Good morning. My old newsroom colleague Jack Clark, now a co-founder of high-flying AI startup Anthropic, recently shared some thoughts on artificial intelligence that I found enlightening.
“What we are dealing with is a real and mysterious creature, not a simple and predictable machine,” he wrote in an essay published Monday. “And like all the best fairytales, the creature is of our own creation.”
With AI showing ever more capability—and crucially, agency—Clark says he’s both optimistic and “deeply afraid” of what that means for the future. But we humans have some work to do, too.
“We need to spend a bit less time talking about the specifics of the technology and trying to convince people of our particular views of how it might go wrong … and more time listening to people and understanding their concerns,” he wrote, adding: “The rest of the world, which will surely want—and deserves—a vote over this.”
Hear, hear. Today’s tech news below. —Andrew Nusca
Want to send thoughts or suggestions to Fortune Tech? Drop a line here.
U.S. seizes $15 billion in bitcoin held by 'pig butchering' scammer
The U.S. Department of Justice said Monday that it had made its largest forfeiture seizure in its history.
The take: About $15 billion worth of Bitcoin.
Prosecutors say the man who once held the crypto allegedly oversaw a massive fraud operation based in Cambodia.
(What kind of scam? “Pig butchering,” as it's known, wherein the perpetrators build trust with victims to convince them to “invest” and ultimately hand over their hard-earned crypto.)
The alleged pig butcher, Chen Zhi, is the founder and chairman of the Prince Group. He has been formally charged with wire fraud conspiracy and money laundering conspiracy.
His conglomerate has been labeled by the Treasury Department as “one of Asia’s largest transnational criminal organizations” for allegedly trafficking hundreds of people and forcing them to work in its facilities.
The 37-year-old remains at large and faces up to 40 years in prison if convicted. —AN
Coming soon: A personality for ChatGPT
Gather round, children, to remember a time when large language models were warm, upbeat…even funny.
It was a generation ago—18 whole months in the land of AI. But widespread concerns (read: lawsuits) about AI chatbots taking advantage of the mental health of some users led to a more defensive (read: boring) personality posture.
That will change in a few weeks, OpenAI CEO Sam Altman says, when its ChatGPT will be updated to allow for more, well, spirited interactions.
“Now that we have been able to mitigate the serious mental health issues and have new tools, we are going to be able to safely relax the restrictions in most cases,” he wrote in a social media post. “If you want your ChatGPT to respond in a very human-like way, or use a ton of emoji, or act like a friend, ChatGPT should do it.”
And what’s more human than sex? Altman added that once OpenAI rolls out age-gating “more fully,” it will loosen its chatbot restrictions enough (in a rather Meta-like move) to allow for “erotica for verified adults.”
Oh, ChatGPT. You’re making me blush. —AN
‘Pixnapping’ attacks are trying to steal your 2FA codes
Use a Google Android device? Best listen up.
A new attack promises to quietly steal your sensitive data in under a minute.
Dubbed “Pixnapping,” the threat requires first downloading a malicious app. But that wayward app doesn’t require system permissions to read the data that any other installed app displays on the screen, according to Ars Technica.
(Why the name? Because the attack involves capturing pixels on your display and mapping them to letters, numbers, or shapes.)
It’s not hard to imagine what else can be stolen with such an attack: Text messages, emails, passwords, proprietary financial information. And because of Pixnapping’s relative speed, even your two-factor authentication codes are at risk.
Though the vulnerability is certain, Google says it has yet to see it occur in the wild. The company partially fixed the issue with a September patch, but another, scheduled for December, is necessary to cover its full extent. —AN
More tech
—Walmart, OpenAI partner. Browse and purchase products from within ChatGPT.
—Intel debuts AI chip. Its “Crescent Island” GPU is tailored for inference.
—Oura is now worth $11 billion, according to the smart ringmaker’s latest fundraise.
—Microsoft ends Windows 10 support. Forty percent of Windows PCs still run on it.
—Spotify and Netflix partner. Select Spotify vodcasts come to the Hollywood streamer next year.
—Meta overhauls Instagram teen accounts. They now align with PG-13 movie ratings, it says.
—“Custom shoppable TV screensavers,” coming soon to a DirecTV near you.
