5x earnings, check out our report on cheapest AI stock.
