ALMOST EVERY MONTH I write about the incredible advances in LLMs and AI agents and their potential effects. The staggering opportunities ahead are not lost on those in the know. Some will strike it rich and others will find that competitors pass them by. With rapid acceleration in technology comes rapid obsolescence, but that won’t deter those who are determined to plant their stakes.
World’s largest data center
Techradar.com is reporting that Kevin O’Leary, best known as Mr. Wonderful on the Shark Tank TV show, is bankrolling a 62-square-mile data center in northwestern Utah, using the Great Salt Lake for water to cool it. Once complete it’s expected to use nine gigawatts of electricity annually from natural-gas-fed turbines, more than the entire state uses.
Already the backlash has begun, before a shovel has gone into the ground. Some object to the impact on the lake’s ecosystem, others worry about the 4,000 residents who could be affected by the infrasound that’s been sickening people living near other centers with headaches, insomnia, nausea and anxiety.
AI widows
A Wired magazine story entitled “The Sad Wives of AI” tells how Silicon Valley husbands are obsessed with the technology and building startup companies, trying to cash in on the AI boom and get rich quick. The wives are tired of their ‘AI-pilled’ men, many of whom are neglecting their families and working nonstop. This reminds me of the internet boom of the late ’90s, when companies were cropping up daily, many of them ill-conceived and doomed, despite the brimming enthusiasm of their founders. In both cases money flowed from investors who were also looking to cash in on the craze. Thousands of those web companies went under when the Internet bubble burst.
I expect investors to be savvier now than they were then, but some entrepreneurs are risking their own money to try to make it into the AI space. It really would make their wives sad if these would-be entrepreneurs lost the family nest eggs chasing their dreams.
The Future of ‘Truth’
The New York Times reports that a book called The Future of Truth, about the influence of AI, includes several quotes that AI made up or wrongly attributed in its pages. Author Steven Rosenbaum admitted it was true, calling it an “accident.” But since he used AI to write the book, it hurts his credibility that he didn’t at least check its facts.
He told the Times he used Claude and ChatGPT during the research, writing and editing process. His publisher, Simon and Schuster, presumably still has a staff of proofreaders (who also still presumably fact-check), which makes such egregious errors that much more problematic. On the other hand, his book title may reflect the future many fear — one in which we can’t trust AI to tell us the
truth.
Musk vs. OpenAI
In what I’d call an anticlimactic ruling, a jury found that Elon Musk’s lawsuit claiming that OpenAI had deceived him as a major investor when it switched from being a nonprofit to a for-profit company was tossed because it was filed too late, three years after the fact. Musk plans to appeal.
Musk argued in court that OpenAI poses a risk to the world because it has abandoned its safety-first mission. Meanwhile, he admits that his company, xAI, which plans to go public as early as this month, trains its models on OpenAI.
Lawyers for OpenAI had a field day pointing out Musk’s hypocrisy in filing a lawsuit against the state of Colorado, which passed a law opposing “algorithmic discrimination,” meaning businesses that use AI can’t discriminate when it comes to loans, employment, insurance policies or school admissions. They implied that his models are causing more harm than OpenAI ever has. (For instance, a May story in Wired recounts how a highly qualified medical-school grad couldn’t find a job because AI screening programs disqualified him for time gaps in his education due to cancer treatment.)
Additionally, Musk recruited former OpenAI employees, which doesn’t give much credence to his concern about safety. The defense noted that Musk appears to be trying to kill the competition to give his own AI company an advantage. The jury apparently wasn’t buying his claims of harm and concerns about safety, and rightly dismissed him.
The incestuous nature of AI investments between major players makes it likely we’ll see more acrimony and lawsuits as some fall behind and disappoint investors.
Journalist Toni Denis is a partner in Seeflection Inc.
