As the stock market reels in the midst of a midsummer correction, analysts continue to warn about air leaking out of the generative AI balloon as well — chatbots and all:

For example, at Wired, Will Knight told us last May,

A week after its algorithms advised people to eat rocks and put glue on pizza, Google admitted Thursday that it needed to make adjustments to its bold new generative AI search feature. The episode highlights the risks of Google’s aggressive drive to commercialize generative AI—and also the treacherous and fundamental limitations of that technology…

“You can get a quick snappy prototype now fairly quickly with an LLM, but to actually make it so that it doesn’t tell you to eat rocks takes a lot of work,” says Richard Socher, who made key contributions to AI for language as a researcher and, in late 2021, launched an AI-centric search engine called You.com.

Yes. Space Odyssey 2001‘s HAL 9000 (c. 1968) would not have told you to eat rocks. The industry is slowly coming to terms with the fact that these types of built-in limitations may be a feature, not a bug. Daniel Griffin, a search consultant and researcher, told Knight, “that it’s important to remember no one has solved the problem of LLMs failing to grasp what is true, or their tendency to fabricate information.”

So if you would never tell anyone to eat rocks, your desk job is probably safe from this generation of AI.

Signs of retrenchment

Overall, there are signs of retrenchment. Meta (Facebook’s and Instagram’s parent) just dropped a big celebrity chatbot project that fell flat with users. Apparently, not very many people wanted to converse with the alter ego versions of celebrities. Instead, Meta is rolling out AI Studio where users can make AI versions of themselves, with the ability to chat with others online:

With AI Studio, Instagram creators can set up an AI as an extension of themselves that can quickly answer common DM questions and story replies. Whether it’s sharing facts about themselves or linking to their favorite brands and past videos, creator AIs can help creators reach more people and fans get responses faster.

Meta, July 29, 2024

It will be interesting to see how many Meta product users want to interact with the chatbots of others as opposed to those who want others to interact with their chatbots.

Another AI winter looming for the tech sector as a whole?

From Reuters, we learn that even mighty Microsoft and Google beginning to worry investors:

Fears that runaway spending by tech giants on data centers would yield little payoff in the short term have dogged the U.S. stock market this month amid signs that Wall Street may have become too optimistic about earnings growth.

Shares of Google-parent Alphabet (GOOGL.O), opens new tab fell more than 5% last week after the company reported a quarterly capital spending that exceeded estimates by nearly $1 billion, while the revenue boost from AI integrations remained modest, sparking a selloff in major tech companies.

Microsoft’s costs in focus as fears rise over slow payoff from AI, July 29, 2024

The smoke is still clearing …

“Will the stock market crash in August?” asked Harry Jones at The Motley Fool on July 31.

Good question. On Monday, August 5, “Wall Street suffers worst day in nearly two years after global sell-off.”

The technology-focused Nasdaq was also dented by Apple, which came under pressure after it emerged Warren Buffett’s Berkshire Hathaway had scaled back its vast stake in the tech giant.

Apple is the world’s largest public company, with a market value of more than $3trn, and its stock’s performance has a significant effect on the wider market. Stocks tied to the boom around artificial intelligence, including Nvidia, have also struggled in recent days. The collective market value of the so-called Magnificent Seven band of tech giants – which also includes Alphabet, owner of Google and YouTube; Amazon; Microsoft; Meta Platforms, owner of Facebook, Instagram and WhatsApp; and Tesla – was set to fall by about $800bn on Monday, according to LSEG data.

Callum Jones, The Guardian, August 5, 2024

Keeping the faith

Some think that continued progress toward human-like thinking is inevitable. Dean W. Ball, writing at the New Atlantis (Summer 2024) thinks that the decline of generative AI points to a more glorious future in which “it may become more appropriate to think of future language models as the product of several AIs reasoning together and conversing among themselves, with the entire written corpus of human knowledge — our tweets and blogs, our poetry and prose, our wisdom and folly — as the foundation.

Really? One can only speculate what they might have to tell us then.

You may also wish to read: Why the artificial intelligence bubble may be about to burst. For one thing, Space Odyssey’s HAL and Prometheus’s David aren’t really happening, for good reasons. Gary Marcus argues that Big Tech got rich, largely on promises, without confronting the known limits of artificial intelligence.

and

Is Big Tech’s AI starting to run out of other people’s money? People are beginning to wonder if all the AI hype is really going to pay off in substantial improvements. Two things we can be sure of: Things that can’t go on forever won’t. And even artificial intelligence can’t make hype into reality.