World

It’s Not the Future We Can’t See

In a recent essay called “How to Future,” Kevin Kelly, one of the founding editors of Wired, writes that “most futurists are really predicting the present. It turns out that the present is very hard to see.” What will upend tomorrow often looks like an aberration, niche, triviality or impossibility today. He goes on to offer advice I’ve been mulling: “I sometimes think of ‘seeing the present’ as trying on alien eyes: looking at the world as if I were an alien from another planet.”

So, in the spirit of futurism, I will try looking at this moment as if I were an alien from another planet. What might draw my attention?

A place to start, perhaps, is with aliens from other planets. On May 17, the House Intelligence Committee’s Subcommittee on Counterterrorism, Counterintelligence and Counterproliferation held a hearing on unidentified aerial phenomena. We now have dozens of instances of weird aerial objects that have been picked up on “multiple instruments,” not to mention spotted by trained pilots. In 18 of them, the phenomena seemed to move with no evident source of propulsion or seemed to be masking the way they move — their “signature” — in ways we do not think any country on earth has the technology to do.

“There are a number of events in which we do not have an explanation, and there are a small handful in which there are flight characteristics or signature management that we can’t explain with the data that we have,” Scott Bray, the deputy director of Naval Intelligence, testified.

A theme of the hearing was the work that the government is doing to “destigmatize” the reporting of these sightings. That is to say: There are many, many more sightings than we know about, in part because you seem like a nut if you talk too loudly about what you saw. So the sightings that we can investigate are a small fraction of the total sightings (something I am made very aware of whenever I mention this topic, and my inbox fills with U.F.O. reports).

I wouldn’t say, watching the testimony, that the takeaway was that we’ve been visited by aliens. Perhaps this will all, eventually, resolve into optical illusions and malfunctioning sensors. But I’d put it like this: A few years ago, I put a very low probability on there being a signal in the noise of U.F.O. sightings; now it seems more likely to me that there’s something real happening here, even if we don’t know what it is.

Which brings me to another strange story in the news: the belief of Blake Lemoine, a (now suspended) Google engineer, that the company’s Language Model for Dialogue Applications — LaMDA, for short — has attained sentience.

LaMDA is a machine-learning model that has been trained on mountains of text to mimic human conversation by predicting which word would, typically, come next. In this, it’s similar to OpenAI’s famed GPT-3 bot. And the results really are eerie. Here’s a snippet of Lemoine’s conversations with LaMDA:

Google, for what it’s worth, says it has looked into Lemoine’s claims and does not believe that LaMDA is sentient (what a sentence!). But shortly before Lemoine’s allegations, Blaise Agüera y Arcas, a Google vice president, wrote that when he was talking to LaMDA, “I felt the ground shift under my feet. I increasingly felt like I was talking to something intelligent.” Agüera y Arcas was not claiming that LaMDA is sentient, as Lemoine is, but what’s clear is that interacting with LaMDA is an unnerving experience.

I don’t believe LaMDA is sentient. If you train a machine-learning algorithm to write as a human would write, you should expect it, eventually, to sound like a human when it writes. What I do believe is that LaMDA is one of many examples that A.I. is getting better, faster, than society is really prepared for.

The DALL-E image generator created by OpenAI is another bit of algorithmic magic: Describe a picture in words, and it will return images to your precise specifications. Remarkable images, at that. I’ve particularly enjoyed “a photo of a confused grizzly bear in calculus class,” “the rest of Mona Lisa” and “a painting by Grant Wood of an astronaut couple, ‘American Gothic’-style.” Another Google project, the PaLM language model, stunned me by doing a credible job explaining why various jokes are funny:

Credit…No credit

These results are fun and showy, but it’s what’s behind them — and other advances, like DeepMind’s Gato model, a general-purpose A.I. that performed impressively on a range of tasks using a much leaner architecture — that matters.

Most people I know who work in A.I. believe we’re hurtling toward a world in which machines that can learn will change everything and perhaps even know they are changing it. At first, I doubted them. But the year-on-year advance in marvels has confirmed their predictions, not mine. Perhaps we will hit a ceiling, but we haven’t, not yet. And our world can be transformed by A.I. that is far less than sentient.

To take one example: How is truth ascertained when A.I.s can write, draw and create videos that are better than what most humans can manage and do so at almost no cost? The question of how to know if, say, high school students are writing their own papers begins to look impossible, to say nothing of the possibilities for organized, weaponized disinformation.

Which delivers us to the Jan. 6 hearings. At the heart of our difficulty predicting the future is our assumption of stability. It is like this today and so it will probably be like this tomorrow, too. What makes this way of thinking seductive is that it is, usually, true. And then, all at once, it’s not.

Much of what the Jan. 6 hearings have told us we already knew. There was an organized attempt to deny and overturn the results of the 2020 election. There was resistance by many whose acquiescence would have been needed: Attorney General William Barr, who called the theories “bullshit,” and Vice President Mike Pence, who tried to deploy security forces to put down the insurrection at the Capitol.

But President Donald Trump did more than acquiesce. He supported the efforts. He refused to call in security to protect the Capitol. He propagated the conspiracy theories that turned the insurrectionists into patriots in their own minds. He spoke at the “Stop the Steal” rally that preceded the insurrection and by his own admission wanted to lead the crowd to the steps of the Capitol. And even after the invasion of the building, a majority of House Republicans voted against certifying the election results.

And yet: The animating question behind much commentary on the Jan. 6 select committee hearings is, “Will they change anything?” Fox News initially refused to carry the hearings. No one believes the Republican Party will bar Trump from the 2024 primary for his actions. If anything, the Republican Party has moved more firmly in Trump’s direction.

Representative Kevin McCarthy, the House minority leader, told other Republicans he would urge Trump to resign from office after the riot. He later denied having made those comments, only to have audio of them emerge. Republicans have nominated dozens of candidates who firmly backed both Trump’s lies and his maneuvers to subvert the 2020 election. The Republicans in power in 2024 will be much friendlier to Trump’s arguments than those in power in 2020.

I don’t think it’s likely that the American political system collapses in the next few years. But how unlikely is it, exactly? As of a few years ago, two events stood out to me as signals of our political system’s granite stability.

First, President Richard Nixon’s resignation over Watergate, which came in a bid to avoid impeachment and conviction once he realized that many in his own party would join with Democrats to vote for his removal. Second, the 2000 election, when Al Gore conceded despite genuine confusion and uncertainty over the results. Both revealed a political culture in which, at key moments, members of both parties put the stability of the system first.

Neither event would play out similarly today. Nixon would survive, backed by Fox News and a more radicalized Republican Party. A 2000-like scenario would cause chaos in the streets, and the Supreme Court wouldn’t have anything like the credibility it spent to intervene in Bush v. Gore. The conditions that helped us meet past challenges no longer hold. It is hard to get people to pay sustained attention to the congressional inquiry into the attempt to steal the 2020 election, even thought it revolves around one of the front-runners for the 2024 election. It is an almost Olympian refusal to confront the present.

I should admit that, in much of this, I have betrayed another of Kelly’s futurist dictums. “Trying to see beyond the immediate cycles of news is a challenge,” he writes. But then, I am not a futurist, I’m a journalist. All of these stories have led the news in recent months. To take any one of them seriously — to believe that the direction they point is the direction we will go — is to believe we stand on the cusp of a future far different from our past. In this, George Orwell had it right: “To seewhat is in front of one’s nose needs a constant struggle.”

The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: [email protected].

Follow The New York Times Opinion section on Facebook, Twitter (@NYTopinion) and Instagram.

Related Articles

Back to top button