Another week, another dizzying spate of A.I. news
The pace of developments is picking up. How does this all shake out?
I’m far from the first to note it, but it does seem like the velocity of the A.I. rollout has been astounding. Things are moving at a frenzied pace. Just since Ariella’s piece on Tuesday, there has been another deluge of news just about data centers proliferating around the world — and the issues they’re causing in local communities — among other A.I. developments.
As a reporter, it’s been interesting to see A.I. dominate the news cycle — and the political arena — in a way that no other tech storyline has before it. For many years, tech was looked at in the news business as a bit of a niche: a subject that usually got housed, at least at newspapers and other media companies, under a broader aegis of a Business or economy-focused section. Tech was an addendum. A sidecar. A quirky little beat from the territories Out West California. The social media boom never moved as quickly — or was covered as aggressively — in the media. Perhaps to great detriment.
This shift from side dish to main course has been in the works for years as technology increasingly permeates our lives and the industries around us. Now, you’d be hard-pressed to find any beat, from culture reporting to sports and beyond, that isn’t touched by a technological angle or aspect. And it is a signifier of its own that A.I. is dominating discussion in Washington.
Still, I can’t help but pick up a feverish feeling underlying the A.I. race — something a bit breathless, glossy with sweat and steam, an intensity to this push that feels novel. The big money in tech is getting even bigger. The concerns are ever-growing but seem to confusingly be blown past.
Is this the result of a handful of powerful companies each competing to stake out uncharted terrain in what may amount to a short window — the four years of Trump’s second term — of exceedingly permissive attitudes in Washington? Or is there also something vaguely frenzied and apocalyptic that is inherent to this new industry?
For this week’s newsletter, I’m doing a link round-up that is heavily A.I.-dominated.
FIRST UP: Google released a new A.I. agent this week that can navigate web browsers for you, in theory saving people some time and energy they spend online. It automatically moves your cursor around the screen, ordering groceries and clothes, and conceivably filling out forms and completing other tasks. But the details about the new product, which is still being developed, do not necessarily inspire confidence, as reported in TechCrunch:
From there, the A.I. agent navigated to a grocery store’s website — in this case, Safeway — and then searched for and added items to a virtual shopping cart. One thing that’s immediately evident is how slow the agent is: There were about 5 seconds of delay between each cursor movement. At times, the agent stopped its task and reverted back to the chat window, asking for clarification about certain items (how many carrots, etc.).
The outlet notes that the terms for using the product require users allow it to take screenshots of their browser windows, a reminder of the age-old adage, particularly apt for tech, that if a product is free for you, then you are the product being sold.
On the heels of the BIG A.I. BONANZA IN D.C. last week, Meta, which has aggressively scaled up A.I. investments, released a letter about “personal superintelligence,” (A.I.) this week from our pal “Mark.”
Zuckerberg sounded the requisite safety concerns that now seem to be required for industry leaders to mention as they work to engender the public’s trust on A.I., while going on to talk about how A.I. will herald an era of personal empowerment as long as it’s put in everyone’s hands.
Despite the apparent dig at competitors, including OpenAI, the note was actually perfectly aligned with the lobbying pitch that Altman/OpenAI trotted out last week in D.C., as we wrote about: that A.I. needs to be put into as many people’s hands as possible for society to benefit.
FAKE FRIENDS: When a study from the Harvard Business Review analyzed the ways that people were using A.I., “therapy / compassion” was not only near the top of the list, but in the number one slot.
Since then, many stories have come out to look at the cases where this kind of dynamic goes horribly off the rails, including a big takeout about AI-induced spiraling in the New York Times, and another shorter feature this week from Futurism about a support group that has been launched for people struggling with A.I.-induced psychosis. That group also runs a website where they collect stories about people suffering in this way, called The Human Line Project.
"There's no playbook," one early member of the support group, a US-based father of young children whose wife has been using ChatGPT to communicate with what she says are spiritual entities. "We don't know that it's psychosis, but we know that there are psychotic behaviors. We know that people are in delusion, that it's fixed belief… you can pull some of these clinical terms, but it's so surreal. It's like an episode of 'Black Mirror,' but it's happening."
NUMBER GOES UP: Compensation details from the A.I. boom are jaw-dropping. According to a report this week in Wired, Meta is offering some A.I. researchers it has been trying to poach compensation packages of as much as $200-$500 million over a four-year period, including $50-$100 million in the first year. One offer was reportedly worth as much as much as $1 billion.
On that note, rents are up, office vacancies are down, and downtown is (maybe?) coming back in post-pandemic San Francisco due to the surge of energy and investment in the A.I sector. A worthwhile read in the Washington Post.
MORE ON DATA CENTERS: This piece from the Financial Times, is a very solid and comprehensive look at the construction of data centers around the U.S., their astronomically high energy and water usage, and other issues, with some fantastic images.
A.I. HALLUCINATIONS in court, which we wrote about a few weeks ago, continue to pop up. Meanwhile, Microsoft researchers determined that A.I. might be most helpful as an assistant not replacement worker, while noting that jobs involving physical labor remain the safest from being made obsolete.
SCHLOCK BOTS: Spotify removed an A.I.-generated track from the platform that someone uploaded, and falsely claimed was from a country singer who died in 1989.
Nearly 65 percent of Gen-Z workers are concerned about A.I.-induced job loss or disruption, according to a new study.
QUOTE OF THE WEEK: “There are enormous ethical and legal concerns…Patients across America should be very worried that their medical records are going to be used in ways that harm them and their families.”
That’s what Lawrence Gostin, a Georgetown University law professor who specializes in public health, told the Associated Press about the Trump administration plan to launch a new private health tracking system that will allow big tech companies to share patient data. Bloomberg has a more upbeat take.
Another piece that very adeptly Explains Everything That is Going On Now and Why: “The rise of end times fascism,” in The Guardian. This one’s from a few months ago; if you haven’t read it already, it’s quite good.
See you next week.