2024 is a year of reckoning for AI

The AI ​​marketing hype, likely sparked by OpenAI’s ChatGPT, has reached a fever pitch, with investors and executives having stratospheric expectations for the technology. But the higher the expectations, the easier it is to disappoint. 2024 will be a year of reckoning for AI, as business leaders take a closer look at what AI can do Actually do it right now.

The Gartner hype cycle suggests that after the peak of inflated expectations comes the valley of disillusionment. Take the most recent earnings of Microsoft and Alphabet, Google’s parent company. Alphabet reported a strong quarter: revenue of $86.2 billion, an increase of more than 13 percent compared to the previous year. In response, shares fell by more than 7 percent the next day. Microsoft similarly reported that its net income rose 33 percent – ​​and its shares also fell, albeit only 2 percent.

Investors did not expect anything good. They were pregnant perfect. The technology sector is on the run, with the so-called Magnificent Seven – Nvidia, Meta, Amazon, Tesla, Apple, Microsoft and Alphabet – far ahead of the market last year. Investments in AI have not yielded as much as the market had hoped.

Business expectations are the natural result of the breathtaking AI hype of the past two years, with Wall Street analysts promising that AI is nothing short of “a generational transformation” that is “analogous to the advent of electricity.”

“We got very frothy very quickly,” said Mark Shmulik, an analyst at Bernstein. Some investors and analysts have compared AI to the early internet, which puts a lot of weight on the shoulders of AI companies. “But we’ve already been through a few of these cycles, with crypto and the metaverse, that we quickly discarded.”

“We got very frothy very quickly.”

There doesn’t seem to be much doubt that there is a need for AI tools in the business world. “The interest in virtually every sector is universal,” says Bret Greenstein of PwC. “However, it is still early stages of adoption for everyone.”

Our Business Leaders expect returns on their AI investments by 2024, with 61 percent of CEOs surveyed expecting generative AI to improve the quality of their products and services, according to PwC’s annual survey. The PwC report warns of a ‘moment of truth’ for trust in AI, where mistakes could have major consequences; This quarter’s earnings figures suggest that the failure to deliver on promised productivity gains could be the most incendiary of all.

The first place where observers like Greenstein believe mass adoption can happen is in IT – especially around software development, customer service and back-office tasks such as processing incoming forms. That makes Microsoft’s Copilot, the “AI companion”, an indicator for the entire field. A soft measure of Microsoft’s dominance is that, just as “Googling” quickly became synonymous with Internet searching, we’re talking about Copilot and not offerings from other giants like Google, Salesforce or Oracle, says Adam Niewinski, co-founder and general manager of Microsoft. partner at OTB Ventures.

“Microsoft is ahead of Google and actually ahead of everyone else,” says Niewinski. Their partnership with OpenAI — more on that later — allows them to source development in ways that may be more important than just a simple return on investment, he told me. “Everyone knows ChatGPT and OpenAI and everyone knows Copilot. But it took me a few seconds to remind myself that Google has Bard, because no one actually uses this name. (Bard has since been renamed Gemini, because nothing is as reliable as the astrological sign most likely to guide you “away from your own life.”)

Microsoft’s leadership creates a positive feedback loop, where more users mean more training, which means better data, which means more users. So seeing how Copilot is deployed could be indicative of how AI is implemented, says PitchBook’s Brendan Burke. “That’s an easy-to-use application that developers test to see if it can meaningfully help them, or distract them,” Burke says.

Despite all the hype surrounding the inclusion of AI in the Bing search engine, Microsoft’s market share grew by less than half a percent

Whether Copilot is actually worth its price remains an open question. To be sure, AI hasn’t helped Microsoft in search: despite all the hype surrounding AI’s inclusion in the Bing search engine, its market share grew by less than half a percent. Microsoft also seems to have looked for a good use of Copilot; the brand started at GitHub and moved to retail products and then to Microsoft Office apps. Even Microsoft’s Super Bowl ad is vague: If a great use case exists, the company hasn’t found it yet.

AI is expensive. Take OpenAI for example; by December 2023, the annual run rate was $2 billion. Since that’s a figure that takes the previous month’s revenue and then multiplies it by 12, we know that means OpenAI made about $167 million that month. The company is nonetheless operating at a loss and will likely need to raise “tens of billions more” to continue operating Financial times reported. Sam Altman, CEO of OpenAI, was looking trillions of dollars of investment to completely transform the chip industry. Meanwhile, ChatGPT’s growth has stalled.

During the era of zero interest rates, big tech could pour endless money into its pet projects — CEO Mark Zuckerberg’s little venture into the metaverse has raised at least $46.5 billion since 2019. Fortune reported last October. If we were still in that era, a company like Google might just put money into AI. “I don’t think Google can burn money to its heart’s content with these initiatives,” Shmulik said. “We are going through a period where investors are increasingly concerned about profitability.” That doesn’t seem like great news to me for Gemini, which so far seems to be a slightly better version of Google Assistant and is lagging behind, according to ChatGPT. Tom’s guide.

“I don’t think Google can set money on fire with these initiatives to its heart’s content.”

Even Google is cutting costs (12,000 jobs last year and more in 2024), and the bar for funding will likely be higher in 2024. The challenge for AI startups now is to create sustainable business models and bring AI to areas where that is not yet the case. yet disturbed, Burke told me. And the high valuations, relative to revenue, assigned to these companies suggest that venture capital funds expect them to become technology giants in the long term.

VCs have tightened their wallets and even the AI ​​sector has been hit, according to a note from PitchBook’s Burke. Excluding deals from Microsoft and Amazon, which aren’t exactly traditional VC investments, there was only $7.9 billion committed to this space by VC investments as of last November, which meant about the same level of spending as in 2021.

There may also be fewer exits for venture capital funds investing in the AI ​​sector. As rumblings about tech antitrust grow louder, there are concerns about whether acquisitions can go ahead. After all, Adobe’s attempt to acquire Figma did not. “Many already believe that these mega-cap companies have way too much power, and AI is clearly a sensitive issue right now,” said Angelo Zino, a technology analyst at CFRA Research. “That’s why you haven’t seen a direct acquisition yet.”

Instead, the spike in AI fundraising in 2023 was the result of big companies dumping money into the space. “We won’t necessarily see the size of a deal between Microsoft and OpenAI or Amazon and Anthropic in the next year, just because those companies have established their leadership in this space,” Burke says.

However, AI companies have a compelling pitch for anyone who wants to use their services. For example, ServiceNow CEO Bill McDermott has said he can make engineers 50 percent more productive. That’s a huge cost savings, says Scott Kessler, head of technology at Third Bridge. Theoretically, AI should be in high demand.

But actual adoption and real gains may not yet be realized. So far, it’s not clear whether Microsoft’s Copilot is useful enough to justify its cost The Wall Street Journal. The recording was slower than expected; Microsoft had to drop the original requirement that companies have a minimum of 300 subscriptions. There also appears to be a decline in use a month after an initial increase. But don’t worry, Microsoft has a plan to address this: it’s going to issue more warnings.

Right now, most AI companies generate money through premium pricing for better services, says PwC’s Greenstein. But that could potentially evolve into results-based pricing, he says. “The idea of ​​paying based on the outcome – you know, ‘I saved your money, I get some, you get some, everyone wins’ – it’s a very interesting model in AI, because there are a lot of things [where] you can guarantee an outcome,” he says. “And I think these will be very attractive commercialization models.”

This can help prevent disappointments. Because if AI tools still require human judgment, they may not be faster or better than just letting someone do it. Greenstein told me on the phone enthusiastically about using AI to summarize information, but he and I have very different experiences with that. Maybe AI works well for summarizing forms, assuming it doesn’t make mistakes.

But I tried using AI to summarize podcasts I don’t really want to listen to, and then listened to them to compare how well the AI ​​worked. I find that while the AI ​​summaries are technically correct, neither Gemini nor ChatGPT have enough context (or social skills?) to pick up subtle digs at rivals, subtext, innuendo, or other important parts of human communication. Anything that required background knowledge was beyond the capabilities of the AI. Using AI actually added time to my job: the time it took to request an AI summary and read it.

Even OpenAI is trying to push back on the hype. In December, OpenAI CEO Brad Lightcap told CNBC that he continues to have to explain to people that AI cannot dramatically reduce costs or bring back growth for struggling companies. Morgan Stanley’s AI chatbot is being bypassed by asset managers because people want to talk to other people. The information reported. News operations that tried to replace journalists with AI-written articles faced backlash because those articles were wrong, offensive or useless.

So 2024 seems to be the place where the rubber meets the road. If there are real use cases for large language models, models that save companies money, AI may be on its way to sustainability. But if these tools become widely used and lead to bad publicity, lawsuits, and congressional hearings, with minimal productivity gains, the bottom of disillusionment can come—and it can be very deep.

Leave a Reply

Your email address will not be published. Required fields are marked *