It’s a mistake to assume Elon Musk is lying all the time—but his Neuralink comment needs verification

Let me start with a kind of defense of Musk: I think some people have gone overboard by assuming that everything that comes out of the man’s mouth is a lie. For an example of what I’m talking about, check out the responses to CNBC’s post on social media service Threads reporting what Musk said. I get that Threads is a place where mad-at-Musk former Twitter users have gone, but the suggestion by one of them that CNBC should have interviewed the patient before reporting Musk’s comment is absurd on privacy as well as logistical grounds.

There is no question that Musk does spout bull, regularly. Sometimes they’re real whoppers, like the infamous “funding secured” tweet about taking Tesla private, or the idea that Tesla vehicles are always just about to become fully autonomous. Sometimes the lie may be less obvious to American observers. As a white South African who also grew up under apartheid, my eyebrows shot up when Musk told a court that “pedo guy” was a commonly used insult in the country during the 1980s (I certainly never heard it, and I heard plenty of offensive language). And they entirely departed my face when he told biographer Walter Isaacson that he once witnessed a corpse “with a knife still sticking out of his brain” while en route to an “anti-apartheid music concert” (South Africa was a police state with heavily enforced racial segregation that overwhelmingly privileged us whites, so anti-apartheid concerts were not a thing, and any streets Musk would have seen were far from being mean.)

However, while I can certainly appreciate why many people are eager to write Musk off as just another one of the narcissistic grifters who have unfortunately become some of the world’s most powerful men, he’s more complex than that. As OpenAI CEO Sam Altman put it last year, “Elon desperately wants the world to be saved. But only if he can be the one to save it.” Musk may lie to make himself seem more impressive or get himself out of a tight spot, but he’s also a genuinely smart person who has managed to build genuinely impressive businesses. And because some of his businesses carry the plausible promise of moving humanity forward in positive ways—Tesla’s electrification of transport, SpaceX’s horizon-expanding rockets—I believe it’s as foolish to reflexively write off everything Musk says as it is to be overly credulous.

Instead, we should approach his statements with healthy skepticism. And in the case of his brain-chip startup, Neuralink, that would be the appropriate response even if it weren’t Musk at the company’s helm. This is medical science we’re talking about, where claims should always be interrogated by third parties before being accepted as fact. That’s called peer review, and its absence is the real problem here.

As Coins2Day reported in early 2022, Neuralink is notorious for not publishing its research. That was always a problem, but now that its implants are going into real live people, it’s closer to horrifying. If Neuralink’s patient has indeed controlled a cursor with his mind, that isn’t a breakthrough in itself—rival brain-computer–interface (BCI) company Blackrock Neurotech won the FDA’s “breakthrough device” designation for its MoveAgain implant all the way back in 2021, for doing just that, and its paralyzed patients are now creating art with the device. Synchron, another rival whose BCI device taps into the brain via a veinal stent rather than being implanted, has even demonstrated smartphone control. But cursor control would demonstrate that Neuralink’s wireless N1 device has successfully taken a step on its own path. If only someone outside Neuralink could verify this!

As University of Washington bioethicist Nancy Jecker and neurologist Andrew Ko wrote a week ago, Neuralink’s lack of transparency does a disservice to other researchers in the BCI field, denying them information that could aid their work. There is little sense of the common good in play here. Jecker and Ko’s piece is well worth reading, as it outlines way more ethical concerns about Neuralink’s approach than I have space to list here. But Hastings Center bioethicists Arthur Caplan and Jonathan Moreno usefully summarize the most fundamental issue in a separate article about Neuralink, also published last week: “Science by press release, while increasingly common, is not science. When the person paying for a human experiment with a huge financial stake in the outcome is the sole source of information, basic ethical standards have not been met.”

So we shouldn’t necessarily assume Musk is lying to us about Neuralink’s progress—but we should be very worried that his cursory comment, made in an X Spaces session for heaven’s sake, is all we have to go on. More news below.

David Meyer

Want to send thoughts or suggestions to Data Sheet? Drop a line here.

NEWSWORTHY

Google debuts open-source AI. Google, which has over the past year stayed in the proprietary AI camp while rivals like Meta and Mistral released open-source models to the public, has finally followed their lead. As Coins2Day’s Jeremy Kahn reports, Google says its new Gemma models are a response to demand from developers who want to combine proprietary and open-source technology when creating their AI applications. Unlike Google’s multimodal Gemini models, which can handle a variety of inputs, Gemma can only handle text. Gemma’s licensing is less restrictive than that for Meta’s Llama 2, though.

X’s Navalnaya “error.” When X yesterday suspended the account of Yulia Navalnaya, the widow of possibly murdered Russian opposition leader Alexei Navalny, many were quick to accuse Musk’s platform of censorship. But, as Reuters reports, X says the brief suspension was an accident, caused by a malfunction of its “defense mechanism against manipulation and spam.”

Encrypted messaging developments. Apple is upgrading iMessage so that its encryption can withstand potential future breakthroughs in quantum computing, which many fear could make current encryption easy to break. In a blog post, Apple’s security team said their new PQ3 protocol had “the strongest security properties of any at-scale messaging protocol in the world” and would be introduced over the next year. Meanwhile, the widely used Signal secure messaging app yesterday launched usernames, so users don’t have to give people their phone numbers to have them get in touch. The feature will roll out in the coming weeks.

SIGNIFICANT FIGURES

26%

—The reduction in data storage by EU firms (relative to their U.S. Counterparts) in response to the General Data Protection Regulation (GDPR), which went into effect in 2018. That’s according to new research, which also found a 15% drop in data processing, and which suggests the “cost of data” in the EU has increased by 20% on average since the privacy law’s introduction.

IN CASE YOU MISSED IT

What the latest flurry of AI news has in common with the Great Tea Race, by Jeremy Kahn

Semiconductor giant ASML is one of Europe’s most ‘over-owned’ stocks, Morgan Stanley says—but it doesn’t necessarily signal the beginning of an AI bubble, by Ryan Hogg

Walmart’s $2.3 billion Vizio acquisition is very much about Amazon envy, by Jason Del Rey

Ex–Yahoo CEO Marissa Mayer on AI, remote work and her new app: ‘I think it is really hard to join an organization that is fully remote because that notion of culture gets lost,’ by the Associated Press

‘The more extreme ends of the spectrum won’t like us very much’: The ousted cofounder of controversial social app Parler launches new ‘civil conversation’ platform Hedgehog, by John Matze (Commentary)

BEFORE YOU GO

Streaming over borders. If you’re used to using your streaming services while traveling across borders, that’s an era that may soon be over. As Bloomberg reports, the World Trade Organization’s moratorium on so-called digital customs charges is set to expire next month, and its renewal could be stymied by just one WTO member. Three—Indonesia, South Africa, and India—have signaled they want the long-standing deal to end.