Why AI Scientists Are Freaking Us Out
Well that’s one way to get attention. According to a statement issued today by a bunch of leading names in AI, “mitigating the risk of extinction from AI should be a global priority.” Right. It’s hard to imagine anyone out there disagrees—who but the most extreme misanthrope wants to extinguish the human race? The question, then, is why the very people responsible for AI’s development are suggesting it should be a priority to stop the new technology from killing off every human. Aren’t they the best people to ensure that doesn’t happen?
Yes and no. AI scientists want governments’ help, including via an “ambitious global coordination,” according to Dan Hendrycks, the director of the Center for AI Safety, which was responsible for the statement’s publication. Otherwise there’s a danger of “unchecked competitive pressures” leading to “bad outcomes,” he told me. That makes sense. But let’s face it: Given the difficulty that climate activists have had in getting nations globally to respond to the threat of global warming, and the current chilling of relations between the U.S., China and Russia, it will take more than this statement to bring about a globally coordinated regulatory response to AI.
Cynics will say we shouldn’t read too much into the statement, that the scientists are essentially virtue-signaling, preempting political calls for regulation by taking the lead in that campaign themselves. An alternative cynical view is that existing AI firms are simply pushing for regulation to make life difficult for future entrants. Either way, those involved with today’s statement are taking a big risk. It’s one thing to go before Congress, as OpenAI CEO Sam Altman has done, and acknowledge the “significant harms” posed by AI while calling for the drafting of legislation as a response. It’s another to describe those risks as extinction-level, on a par with “pandemics and nuclear war.” The only logical response to that suggestion is to ban AI—or regulate it so tightly that even firms already in the business suffer.
That’s surely not what these scientists are looking for. But what do they want? Perhaps instead of issuing a single-sentence statement meant to freak everyone out, AI scientists should use their considerable skills to figure out a solution to the problem they have wrought.
Crypto Spring? Hold That Thought
For much of 2023, bitcoin has been up and to the right—by the end of April, the world’s dominant cryptocurrency was up roughly 76%, which helped drive all kinds of silliness in the sector. May, however, has turned out differently. Bitcoin is on track to close the month down from where it ended April. And while bitcoin has directionally tracked the equities market over the past 18 months, its performance in May was worse than that of the stock market. The S&P 500, for comparison, is on track to close up nearly 1 percentage point for May.
And bitcoin’s weak performance isn’t the only indicator that crypto fans’ predictions of the sector’s recovery were premature. Consider this:
Yet another prominent crypto startup laid off 30% of its staff today.
Crypto venture funding fell 87% in the first quarter, according to PitchBook. The Block reported last week that one of crypto’s most prominent investors is now expanding its focus to include AI.
Trading volumes of non-fungible tokens have trended downward since early February, when a zero-fee marketplace for the digital collectibles drove a short-lived pop, according to data from Hildebert Moulié, a data scientist at Dragonfly Capital. That explains why at least one well-funded NFT startup decided to shut down two weeks ago.
That list of challenges doesn’t even include all the added scrutiny from regulators, complicating crypto company endeavors this year. The groundhog has seen its own shadow. The crypto winter isn’t over yet.—Akash Pasricha