Is AI Headed for a Trough of Disillusionment?

0:01Today on the AI Breakdown, we're looking at the first day of market performance for tech stocks and wondering if it suggests that the AI hype hype is wearing off. Before that on the brief, OpenAI's ARR has gone up, John Roberts is talking about AI in his note, and the upcoming CES event is all about AI. The AI Breakdown is a daily podcast and video about the most important news and discussions in AI. Go to breakdown.network for more information about our Discord, our newsletter, and our YouTube channel. Welcome back to the AI Breakdown Brief, all the AI headline news you need in around five minutes.

0:34One of the big questions following Sam Altman's firing and then rehiring last year was how it would impact customers' relationships with OpenAI. In particular, would developers get nervous about building on OpenAI's APIs thinking that maybe something like that could happen again? Well, at first glance, it doesn't seem to have done much because according to the information sources, revenue grew about 20 % between October and December, up from a$1.3 billion annualized revenue rate to a$1.6 billion revenue rate. As the information puts it, this suggests that the company was able to hold on to its business momentum in selling AI to enterprises despite a leadership crisis that provided an opening for rivals to go after its customers.

1:18Now, these numbers are not confirmed. They are from a set of sources that the information believes has access to this information, but it still gives a pretty good idea of where things stand. We got a report last week that Anthropic has also upped its revenue projections for next year, suggesting that by the end of the year they will be at an$850 million annualized revenue rate. Now, on top of that$1.6 billion ARR, we also got reports recently that OpenAI was raising another round of funding at a north of$100 billion valuation, which, if that comes to completion shows just how much demand, at least from investors, there still is for artificial intelligence.

1:52Indeed, the amount that AI companies need to raise has dramatically shifted the nature of venture capital in some ways and created an opening for big tech companies who have much bigger war chests than even the biggest VCs to have preferential access to leading technology innovators in a way that they wouldn't have in previous tech cycles. Now, of course, in 2024, one of the things that we expect to see is even more competition for open AIs thrown at the top of the LLM heap, and one person who is pretty determined to be a contender in that race is Elon Musk. Musk's XAI has of course started to integrate the Grok chatbot into Twitter for their paying premium users, with the promised benefit being access to Twitter's real-time information stream and thus a chatbot that is theoretically much more up-to-date than something that isn't natively plugged into that information source.

2:38Well, one small but interesting piece of news, XAI has now been incorporated as a public benefit corporation in Nevada. This is effectively a corporate structure where a company involved pledges not only to maximize value for shareholders, but for the world at large. It's not a totally PR meaningless designation either. The company has taken on an obligation to, quote, create a material positive impact on society and the environment taken as a whole, and as part of their setup, requires XAI to have an outsider come in and look each year to see whether they're actually living up to those social and environmental obligations.

3:11It also has to publish an annual report outlining how it's positively impacted the public. Now, of course, this sort of structure does not come with as many governance weirdnesses as does, for example, OpenAI's setup, but it does suggest that Elon at least wants to compete from a perception perspective in that positive impact type of space. Moving on to our next story and speaking of competition, one of the areas that I anticipate to be a very brutal competition this year is around enterprise AI adoption. A little bit later in today's main episode, you'll hear me discuss the potential that we're going into a trough of disillusionment when it comes to adoption of AI, but I actually think that enterprises have been relatively slow to adopt and that we're likely to see pretty steady inclines driven in large part by bottoms-up adoption from individual employees.

3:55Well, one new contender in the enterprise AI adoption race is a new firm called Articulate AI that is spinning out of Intel. The work that would eventually become Articulate started as a project collaboration between Intel and the Boston Consulting Group, or BCG, where Intel used one of its own supercomputers to develop effectively its own generative AI system. Intel then customized and modified that system to run inside BCG's own data centers in a way that alleviated their concerns around security and data integrity. The new spin-out company has attracted investment from Digital Bridge Group and others, and will be led by the former vice president and general manager of Intel's data center and AI group.

4:31I think it'll be interesting to see whether this sort of structure offers a best-of-both-worlds approach for enterprises, where they feel like they get the nimbleness of a startup, but effectively with a big Intel-backed package. But we'll just have to wait and see. Now, one of the big themes from the last two weeks is AI showing up in every sort of end-of-year report or 2024 prediction report, and in the legal sector, it was no different. The Chief Justice of the United States Supreme Court, Judge John Roberts, extensively discussed artificial intelligence in his annual end-of-year note. Reuters described his tone in the 13-page report as ambivalent.

5:04Quote, He said AI had potential to increase access to justice for indigent litigants, revolutionize legal research, and assist courts in resolving cases more quickly and cheaply, but also pointed to privacy concerns and the current technology's inability to replicate human discretion. Wrote Roberts, I predict that human judges will be around for a while, but with equal confidence I predict that judicial work, particularly at the trial level, will be significantly affected by AI. Now, of course, AI and the legal system has been in the news frequently recently. There is, of course, the big pending case between the New York Times and OpenAI and Microsoft, where the New York Times has filed the most high-profile lawsuit yet, claiming that Microsoft and OpenAI violated its copyright by training their GPT models on old times material.

5:45But we also got reports recently that Michael Cohen, the former lawyer for Donald Trump, has said that he mistakenly gave his attorney fake case citations that were generated by AI that made their way into an official court filing. It's fair to say that we're going to see AI in the courts not only in the context of landmark cases like the New York Times suit, but also as just a tool that's more and more integrated into the legal process. Finally today, the first big tech event of the year is the Consumer Electronics Show, or CES, which happens every January in Las Vegas. To the surprise of no one, AI is expected to be at the very center of everything.

6:21Now, interestingly, CES is not just about computers and software in a traditional sense. It's really about devices of all types. This year, they expect more than 130 ,000 attendees and will feature more than 3 ,500 exhibitors. Investors Business Daily writes that AI will be in everything at CES 2024. They write, Consumer electronics firms such as LG Electronics, Samsung, and Sony will be on hand to unveil their latest audio and video products and other gadgets. These will include smart televisions, home appliances, home security gear, wearable devices, music speakers, and kitchen gadgets, and everything will be infused with AI or at least marketed as such.

6:56Now the thing that I'm watching most closely is this trend around so-called AI PCs. Market analysts have been discussing the potential that new computers designed for the AI age could spur what they're calling an upgrade cycle in the coming year. And whether consumers take hold of this trend or not, companies are certainly going to be pushing it. Intel, for example, will be holding an AI PC event on January 8th in combination with Dell, HP, Lenovo, and Microsoft. I will of course be keeping an eye out for any interesting announcements from CES, but for now, that will do it for the AI Breakdown Brief.

7:28Next up, the main AI breakdown. Welcome back to the AI Breakdown. One of the interesting phenomenons about artificial intelligence last year was the extent to which it propped up markets in spite of incredible other forces. One of the most dramatic moments that I remember last year of this was during the banking crisis in May, as the Fed was gearing up to put SVB in receivership and looking for other dominoes to fall like First Republic, and markets didn't react nearly as much as you would have thought. Another moment was during the debt ceiling standoff, when again, you would expect markets to be going down given the political turmoil, but the hype and excitement around artificial intelligence just kept things moving forward.

8:13Overall, the Nasdaq was up 43 % last year. That was its second best performance in 15 years. The tech and e-commerce sector of the S &P 500 were up an average of 57 % on the year, which was more than double the S &P 500's performance overall. Chips and software saw their biggest annual gain since the year after the global financial crisis. Indeed, as the Wall Street Journal puts it, excitement for generative artificial intelligence sparked by OpenAI's chatbot was the dominant theme for investors in 2023. However, as they put it, the new year might already be ushering in more sober perspectives.

8:47Yesterday, we saw companies including NVIDIA, Intel, AMD, Salesforce, Adobe, and ServiceNow all down by 1 % or more. The so-called Magnificent Seven tech companies averaged nearly a 2 % loss for the day, wiping out around$238 billion in combined market cap. Again, as the journal puts it, one thing already seems certain about 2024. AI is going to need to start showing the money. Whether it can is a whole other question. So basically the discussion in this piece, which is called Tech's AI Hangover May Just Be Getting Started, is around at what point Wall Street is going to expect AI to start driving actual financial gains, not just excitement and hype.

9:26For example, in spite of Microsoft being perceived as an early leader in the AI arms race, the journal writes, Microsoft already has a massive business throwing off more than$218 billion in annual revenue, against which AI will need to produce a lot of new growth to move the needle. In a survey of chief information officers last month, Brent Thrill of Jefferies noted that AI and machine learning are not major drivers behind why customers intend to increase cloud spend. Bernstein analysts noted in a December 19th report that, quote, CIOs are generally still in the exploration phase on AI. The piece also points out, though we've already started to get some examples of what an AI letdown might look like.

10:03Specifically, they point to Adobe, who had seen 85 % stock price gains ahead of their fourth quarter reporting in December, on the basis that their new Gen.AI tools, including Firefly, would increase demand significantly. However, when all was said and done, Adobe ended up projecting only 10 % revenue growth for the next fiscal year, which was effectively flat compared to the previous year's performance. Now, analysts estimate, in the wake of that report, Adobe's stock has fallen 7%, and analysts expect as more fourth quarter reports come in, this is going to be a common story. Writes Scotia Capital software analyst Patrick Coville, we worry AI benefits may materialize later than many expect.

10:39Alex Zucin of Wolf Research wrote, We are at peak AI hype with a likely slide into the trough of disillusionment, as actual Gen AI revenue dollars take longer to materialize, yielding at most low single-digit upside to CY24 revenue estimates. Now, if you've been paying attention closely, you might have seen a number of different versions of this sort of assessment that AI is headed towards a trough of disillusionment. The Next Web published a piece at the end of December called After a Year of Breathless Hype, AI Will Face Reality in 2024. Now, the trough of disillusionment is a reference to the Gardner hype cycle for emerging technologies.

11:13Gardner assesses generative AI at fairly near the top of the peak of inflated expectations. This is the point on the curve where people expect huge things, but those huge things have yet to materialize. What follows next is that trough of disillusionment, after which a more reasonable set of expectations takes hold, which Gardner calls the slope of enlightenment, finally reaching the plateau of productivity. At the heart of a lot of these assessments are flagging statistics around enterprise adoption. For example, the NextWeb points out recent research from Infosys, which found that only 6 % of European companies were currently producing business value with their generative AI use cases, and other statistics suggest this as well.

11:50Back at the beginning of December, NBC published a piece called There's a Gap Between AI Talk and Businesses Actually Using It. The central crux of the story was that although about half of the companies in the S &P 500 had talked about AI during an earnings call since last May, a November survey from the Census Bureau found only 4.4 % of businesses nationwide using AI to produce goods or services. What's more, only 6.9 % said that they think they will use AI in the next six months. Now, when it comes to the explanation for why there's such a gap, many point to a lack of resources specifically around education and talent.

12:24Said Christina McElleran, an assistant professor at the University of Toronto, when you get some exciting new technological change, it does not magically appear in the economy. Instead, writes NBC, McElleran believes that it is education, access to talented employees, and money that help companies make use of new technologies. Now, of course, there are other reasons why enterprise adoption has been slower than it might have seemed as well. There are, of course, still major questions of hallucinations and false information. There are concerns around security and the integrity of data in publicly available systems like ChatGPT, even in the enterprise versions, which promise to not incorporate companies' private data into their training models, but still feel a little bit like entering information into a black box.

13:06Others I've seen have noted that a big part of the value of artificial intelligence has to do with enabling better insights from companies' data, but many companies don't even have the sort of data practices necessary to input all of that information for AI to actually generate those insights. There is also, I believe, probably a bit of a gap between formal enterprise adoption, i.e. a business making a specific decision to implement some new AI-powered system, and individual employees starting to incorporate AI tools into their workflows on a one-by-one sort of basis. I would venture a guess that there is far more of that activity and that indeed, the bottoms-up usage of AI inside companies by employees who are just finding ways for it to increase their own productivity is the way in which AI will tend to infiltrate the mainstream enterprise sector.

13:51I also think it's likely the case that enterprises are being a little bit more deliberate with their adoption of new technology following previous hype cycle waves, such as crypto and NFTs. To me, that suggests that whatever trough of disillusionment we go through may be less pronounced than in previous cycles because the adoption cycle hasn't actually started. In other words, we won't necessarily see the same sort of adoption then abandonment that we've seen in other cases, but just the slow, steady accretion of adoption over time. Now, I wouldn't be surprised to see a lot more of this discussion in mainstream news outlets in the weeks to come because it's a very useful counter-narrative and a nice shift from what we experienced in 2023.

14:32And ultimately, narrative creation is a lot of what media does. but my bet is still firmly in the camp of massive transformation of business across basically every knowledge sector and it happening in a pretty timely fashion. But of course, we will have lots of chances to see how this all plays out and I will be happy to share all the evidence as it happens with you all. For now, that is going to do it for the AI Breakdown. Until next time, peace.

15:05Thank you.