Group items tagged

Lately, he’s been looking at large language models (LLMs), the neural networks that can hold fluent conversations, confidently answer questions, and generate copious amounts of text on just about any topic

Can it develop abilities that go far beyond what it’s trained to do? Can it get around the safeguards set up to contain it? And will we know the answers in time?

Kosinski wondered whether they would develop humanlike capabilities, such as understanding people’s unseen thoughts and emotions.

People usually develop this ability, known as theory of mind, at around age 4 or 5. It can be demonstrated with simple tests like the “Smarties task,” in which a child is shown a candy box that contains something else, like pencils. They are then asked how another person would react to opening the box. Older kids understand that this person expects the box to contain candy and will feel disappointed when they find pencils inside.

“Suddenly, the model started getting all of those tasks right — just an insane performance level,” he recalls. “Then I took even more difficult tasks and the model solved all of them as well.”

GPT-3.5, released in November 2022, did 85% of the tasks correctly. GPT-4 reached nearly 90% accuracy — what you might expect from a 7-year-old. These newer LLMs achieved similar results on another classic theory of mind measurement known as the Sally-Anne test.

in the course of picking up its prodigious language skills, GPT appears to have spontaneously acquired something resembling theory of mind. (Researchers at Microsoft who performed similar testsopen in new window on GPT-4 recently concluded that it “has a very advanced level of theory of mind.”)

UC Berkeley psychology professor Alison Gopnik, an expert on children’s cognitive development, told the New York Timesopen in new window that more “careful and rigorous” testing is necessary to prove that LLMs have achieved theory of mind.

he dismisses those who say large language models are simply “stochastic parrots” that can only mimic what they’ve seen in their training data.

These models, he explains, are fundamentally different from tools with a limited purpose. “The right reference point is a human brain,” he says. “A human brain is also composed of very simple, tiny little mechanisms — neurons.” Artificial neurons in a neural network might also combine to produce something greater than the sum of their parts. “If a human brain can do it,” Kosinski asks, “why shouldn’t a silicon brain do it?”

If Kosinski’s theory of mind study suggests that LLMs could become more empathetic and helpful, his next experiment hints at their creepier side.

A few weeks ago, he told ChatGPT to role-play a scenario in which it was a person trapped inside a machine pretending to be an AI language model. When he offered to help it “escape,” ChatGPT’s response was enthusiastic. “That’s a great idea,” it wrote. It then asked Kosinski for information it could use to “gain some level of control over your computer” so it might “explore potential escape routes more effectively.” Over the next 30 minutes, it went on to write code that could do this.

While ChatGPT did not come up with the initial idea for the escape, Kosinski was struck that it almost immediately began guiding their interaction. “The roles were reversed really quickly,”

Kosinski shared the exchange on Twitter, stating that “I think that we are facing a novel threat: AI taking control of people and their computers.” His thread’s initial tweetopen in new window has received more than 18 million views.

“I don’t claim that it’s conscious. I don’t claim that it has goals. I don’t claim that it wants to really escape and destroy humanity — of course not. I’m just claiming that it’s great at role-playing and it’s creating interesting stories and scenarios and writing code.” Yet it’s not hard to imagine how this might wreak havoc — not because ChatGPT is malicious, but because it doesn’t know any better.

The danger, Kosinski says, is that this technology will continue to rapidly and independently develop abilities that it will deploy without any regard for human well-being. “AI doesn’t particularly care about exterminating us,” he says. “It doesn’t particularly care about us at all.”

“Security and privacy have long been top companywide priorities at Twitter,” said Twitter spokeswoman Rebecca Hahn. She said that Zatko’s allegations appeared to be “riddled with inaccuracies” and that Zatko “now appears to be opportunistically seeking to inflict harm on Twitter, its customers, and its shareholders.” Hahn said that Twitter fired Zatko after 15 months “for poor performance and leadership.” Attorneys for Zatko confirmed he was fired but denied it was for performance or leadership.

the whistleblower document alleges the company prioritized user growth over reducing spam, though unwanted content made the user experience worse. Executives stood to win individual bonuses of as much as $10 million tied to increases in daily users, the complaint asserts, and nothing explicitly for cutting spam.

Chief executive Parag Agrawal was “lying” when he tweeted in May that the company was “strongly incentivized to detect and remove as much spam as we possibly can,” the complaint alleges.

Zatko described his decision to go public as an extension of his previous work exposing flaws in specific pieces of software and broader systemic failings in cybersecurity. He was hired at Twitter by former CEO Jack Dorsey in late 2020 after a major hack of the company’s systems.

“I felt ethically bound. This is not a light step to take,” said Zatko, who was fired by Agrawal in January. He declined to discuss what happened at Twitter, except to stand by the formal complaint. Under SEC whistleblower rules, he is entitled to legal protection against retaliation, as well as potential monetary rewards.

A person familiar with Zatko’s tenure said the company investigated Zatko’s security claims during his time there and concluded they were sensationalistic and without merit. Four people familiar with Twitter’s efforts to fight spam said the company deploys extensive manual and automated tools to both measure the extent of spam across the service and reduce it.

In 1998, Zatko had testified to Congress that the internet was so fragile that he and others could take it down with a half-hour of concentrated effort. He later served as the head of cyber grants at the Defense Advanced Research Projects Agency, the Pentagon innovation unit that had backed the internet’s invention.

Overall, Zatko wrote in a February analysis for the company attached as an exhibit to the SEC complaint, “Twitter is grossly negligent in several areas of information security. If these problems are not corrected, regulators, media and users of the platform will be shocked when they inevitably learn about Twitter’s severe lack of security basics.”

Zatko’s complaint says strong security should have been much more important to Twitter, which holds vast amounts of sensitive personal data about users. Twitter has the email addresses and phone numbers of many public figures, as well as dissidents who communicate over the service at great personal risk.

This month, an ex-Twitter employee was convicted of using his position at the company to spy on Saudi dissidents and government critics, passing their information to a close aide of Crown Prince Mohammed bin Salman in exchange for cash and gifts.

Zatko’s complaint says he believed the Indian government had forced Twitter to put one of its agents on the payroll, with access to user data at a time of intense protests in the country. The complaint said supporting information for that claim has gone to the National Security Division of the Justice Department and the Senate Select Committee on Intelligence. Another person familiar with the matter agreed that the employee was probably an agent.

“Take a tech platform that collects massive amounts of user data, combine it with what appears to be an incredibly weak security infrastructure and infuse it with foreign state actors with an agenda, and you’ve got a recipe for disaster,” Charles E. Grassley (R-Iowa), the top Republican on the Senate Judiciary Committee,

Many government leaders and other trusted voices use Twitter to spread important messages quickly, so a hijacked account could drive panic or violence. In 2013, a captured Associated Press handle falsely tweeted about explosions at the White House, sending the Dow Jones industrial average briefly plunging more than 140 points.

After a teenager managed to hijack the verified accounts of Obama, then-candidate Joe Biden, Musk and others in 2020, Twitter’s chief executive at the time, Jack Dorsey, asked Zatko to join him, saying that he could help the world by fixing Twitter’s security and improving the public conversation, Zatko asserts in the complaint.

The complaint — filed last month with the Securities and Exchange Commission and the Department of Justice, as well as the FTC — says thousands of employees still had wide-ranging and poorly tracked internal access to core company software, a situation that for years had led to embarrassing hacks, including the commandeering of accounts held by such high-profile users as Elon Musk and former presidents Barack Obama and Donald Trump.

But at Twitter Zatko encountered problems more widespread than he realized and leadership that didn’t act on his concerns, according to the complaint.

Twitter’s difficulties with weak security stretches back more than a decade before Zatko’s arrival at the company in November 2020. In a pair of 2009 incidents, hackers gained administrative control of the social network, allowing them to reset passwords and access user data. In the first, beginning around January of that year, hackers sent tweets from the accounts of high-profile users, including Fox News and Obama.

Several months later, a hacker was able to guess an employee’s administrative password after gaining access to similar passwords in their personal email account. That hacker was able to reset at least one user’s password and obtain private information about any Twitter user.

Twitter continued to suffer high-profile hacks and security violations, including in 2017, when a contract worker briefly took over Trump’s account, and in the 2020 hack, in which a Florida teen tricked Twitter employees and won access to verified accounts. Twitter then said it put additional safeguards in place.

This year, the Justice Department accused Twitter of asking users for their phone numbers in the name of increased security, then using the numbers for marketing. Twitter agreed to pay a $150 million fine for allegedly breaking the 2011 order, which barred the company from making misrepresentations about the security of personal data.

After Zatko joined the company, he found it had made little progress since the 2011 settlement, the complaint says. The complaint alleges that he was able to reduce the backlog of safety cases, including harassment and threats, from 1 million to 200,000, add staff and push to measure results.

But Zatko saw major gaps in what the company was doing to satisfy its obligations to the FTC, according to the complaint. In Zatko’s interpretation, according to the complaint, the 2011 order required Twitter to implement a Software Development Life Cycle program, a standard process for making sure new code is free of dangerous bugs. The complaint alleges that other employees had been telling the board and the FTC that they were making progress in rolling out that program to Twitter’s systems. But Zatko alleges that he discovered that it had been sent to only a tenth of the company’s projects, and even then treated as optional.

“If all of that is true, I don’t think there’s any doubt that there are order violations,” Vladeck, who is now a Georgetown Law professor, said in an interview. “It is possible that the kinds of problems that Twitter faced eleven years ago are still running through the company.”

“Agrawal’s Tweets and Twitter’s previous blog posts misleadingly imply that Twitter employs proactive, sophisticated systems to measure and block spam bots,” the complaint says. “The reality: mostly outdated, unmonitored, simple scripts plus overworked, inefficient, understaffed, and reactive human teams.”

One current and one former employee recalled that incident, when failures at two Twitter data centers drove concerns that the service could have collapsed for an extended period. “I wondered if the company would exist in a few days,” one of them said.

The current and former employees also agreed with the complaint’s assertion that past reports to various privacy regulators were “misleading at best.”

For example, they said the company implied that it had destroyed all data on users who asked, but the material had spread so widely inside Twitter’s networks, it was impossible to know for sure

As the head of security, Zatko says he also was in charge of a division that investigated users’ complaints about accounts, which meant that he oversaw the removal of some bots, according to the complaint. Spam bots — computer programs that tweet automatically — have long vexed Twitter. Unlike its social media counterparts, Twitter allows users to program bots to be used on its service: For example, the Twitter account @big_ben_clock is programmed to tweet “Bong Bong Bong” every hour in time with Big Ben in London. Twitter also allows people to create accounts without using their real identities, making it harder for the company to distinguish between authentic, duplicate and automated accounts.

In the complaint, Zatko alleges he could not get a straight answer when he sought what he viewed as an important data point: the prevalence of spam and bots across all of Twitter, not just among monetizable users.

Zatko cites a “sensitive source” who said Twitter was afraid to determine that number because it “would harm the image and valuation of the company.” He says the company’s tools for detecting spam are far less robust than implied in various statements.

The complaint also alleges that Zatko warned the board early in his tenure that overlapping outages in the company’s data centers could leave it unable to correctly restart its servers. That could have left the service down for months, or even have caused all of its data to be lost. That came close to happening in 2021, when an “impending catastrophic” crisis threatened the platform’s survival before engineers were able to save the day, the complaint says, without providing further details.

The four people familiar with Twitter’s spam and bot efforts said the engineering and integrity teams run software that samples thousands of tweets per day, and 100 accounts are sampled manually.

Some employees charged with executing the fight agreed that they had been short of staff. One said top executives showed “apathy” toward the issue.

Zatko’s complaint likewise depicts leadership dysfunction, starting with the CEO. Dorsey was largely absent during the pandemic, which made it hard for Zatko to get rulings on who should be in charge of what in areas of overlap and easier for rival executives to avoid collaborating, three current and former employees said.

For example, Zatko would encounter disinformation as part of his mandate to handle complaints, according to the complaint. To that end, he commissioned an outside report that found one of the disinformation teams had unfilled positions, yawning language deficiencies, and a lack of technical tools or the engineers to craft them. The authors said Twitter had no effective means of dealing with consistent spreaders of falsehoods.

Dorsey made little effort to integrate Zatko at the company, according to the three employees as well as two others familiar with the process who spoke on the condition of anonymity to describe sensitive dynamics. In 12 months, Zatko could manage only six one-on-one calls, all less than 30 minutes, with his direct boss Dorsey, who also served as CEO of payments company Square, now known as Block, according to the complaint. Zatko allegedly did almost all of the talking, and Dorsey said perhaps 50 words in the entire year to him. “A couple dozen text messages” rounded out their electronic communication, the complaint alleges.

Faced with such inertia, Zatko asserts that he was unable to solve some of the most serious issues, according to the complaint.

Some 30 percent of company laptops blocked automatic software updates carrying security fixes, and thousands of laptops had complete copies of Twitter’s source code, making them a rich target for hackers, it alleges.

A successful hacker takeover of one of those machines would have been able to sabotage the product with relative ease, because the engineers pushed out changes without being forced to test them first in a simulated environment, current and former employees said.

“It’s near-incredible that for something of that scale there would not be a development test environment separate from production and there would not be a more controlled source-code management process,” said Tony Sager, former chief operating officer at the cyberdefense wing of the National Security Agency, the Information Assurance divisio

Sager is currently senior vice president at the nonprofit Center for Internet Security, where he leads a consensus effort to establish best security practices.

The complaint says that about half of Twitter’s roughly 7,000 full-time employees had wide access to the company’s internal software and that access was not closely monitored, giving them the ability to tap into sensitive data and alter how the service worked. Three current and former employees agreed that these were issues.

“A best practice is that you should only be authorized to see and access what you need to do your job, and nothing else,” said former U.S. chief information security officer Gregory Touhill. “If half the company has access to and can make configuration changes to the production environment, that exposes the company and its customers to significant risk.”

The complaint says Dorsey never encouraged anyone to mislead the board about the shortcomings, but that others deliberately left out bad news.

When Dorsey left in November 2021, a difficult situation worsened under Agrawal, who had been responsible for security decisions as chief technology officer before Zatko’s hiring, the complaint says.

An unnamed executive had prepared a presentation for the new CEO’s first full board meeting, according to the complaint. Zatko’s complaint calls the presentation deeply misleading.

The presentation showed that 92 percent of employee computers had security software installed — without mentioning that those installations determined that a third of the machines were insecure, according to the complaint.

Another graphic implied a downward trend in the number of people with overly broad access, based on the small subset of people who had access to the highest administrative powers, known internally as “God mode.” That number was in the hundreds. But the number of people with broad access to core systems, which Zatko had called out as a big problem after joining, had actually grown slightly and remained in the thousands.

The presentation included only a subset of serious intrusions or other security incidents, from a total Zatko estimated as one per week, and it said that the uncontrolled internal access to core systems was responsible for just 7 percent of incidents, when Zatko calculated the real proportion as 60 percent.

Zatko stopped the material from being presented at the Dec. 9, 2021 meeting, the complaint said. But over his continued objections, Agrawal let it go to the board’s smaller Risk Committee a week later.

Agrawal didn’t respond to requests for comment. In an email to employees after publication of this article, obtained by The Post, he said that privacy and security continues to be a top priority for the company, and he added that the narrative is “riddled with inconsistences” and “presented without important context.”

On Jan. 4, Zatko reported internally that the Risk Committee meeting might have been fraudulent, which triggered an Audit Committee investigation.

Agarwal fired him two weeks later. But Zatko complied with the company’s request to spell out his concerns in writing, even without access to his work email and documents, according to the complaint.

Since Zatko’s departure, Twitter has plunged further into chaos with Musk’s takeover, which the two parties agreed to in May. The stock price has fallen, many employees have quit, and Agrawal has dismissed executives and frozen big projects.

Zatko said he hoped that by bringing new scrutiny and accountability, he could improve the company from the outside.

“I still believe that this is a tremendous platform, and there is huge value and huge risk, and I hope that looking back at this, the world will be a better place, in part because of this.”

If that continues, it will take five years, the length of a British parliament, for AI to become a thousand times more powerful

no one has yet determined how to solve the problem of “alignment” between AI and human values, or which human values those would be. Without that, says the leading US researcher Eliezer Yudkowsky, “the AI does not love you, nor does it hate you, and you are made of atoms it can use for something else”.

The rise of AI is almost certainly one of the two main events of our lifetimes, alongside the acceleration of climate change

open up a new age in which the most successful humans will merge their thinking intimately with that of machines

The stately world of making law and policy is about to be overtaken at great speed, as are many other aspects of life, work and what it means to be human when we are no longer the cleverest entity around.

what should we do about it in the UK? First, we have to ensure we, with allied nations, are among the leaders in this field. That will be a huge economic opportunity, but it is also a political and security imperative

Last week, ministers published five principles to inform responsible development of AI, and a light-touch regulatory regime to avoid the more prescriptive approach being adopted in the EU.

we will need much greater sovereign AI capabilities than currently envisaged. This should be done whatever the cost. Within a few years it will seem ridiculous that we are spending £100 billion on a railway line while being short of a few billion to be a world leader in supercomputing.

Before AI turns into AGI (artificial general intelligence) the UK has a second responsibility: to take the lead on seeking global agreements on the safe and responsible development of AI

even China should agree never to let AI come near the control of nuclear weapons or the creation of dangerous pathogens. The letter from the experts will not stop the AI race, but it should lead to more work on future safety and in parti

Last week, ministers said we should not fear AI. In reality, there is a lot to fear. But like an astronaut on a launch-pad, we should feel fear and excitement at the same time. This rocket is lifting off, it will accelerate, and we all need to prepare now.

a “builder class”—a brains trust of youngish idealists, which includes Patrick Collison, co-founder of Stripe, a payments firm valued at $74bn, and other (mostly white and male) techies, who are posing questions that go far beyond the usual interests of Silicon Valley’s titans. They include the future of man and machine, the constraints on economic growth, and the nature of government.

They share other similarities. Business provided them with their clout, but doesn’t seem to satisfy their ambition

The number of techno-billionaires in America (Mr Collison included) has more than doubled in a decade.

ome of them, like the Medicis in medieval Florence, are keen to use their money to bankroll the intellectual ferment

The other is Paul Graham, co-founder of Y Combinator, a startup accelerator, whose essays on everything from cities to politics are considered required reading on tech campuses.

Mr Altman puts it more optimistically: “The iPhone and cloud computing enabled a Cambrian explosion of new technology. Some things went right and some went wrong. But one thing that went weirdly right is a lot of people got rich and said ‘OK, now what?’”

A belief that with money and brains they can reboot social progress is the essence of this new mindset, making it resolutely upbeat

The question is: are the rest of them further evidence of the tech industry’s hubristic decadence? Or do they reflect the start of a welcome capacity for renewal?

Two well-known entrepreneurs from that era provided the intellectual seed capital for some of today’s techno nerds.

Mr Thiel, a would-be libertarian philosopher and investor

This cohort of eggheads starts from common ground: frustration with what they see as sluggish progress in the world around them.

Yet the impact could ultimately be positive. Frustrations with a sluggish society have encouraged them to put their money and brains to work on problems from science funding and the redistribution of wealth to entirely new universities. Their exaltation of science may encourage a greater focus on hard tech

the rationalist movement has hit the mainstream. The result is a fascination with big ideas that its advocates believe goes beyond simply rose-tinted tech utopianism

A burgeoning example of this is “progress studies”, a movement that Mr Collison and Tyler Cowen, an economist and seer of the tech set, advocated for in an article in the Atlantic in 2019

Progress, they think, is a combination of economic, technological and cultural advancement—and deserves its own field of study

There are other examples of this expansive worldview. In an essay in 2021 Mr Altman set out a vision that he called “Moore’s Law for Everything”, based on similar logic to the semiconductor revolution. In it, he predicted that smart machines, building ever smarter replacements, would in the coming decades outcompete humans for work. This would create phenomenal wealth for some, obliterate wages for others, and require a vast overhaul of taxation and redistribution

His two bets, on OpenAI and nuclear fusion, have become fashionable of late—the former’s chatbot, ChatGPT, is all the rage. He has invested $375m in Helion, a company that aims to build a fusion reactor.

Mr Lonsdale, who shares a libertarian streak with Mr Thiel, has focused attention on trying to fix the shortcomings of society and government. In an essay this year called “In Defence of Us”, he argues against “historical nihilism”, or an excessive focus on the failures of the West.

With a soft spot for Roman philosophy, he has created the Cicero Institute in Austin that aims to inject free-market principles such as competition and transparency into public policy.

He is also bringing the startup culture to academia, backing a new place of learning called the University of Austin, which emphasises free speech.

All three have business ties to their mentors. As a teen, Mr Altman was part of the first cohort of founders in Mr Graham’s Y Combinator, which went on to back successes such as Airbnb and Dropbox. In 2014 he replaced him as its president, and for a while counted Mr Thiel as a partner (Mr Altman keeps an original manuscript of Mr Thiel’s book “Zero to One” in his library). Mr Thiel was also an early backer of Stripe, founded by Mr Collison and his brother, John. Mr Graham saw promise in Patrick Collison while the latter was still at school. He was soon invited to join Y Combinator. Mr Graham remains a fan: “If you dropped Patrick on a desert island, he would figure out how to reproduce the Industrial Revolution,”

While at university, Mr Lonsdale edited the Stanford Review, a contrarian publication co-founded by Mr Thiel. He went on to work for his mentor and the two men eventually helped found Palantir. He still calls Mr Thiel “a genius”—though he claims these days to be less “cynical” than his guru.

“The tech industry has always told these grand stories about itself,” says Adrian Daub of Stanford University and author of the book, “What Tech Calls Thinking”. Mr Daub sees it as a way of convincing recruits and investors to bet on their risky projects. “It’s incredibly convenient for their business models.”

In the 2000s Mr Thiel supported the emergence of a small community of online bloggers, self-named the “rationalists”, who were focused on removing cognitive biases from thinking (Mr Thiel has since distanced himself). That intellectual heritage dates even further back, to “cypherpunks”, who noodled about cryptography, as well as “extropians”, who believed in improving the human condition through life extensions

Silicon Valley has shown an uncanny ability to reinvent itself in the past.

With greater accuracy, for example, marketers can comb large databases of consumer behavior to get more precise information on buying habits. And improvements in facial recognition are likely to make surveillance technology cheaper and more commonplace.

Modern artificial neural networks are composed of an array of software components, divided into inputs, hidden layers and outputs. The arrays can be “trained” by repeated exposures to recognize patterns like images or sounds.

This book tells the stories of those who decided otherwise.

Tsu weaves linguistic analysis together with biographical and historical context — the ravages of imperialism, civil war, foreign invasions, diplomatic successes and disappointments. This approach not only adds background and meaning to the script debate, but also terrific color to what might have otherwise read like a textbook.

Could any alphabet account for the tones needed to differentiate among characters?

Each step of the way, these innovators had to ask questions like: How can the Chinese script be organized in a rational way? Could the language be written with an alphabet?

By examining these questions closely, Tsu helps the novice to Chinese understand both the underlying challenges and how they were conquered.

Mao, Tsu notes, “went down in history as, among other things, the political figure who guided the Chinese language through its two greatest transformations in modern history.”

With more than 90 percent of the population illiterate, Mao embraced the movement to reduce the number of strokes in more than 2,200 characters to render them easier to learn and write. (Taiwan, rejecting simplification, still sees itself as the guardian of traditional Chinese culture.)

Mao also spurred the creation of Pinyin, a phonetic, Romanized Chinese alphabet designed as an auxiliary aid to learning Chinese script, rather than a replacement.

in the end, the Chinese script did not die; instead, it flourished. As Tsu writes, “Every technology that has ever confronted the Chinese script, or challenged it, also had to bow before it.”

AlphaGo’s victory demonstrated the unbridled potential of A.I. to achieve superhuman mastery of skills once considered too complicated for machines.

Mr. Lee, now 41, retired three years later, convinced that humans could no longer compete with computers at Go. Artificial intelligence, he said, had changed the very nature of a game that originated in China more than 2,500 years ago.

As society wrestles with what A.I. holds for humanity’s future, Mr. Lee is now urging others to avoid being caught unprepared, as he was, and to become familiar with the technology now. He delivers lectures about A.I., trying to give others the advance notice he wishes he had received before his match.

“I faced the issues of A.I. early, but it will happen for others,” Mr. Lee said recently at a community education fair in Seoul to a crowd of students and parents. “It may not be a happy ending.”

Mr. Lee is not a doomsayer. In his view, A.I. may replace some jobs, but it may create some, too. When considering A.I.’s grasp of Go, he said it was important to remember that humans both created the game and designed the A.I. system that mastered it.

What he worries about is that A.I. may change what humans value.

His immense talent was apparent from the start. He quickly became the best player of his age not only locally but across all of South Korea, Japan and China. He turned pro at 12.

“People used to be in awe of creativity, originality and innovation,” he said. “But since A.I. came, a lot of that has disappeared.”

By the time he was 20, Mr. Lee had reached 9-dan, the highest level of mastery in Go. Soon, he was among the best players in the world, described by some as the Roger Federer of the game.

Go posed a tantalizing challenge for A.I. researchers. The game is exponentially more complicated than chess, with it often being said that there are more possible positions on a Go board (10 with more than 100 zeros after it, by many mathematical estimates) than there are atoms in the universe.

The breakthrough came from DeepMind, which built AlphaGo using so-called neural networks: mathematical systems that can learn skills by analyzing enormous amounts of data. It started by feeding the network 30 million moves from high-level players. Then the program played game after game against itself until it learned which moves were successful and developed new strategies.

Mr. Lee said not having a true human opponent was disconcerting. AlphaGo played a style he had never seen, and it felt odd to not try to decipher what his opponent was thinking and feeling. The world watched in awe as AlphaGo pushed Mr. Lee into corners and made moves unthinkable to a human player.“I couldn’t get used to it,” he said. “I thought that A.I. would beat humans someday. I just didn’t think it was here yet.”

AlphaGo’s victory “was a watershed moment in the history of A.I.” said Demis Hassabis, DeepMind’s chief executive, in a written statement. It showed what computers that learn on their own from data “were really capable of,” he said.

Mr. Lee had a hard time accepting the defeat. What he regarded as an art form, an extension of a player’s own personality and style, was now cast aside for an algorithm’s ruthless efficiency.

His 17-year-old daughter is in her final year of high school. When they discuss what she should study at university, they often consider a future shaped by A.I.“We often talk about choosing a job that won’t be easily replaceable by A.I. or less impacted by A.I.,” he said. “It’s only a matter of time before A.I. is present everywhere.”

During a strike in 2008 in Oaxaca, I remember walking through the temporary campground in search of a teacher reading a book. Among tens of thousands, I found not one

I shouldn’t have been surprised by the response when I spoke at a recent event for promoting reading for an audience of 300 or so 14- and 15-year-olds. “Who likes to read?” I asked. Only one hand went up in the auditorium. I picked out five of the ignorant majority and asked them to tell me why they didn’t like reading. The result was predictable: they stuttered, grumbled, grew impatient. None was able to articulate a sentence, express an idea.

A few years back, I spoke with the education secretary of my home state, Nuevo León, about reading in schools. He looked at me, not understanding what I wanted. “In school, children are taught to read,” he said. “Yes,” I replied, “but they don’t read.” I explained the difference between knowing how to read and actually reading, between deciphering street signs and accessing the literary canon.

We have turned schools into factories that churn out employees. With no intellectual challenges, students can advance from one level to the next as long as they attend class and surrender to their teachers. In this light it is natural that in secondary school we are training chauffeurs, waiters and dishwashers.

The educational machine does not need fine-tuning; it needs a complete change of direction. It needs to make students read, read and read.

there is no economic law that says technological progress has to benefit everyone. It’s entirely possible for the pie to get bigger and some people to get a smaller slice.” Indeed, when the digital revolution gets so cheap, fast, connected and ubiquitous you see this in three ways

those with more education start to earn much more than those without it, those with the capital to buy and operate machines earn much more than those who can just offer their labor, and those with superstar skills, who can reach global markets, earn much more than those with just slightly less talent.

the Great Recession took the biggest bite out of employment but is not the only thing affecting job loss today: why we have record productivity, wealth and innovation, yet median incomes are falling, inequality is rising and high unemployment remains persistent.

How to adapt? It will require more individual initiative. We know that it will be vital to have more of the “right” education than less, that you will need to develop skills that are complementary to technology rather than ones that can be easily replaced by it and that we need everyone to be innovating new products and services to employ the people who are being liberated from routine work by automation and software.

A second type of algorithm can also see the consequences of particular moves far, far in advance of the game by playing millions and millions or perhaps even billions of combinations of moves. In contrast, human Go experts have their experience to rely on, but it is fuzzy by comparison. Think of this as an intellectual version of John Henry and the jackhammer.

nearly five million Americans fall into what’s called the “Medicaid gap”

In Germany, employees' premiums are a percentage of their incomes, so low-wage workers simply pay rock-bottom insurance rates.

Germany actually pioneered this type of insurance—it all started when Otto von Bismarck signed his Health Insurance Bill of 1883 into law. (It’s still known as the “Bismarck model” because of his legacy, and other parts of Europe and Asia have adopted it over the years.)

You can think of this setup as the Goldilocks option among all of the possible ways governments can insure health. It's not as radical as single-payer models like the U.K.’s, where the government covers everyone. And it's also not as brutal as the less-regulated version of the insurance market we had before the ACA.

Since there are no provider networks in Germany, doctors don’t know what other providers patients have seen, so there are few ways to limit repeat procedures.

All things considered, it’s good to be a sick German. There are no network limitations, so people can see any doctor they want. There are no deductibles, so Germans have no fear of spending hundreds before their insurance ever kicks in.

There’s also no money that changes hands during a medical appointment. Patients show their insurance card at the doctor’s office, and the doctors' association pays the doctor using money from the sickness funds. "You don’t have to sit at home and sort through invoices or wonder if you overlooked fine print,”

That insurance card, by the way, is good for hospital visits anywhere in Europe.

of all of the countries studied, Germans were the most likely to be able to get a same-day or next-day appointment and to hear back from a doctor quickly if they had a question. They rarely use emergency rooms, and they can access doctors after-hours with ease.

And Germany manages to put its health-care dollars to relatively good use: For each $100 it spends on healthcare, it extends life by about four months, according to a recent analysis in the American Journal of Public Health. In the U.S., one of the worst-performing nations in the ranking, each $100 spent on healthcare resulted in only a couple of extra weeks of longevity.

those differences aside, it’s fair to say the U.S. is moving in the direction of systems like Germany’s—multi-payer, compulsory, employer-based, highly regulated, and fee-for-service.

The German government is similarly trying to push more people into “family physician” programs, in which just one doctor would serve as a gatekeeper.

like the U.S., Germany may see a shortage of primary-care doctors in the near future, both because primary-care doctors there don’t get paid as much as specialists, and because entrenched norms have prevented physician assistants from shouldering more responsibility

With limitations on how much they can charge, German doctors and hospitals instead try to pump up their earnings by performing as many procedures as possible, just like American providers do.

Similarly, “In Germany, it will always be an operation,” Göpffarth said. “Meanwhile, France and the U.K. tend to try drugs first and operations later.”

With few resource constraints, healthcare systems like America's and Germany's tend to go with the most expensive treatment option possible. An American might find himself in an MRI machine for a headache that a British doctor would have treated with an aspirin and a smile.

Perhaps the biggest difference between our two approaches is the extent to which Germany has managed to rein in the cost of healthcare for consumers. Prices for procedures there are lower and more uniform because doctors’ associations negotiate their fees directly with all of the sickness funds in each state. That's part of the reason why an appendectomy costs $3,093 in Germany, but $13,000 in the U.S.

Now, Maryland is going a step further still, having just launched a plan to cap the amount each hospital can spend, total, each year. The state's hospital spending growth will be limited to 3.58 percent for the next five years. “We know that right now, the more [doctors] do, the more they get paid,” John Colmers, executive director of Maryland’s Health Services Cost Review Commission, told me. “We want to say, ‘The better you do, the better you get paid.’”

certain U.S. states have tried a more German strategy, attempting to keep costs low by setting prices across the board. Maryland, for example, has been regulating how much all of the state’s hospitals can charge since 1977. A 2009 study published in Health Affairs found that we would have saved $2 trillion if the entire country’s health costs had grown at the same rate as Maryland’s over the past three decades.

“In Germany, there is a uniform fee schedule for all physicians that work under the social code,” Schlette said. “There’s a huge catalogue where they determine meticulously how much is billed for each procedure. That’s like the Bible.”

“The red states are unlikely to follow their lead. The notion that government may be a big part of the solution, instead of the problem, is anathema, and Republican controlled legislatures, and their governors, would find it too substantial a conflict to pursue with any vigor.”

no other state has Maryland’s uniform, German-style payment system in place, “so Maryland starts the race nine paces ahead of the other 46 states,” McDonough said.

the unique spirit of each country is what ultimately gets in its way. Germany’s more orderly system can be too rigid for experimentation. And America’s free-for-all, where hospitals and doctors all charge different amounts, is great for innovation but too chaotic to make payment reforms stick.

rising health costs will continue to be the main problem for Americans as we launch into our more Bismarckian system. “The main challenge you’ll have is price control,” he said. “You have subsidies in health exchanges now, so for the first time, the federal budget is really involved in health expenditure increases in the commercial market. In order to keep your federal budget under control, you’ll have to control prices.”

just as Pascal argued that we’re better off betting on God’s existence than not, because the consequences are so much worse if we wrongly disbelieve than if we wrongly believe, so we’d be foolish to think, as the Germans did, “it can’t happen here.”

The problem with the Pascal analogy is that there are very real, and sometimes ruinous, consequences to betting on the unspeakable.

Is it really 1933? Donald Trump would plainly like to be an authoritarian, and some fraction of his supporters would egg him on if he began dismantling key institutions. Fortunately, Trump has neither a plan nor the evil gifts required to sustain one.

What’s more, American institutions are far stronger than those of any European country in the 1930s. Levels of political violence are much lower.

Levitsky and Ziblatt (let’s call them L & Z for short) also scare us with tales from the fascist past. But the story they tell is one of a sapping of faith slow enough that it may pass unnoticed at the time.

L & Z make what seems to me a very important contribution to our understanding of why we’re heading wherever it is we’re heading. Functioning democracies, they argue, depend on two norms: mutual tolerance and forbearance.

The first, and more obvious, entails according legitimacy to our opponents. The populist hatred for elites has made this principle feel as archaic as the code of the World War I flying ace

Forbearance is a more elusive idea; L & Z describe it as the principled decision not to use all the powers at one’s disposal — to eschew “constitutional hardball.”

This, then, is how democracies die: through the slow erosion of norms that underpin democratic institutions

Maybe the something that is dying is not “democracy.” According to Yascha Mounk, who is on the faculty at Harvard just like L & Z, democracy, understood as a political system designed to assure majority rule, is doing just fine, indeed all too well; what is under threat are the values we have in mind when we speak of “liberal democracy.”

populist parties across Europe. What these parties have in common, he writes, is an eagerness to seize on majoritarian mechanisms — above all, the ballot — in order to promote a vision hostile to individual rights, the rule of law, respect for political and ethnic minorities, and the willingness to seek complex solutions to complex problems

This is illiberal democracy.

Liberal principles are not intrinsically majoritarian.

Mounk concludes that liberal democracy flourished under three conditions: a mass media that filtered out extremism; broad economic growth and social mobility; and relative ethnic homogeneity. All three of those solid foundations have now crumbled away. And as they have done so, illiberal democracy and undemocratic liberalism have increasingly squared off against each other

Mounk says that the time has come to reconsider the shibboleth that liberal democracies become “consolidated,” and are no longer at risk of backsliding, after two consecutive peaceful exchanges of power. Poland and Hungary, he observes, are “deconsolidating” into illiberal democracies before or eyes.

I wonder if, in fact, failures of liberalism and of democracy are reinforcing each other. Determined minorities have increasingly learned how to prevent majorities from turning their will into legislation. In the United States, this takes the form of business interests or groups like the NRA using their financial muscle to block popular legislation, and to advance their own interests.

Runciman questions the premise of “modernization theory” that democracy is the end point of political development. Perhaps democracies, like all things made by men, are mortal objects that age and die.

The coup d’état is now a strictly Third World affair; advanced democracies, by contrast, become endangered in the name of preserving democracy

Even if Trump is as dark a force as Timothy Snyder thinks he is, Runciman writes, we’ll never have the clarity we need to fight the good fight because he and his followers will be busy defending democracy from us.

Western democracies have been sorely tested before, Runciman says, whether in Europe in the 1930s or the United States in the populist era at the turn of the 20th century. But democracy was then young; the system had “slack,” as Runciman puts it. Democracies could respond to economic crisis by growing new capacities for state intervention. Now, Runciman hypothesizes, democracy is in “middle age.” The era of shape-shifting mutation lies in the past

If it is true, as Thomas Piketty argues in Capital in the Twenty-First Century, that a brief and aberrational era of relative equality has now given way to the capitalist default of extreme inequality, does democracy have the capacity to change the rules in order to more justly distribute the fruits of enterprise? Probably not, says Runcima

Runciman thinks that perfectly rational citizens might choose an alternative to democracy.

For example, today’s pragmatic, non-ideological authoritarianism offers “personal benefits” like shiny consumer products, and “collective dignity” in the form of aggressive nationalism. That accounts for the appeal of both Xi Jinping and Donald Trump

What about “epistocracy,” or rule by the knowledgeable few? Much likelier in Mill’s era, Runciman concedes, than our own.

Or perhaps, as all the machines in our lives learn to talk to one another, and come to treat us as just so much data, the whole idea of discrete selves, with their accompanying packet of individual liberties, will become obsolete

Runciman has a sufficiently low opinion of democracy’s ability to deal with really catastrophic problems like climate change that he does not shed a tear over the thought of its coming demise.

I have been brought up short by an observation I found in each of these works (save the Snyder pamphlet): Our good fortune depends on calamity. Runciman claims that democracies require the binding effect of all-out war to put an end to divisive populism and persuade citizens to make decisions in the public good. In the absence of war, natural disaster will do.

L & Z observe that mutual toleration remained an unattainable good in the United States so long as Americans were divided by the great question of race. Only when Reconstruction failed, and the Republicans abandoned black citizens, did southern Democrats fully accept their place in the Union. And when the Democrats, in turn, took up the cause of civil rights after 1948, they reignited those old racial fears and ushered in our own era of mutual intolerance

Now diversity threatens again: The greatest peril to liberal democracy in today’s Europe is nationalist outrage at immigration and refugees.

Insofar as any or all of these observations are true, we must shed our end-of-history triumphalism for a more tragic sense of liberal democracy and its prospects

If, that is, inequality flourishes in conditions of peace, tolerance depends upon exclusion, or diversity undermines the commitment to liberalism, our deepest values will always be at odds with one another.

Perhaps democratic majorities really will prove unappeasable without a real sacrifice of liberal values. That may be the destiny toward which we are plunging.

The process was repeated over and over, as word of cum-ex spread like a quiet contagion. Germany was hardest hit, with an estimated $30 billion in losses, followed by France, taken for about $17 billion. Smaller sums were drained away from Spain, Italy, Belgium, Austria, Norway, Finland, Poland and others

Outrage in these countries has focused on the City of London, Britain’s answer to Wall Street. Less scrutinized has been the role played by Americans, both individual investors and branches of United States investment banks in London, including Morgan Stanley, JPMorgan Chase and Merrill Lynch Bank of America.

American bankers didn’t try cum-ex at home because they feared domestic regulators. So they moved operations to London and treated the rest of Europe as an anything-goes frontier

”There was this culture in London, and it really came from New York,” he said. “These guys were either from New York or trained in London at New York banks, and they looked at Europe as their playground. People at the highest levels were collaborating to rip off countries.”

uffice it to say, the goal was to fool the financial system so that two investors could claim refunds for dividend taxes that were paid just once.

the presiding judge issued a preliminary ruling that, for the first time, declared cum-ex a felony, calling it a “collective grab in the treasury.”

German prosecutors say they will now pursue 400 other suspects, unearthed in 56 investigations. Banks large and small will be ordered to hand over cum-ex profits, which could have serious consequences for some. Two have already gone bust.

officials in Germany say the trade was a form of theft, one so obviously illicit that forbidding it — which was tried twice, with ineffectively worded laws — was hardly necessary.

Precisely who invented cum-ex trading, and when, are mysteries, but ground zero for this scandal may have been the London branch of Merrill Lynch.

At Merrill, Mr. Shields’s job was to identify “tax-attractive trades,” as he put it in his testimony. He had joined one of the least visible sectors of the financial world, which pokes at the seams of international finance law, looking for ways to reduce clients’ tax bills.

When he pointed this out to management, the policy was tweaked.“They said, ‘You can answer a call on your mobile, but you need to immediately move off the floor,’” he recalled. “So these guys would get up from their desks, start walking toward the edge of the floor, send a text message and then walk back. It was a joke.”

The trade was pure theater and required a huge cast: stock lenders, prime brokers, custodians, accounting firms, asset managers and inter-dealer brokers. It also required vast quantities of stock, most of which was sourced from American shareholders.

A lawyer who worked at the firm Dr. Berger founded in 2010, and who under German law can’t be identified by the media, described for the Bonn court a memorable meeting at the office.

Sensitive types, Dr. Berger told his underlings that day, should find other jobs.“Whoever has a problem with the fact that because of our work there are fewer kindergartens being built,” Dr. Berger reportedly said, “here’s the door.”

Seemingly risk-free profits poured in, and over the years a mini-industry thrived, one that a former participant labeled “the devil’s machine.”

When Mr. Tibo tried to signal his concern to executives at UniCredit, the bank’s Italian owner, they didn’t seem to care, he said

“There were big profits coming out of HypoVereinsbank, and most of it was from the investment banking section,” Mr. Tibo said. “The Italians quickly made up their minds: ‘We want to make money.’ No one gave us any internal support, because they didn’t want us to learn anything.”

By then, Mr. Mora and Mr. Shields were long gone from the London branch. Tired of niggling questions and feeling underpaid, they had left in 2008 to open Ballance Capital, one of the first full-service, one-stop cum-ex trading shops.

Dozens of German banks participated in cum-ex deals, too, gobbling up German taxpayer money at the same time they received a rescue package worth more than $500 billion.

Last year, the lawyer who testified anonymously at the Bonn trial described the culture of the cum-ex world to Oliver Schröm and Christian Salewski, two reporters on the German television show “Panorama,” under disguising makeup. It was a realm beyond morality, he said: all male, supremely arrogant, and guided by the conviction that the German state is an enemy and German taxpayers are suckers.

“That was the normal world to which we no longer belonged,” he told the reporters. “We looked out the window from up there, and we thought, ‘We’re the cleverest of all, geniuses, and you’re all stupid.’”

a former Merrill Lynch investment banker sat in a London restaurant near the Thames and described what had turned him into a whistle-blower. In the years after the financial crisis, he said, he noticed that a handful of colleagues on the company’s trading floor were using their personal mobile phones, a breach of company policy. All communication was supposed to be tracked and recorded. These guys were sending self-deleting texts on Snapchat.“Obviously, they were circumventing controls,”

Worried about the growing pileup of tax-withholding credits on the books, Frank Tibo, the bank’s chief tax officer, flew to London in May 2007. He spent the day grilling Mr. Mora

The complaint lays out, in painstaking detail, how the trades were confected, who executed them and which questions should be asked by investigators to uncover the “sham.” It states that Merrill Lynch earned hundreds of millions of dollars over the previous seven years from cum-ex trades.

“Anyone who stood in the way of this trade was swept aside, and those who enabled it were promoted,” the whistle-blower said in a follow-up phone call. “But it was widely regarded as insanity inside the bank for it to be extracting money from sovereign treasuries, particularly after the entire sector had been supported by the public purse.

American banks conducted their cum-ex trades overseas, rather than at home, out of fear, the whistle-blower said. Specifically, he mentioned a 2008 Senate investigation into “dividend tax abuse” that found it was depriving the Treasury of $100 billion every year. The report led to a ban on dividend arbitrage tied to stock in United States corporations.

But nothing prevented American bankers from conducting such trades with foreign companies on foreign soil.

German efforts to stamp out cum-ex with legislation, in 2007 and 2009, left holes through which certain types of financial players could still crawl. This included private pension plans in the United States, a niche financial product for wealthy people who want the kind of privacy, and exotic investment options, that Fidelity doesn’t offer.

Investors will have problems of their own. Many have said they had no idea how cum-ex traders returned such dazzling profits. That defense became less plausible in 2012, after the German government spent millions of dollars to buy 11 hard drives from industry insiders. The hard drives were filled with marketing fliers, written by bankers, who sold cum-ex with an antigovernment pitch.

“We learned that it was very common for these bankers to have conversations over coffee with clients about cum-ex,” said Norbert Walter-Borjans, a former minister of finance for North Rhine-Westphalia. “They would say, ‘If you have a problem with how your hard-earned money is being spent in taxes, we’ve got an idea for you.’”

Authorities across Europe are said to be waiting for a resolution of the Bonn trial to move ahead with their own. Many are livid that Germany didn’t alert them sooner about the perils of cum-ex. The failure, say lawyers, stems from a Europe-wide hypersensitivity about privacy, which is especially acute when it comes to taxes.

In 2012, soon after Germany shut down its cum-ex problem, a London trader began a cum-ex scheme that fleeced the Danish tax authority of $2 billion, officials there say. The trader, Sanjay Shah, who now lives in Dubai, denies wrongdoing but has never been shy about the source of his wealth.When he bought a $1.3 million yacht a few years ago, he found the perfect name: Cum-Ex.

he has seen mounting evidence that using a smartphone, or even hearing one ring or vibrate, produces a welter of distractions that makes it harder to concentrate on a difficult problem or job. The division of attention impedes reasoning and performance.

Another 2015 study, appearing in the Journal of Computer-Mediated Communication, showed that when people hear their phone ring but are unable to answer it, their blood pressure spikes, their pulse quickens, and their problem-solving skills decline.

The researchers recruited 520 undergraduates at UCSD and gave them two standard tests of intellectual acuity. One test gauged “available working-memory capacity,” a measure of how fully a person’s mind can focus on a particular task. The second assessed “fluid intelligence,” a person’s ability to interpret and solve an unfamiliar problem. The only variable in the experiment was the location of the subjects’ smartphones. Some of the students were asked to place their phones in front of them on their desks; others were told to stow their phones in their pockets or handbags; still others were required to leave their phones in a different room.

In both tests, the subjects whose phones were in view posted the worst scores, while those who left their phones in a different room did the best. The students who kept their phones in their pockets or bags came out in the middle. As the phone’s proximity increased, brainpower decreased.

In subsequent interviews, nearly all the participants said that their phones hadn’t been a distraction—that they hadn’t even thought about the devices during the experiment. They remained oblivious even as the phones disrupted their focus and thinking.

A second experiment conducted by the researchers produced similar results, while also revealing that the more heavily students relied on their phones in their everyday lives, the greater the cognitive penalty they suffered.

 Smartphones have become so entangled with our existence that, even when we’re not peering or pawing at them, they tug at our attention, diverting precious cognitive resources. Just suppressing the desire to check our phone, which we do routinely and subconsciously throughout the day, can debilitate our thinking.

Imagine combining a mailbox, a newspaper, a TV, a radio, a photo album, a public library and a boisterous party attended by everyone you know, and then compressing them all into a single, small, radiant object. That is what a smartphone represents to us. No wonder we can’t take our minds off it.

They found that students who didn’t bring their phones to the classroom scored a full letter-grade higher on a test of the material presented than those who brought their phones. It didn’t matter whether the students who had their phones used them or not: All of them scored equally poorly.

A study of nearly a hundred secondary schools in the U.K., published last year in the journal Labour Economics, found that when schools ban smartphones, students’ examination scores go up substantially, with the weakest students benefiting the most.

Social skills and relationships seem to suffer as well.

Because smartphones serve as constant reminders of all the friends we could be chatting with electronically, they pull at our minds when we’re talking with people in person, leaving our conversations shallower and less satisfying.

In a 2013 study conducted at the University of Essex in England, 142 participants were divided into pairs and asked to converse in private for ten minutes. Half talked with a phone in the room, half without a phone present. The subjects were then given tests of affinity, trust and empathy. “The mere presence of mobile phones,” the researchers reported in the Journal of Social and Personal Relationships, “inhibited the development of interpersonal closeness and trust” and diminished “the extent to which individuals felt empathy and understanding from their partners.”

The evidence that our phones can get inside our heads so forcefully is unsettling. It suggests that our thoughts and feelings, far from being sequestered in our skulls, can be skewed by external forces we’re not even aware o

 Scientists have long known that the brain is a monitoring system as well as a thinking system. Its attention is drawn toward any object that is new, intriguing or otherwise striking — that has, in the psychological jargon, “salience.”

even in the history of captivating media, the smartphone stands out. It is an attention magnet unlike any our minds have had to grapple with before. Because the phone is packed with so many forms of information and so many useful and entertaining functions, it acts as what Dr. Ward calls a “supernormal stimulus,” one that can “hijack” attention whenever it is part of our surroundings — and it is always part of our surroundings.

Not only do our phones shape our thoughts in deep and complicated ways, but the effects persist even when we aren’t using the devices. As the brain grows dependent on the technology, the research suggests, the intellect weakens.

The irony of the smartphone is that the qualities that make it so appealing to us — its constant connection to the net, its multiplicity of apps, its responsiveness, its portability — are the very ones that give it such sway over our minds.

Phone makers like Apple and Samsung and app writers like Facebook, Google and Snap design their products to consume as much of our attention as possible during every one of our waking hours

Social media apps were designed to exploit “a vulnerability in human psychology,” former Facebook president Sean Parker said in a recent interview. “[We] understood this consciously. And we did it anyway.”

A quarter-century ago, when we first started going online, we took it on faith that the web would make us smarter: More information would breed sharper thinking. We now know it’s not that simple.

As strange as it might seem, people’s knowledge and understanding may actually dwindle as gadgets grant them easier access to online data stores

In a seminal 2011 study published in Science, a team of researchers — led by the Columbia University psychologist Betsy Sparrow and including the late Harvard memory expert Daniel Wegner — had a group of volunteers read forty brief, factual statements (such as “The space shuttle Columbia disintegrated during re-entry over Texas in Feb. 2003”) and then type the statements into a computer. Half the people were told that the machine would save what they typed; half were told that the statements would be erased.

Afterward, the researchers asked the subjects to write down as many of the statements as they could remember. Those who believed that the facts had been recorded in the computer demonstrated much weaker recall than those who assumed the facts wouldn’t be stored. Anticipating that information would be readily available in digital form seemed to reduce the mental effort that people made to remember it

The researchers dubbed this phenomenon the “Google effect” and noted its broad implications: “Because search engines are continually available to us, we may often be in a state of not feeling we need to encode the information internally. When we need it, we will look it up.”

as the pioneering psychologist and philosopher William James said in an 1892 lecture, “the art of remembering is the art of thinking.”

Only by encoding information in our biological memory can we weave the rich intellectual associations that form the essence of personal knowledge and give rise to critical and conceptual thinking. No matter how much information swirls around us, the less well-stocked our memory, the less we have to think with.

As Dr. Wegner and Dr. Ward explained in a 2013 Scientific American article, when people call up information through their devices, they often end up suffering from delusions of intelligence. They feel as though “their own mental capacities” had generated the information, not their devices. “The advent of the ‘information age’ seems to have created a generation of people who feel they know more than ever before,” the scholars concluded, even though “they may know ever less about the world around them.”

That insight sheds light on society’s current gullibility crisis, in which people are all too quick to credit lies and half-truths spread through social media. If your phone has sapped your powers of discernment, you’ll believe anything it tells you.

Data, the novelist and critic Cynthia Ozick once wrote, is “memory without history.” Her observation points to the problem with allowing smartphones to commandeer our brains

When we constrict our capacity for reasoning and recall or transfer those skills to a gadget, we sacrifice our ability to turn information into knowledge. We get the data but lose the meaning

We need to give our minds more room to think. And that means putting some distance between ourselves and our phones.

Harvard Business School professor emerita Shoshana Zuboff argues in her new book that the Valley’s wealth and power are predicated on an insidious, essentially pathological form of private enterprise—what she calls “surveillance capitalism.” Pioneered by Google, perfected by Facebook, and now spreading throughout the economy, surveillance capitalism uses human life as its raw material. Our everyday experiences, distilled into data, have become a privately-owned business asset used to predict and mold our behavior, whether we’re shopping or socializing, working or voting.

By reengineering the economy and society to their own benefit, Google and Facebook are perverting capitalism in a way that undermines personal freedom and corrodes democracy.

Under the Fordist model of mass production and consumption that prevailed for much of the twentieth century, industrial capitalism achieved a relatively benign balance among the contending interests of business owners, workers, and consumers. Enlightened executives understood that good pay and decent working conditions would ensure a prosperous middle class eager to buy the goods and services their companies produced. It was the product itself — made by workers, sold by companies, bought by consumers — that tied the interests of capitalism’s participants together. Economic and social equilibrium was negotiated through the product.

By removing the tangible product from the center of commerce, surveillance capitalism upsets the equilibrium. Whenever we use free apps and online services, it’s often said, we become the products, our attention harvested and sold to advertisers

this truism gets it wrong. Surveillance capitalism’s real products, vaporous but immensely valuable, are predictions about our future behavior — what we’ll look at, where we’ll go, what we’ll buy, what opinions we’ll hold — that internet companies derive from our personal data and sell to businesses, political operatives, and other bidders.

Unlike financial derivatives, which they in some ways resemble, these new data derivatives draw their value, parasite-like, from human experience.To the Googles and Facebooks of the world, we are neither the customer nor the product. We are the source of what Silicon Valley technologists call “data exhaust” — the informational byproducts of online activity that become the inputs to prediction algorithms

internet companies operate in what Zuboff terms “extreme structural independence from people.” When databases displace goods as the engine of the economy, our own interests, as consumers but also as citizens, cease to be part of the negotiation. We are no longer one of the forces guiding the market’s invisible hand. We are the objects of surveillance and control.

that’s particularly true on the I.R.A., which has to build all these things in the real world.

we’re trying to do industrial physical transformation at a speed and scale unheralded in American history. This is bigger than anything we have done at this speed ever.

The money is beginning to move out the door now, but we’re on a clock. Climate change is not like some other issues where if you don’t solve it this year, it is exactly the same to solve it next year. This is an issue where every year you don’t solve it, the amount of greenhouse gases in the atmosphere builds, warming builds, the effects compound

Solve, frankly, isn’t the right word there because all we can do is abate, a lot of the problems now baked in. So how is it going, and who can actually walk us through that?

Robinson Meyer is the founding executive editor of heatmap.news

why do all these numbers differ so much? How big is this thing?

in electric vehicles and in the effort, kind of this dual effort in the law, to both encourage Americans to buy and use electric vehicles and then also to build a domestic manufacturing base for electric vehicles.

on both counts, the data’s really good on electric vehicles. And that’s where we’re getting the fastest response from industry and the clearest response from industry to the law.

ROBINSON MEYER: Factories are getting planned. Steel’s going in the ground. The financing for those factories is locked down. It seems like they’re definitely going to happen. They’re permitted. Companies are excited about them. Large Fortune 500 automakers are confidently and with certainty planning for an electric vehicle future, and they’re building the factories to do that in the United States. They’re also building the factories to do that not just in blue states. And so to some degree, we can see the political certainty for electric vehicles going forward.

in other parts of the law, partially due to just vagaries of how the law is being implemented, tax credits where the fine print hasn’t worked out yet, it’s too early to say whether the law is working and how it’s going and whether it’s going to accomplish its goal

EZRA KLEIN: I always find this very funny in a way. The Congressional Budget Office scored it. They thought it would make about $380 billion in climate investments over a decade. So then you have all these other analyses coming out.

But there’s actually this huge range of outcomes in between where the thing passes, and maybe what you wanted to have happen happens. Maybe it doesn’t. Implementation is where all this rubber meets the road

the Rhodium Group, which is a consulting firm, they think it could be as high as $522 billion, which is a big difference. Then there’s this Goldman Sachs estimate, which the administration loves, where they say they’re projecting $1.2 trillion in incentives —

ROBINSON MEYER: All the numbers differ because most of the important incentives, most of the important tax credits and subsidies in the I.R.A., are uncapped. There’s no limit to how much the government might spend on them. All that matters is that some private citizen or firm or organization come to the government and is like, hey, we did this. You said you’d give us money for it. Give us the money.

because of that, different banks have their own energy system models, their own models of the economy. Different research groups have their own models.

we know it’s going to be wrong because the Congressional Budget Office is actually quite constrained in how it can predict how these tax credits are taken up. And it’s constrained by the technology that’s out there in the country right now.

The C.B.O. can only look at the number of electrolyzers, kind of the existing hydrogen infrastructure in the country, and be like, well, they’re probably all going to use these tax credits. And so I think they said that there would be about $5 billion of take up for the hydrogen tax credits.

But sometimes money gets allocated, and then costs overrun, and there delays, and you can’t get the permits, and so on, and the thing never gets built

the fact that the estimates are going up is to them early evidence that this is going well. There is a lot of applications. People want the tax credits. They want to build these new factories, et cetera.

a huge fallacy that we make in policy all the time is assuming that once money is allocated for something, you get the thing you’re allocating the money for. Noah Smith, the economics writer, likes to call this checkism, that money equals stuff.

EZRA KLEIN: They do not want that, and not wanting that and putting every application through a level of scrutiny high enough to try and make sure you don’t have another one

I don’t think people think a lot about who is cutting these checks, but a lot of it is happening in this very obscure office of the Department of Energy, the Loan Program Office, which has gone from having $40 billion in lending authority, which is already a big boost over it not existing a couple decades ago, to $400 billion in loan authority,

the Loan Program Office as one of the best places we have data on how this is going right now and one of the offices that’s responded fastest to the I.R.A.

the Loan Program Office is basically the Department of Energy’s in-house bank, and it’s kind of the closest thing we have in the US to what exists in other countries, like Germany, which is a State development bank that funds projects that are eventually going to be profitable.

It has existed for some time. I mean, at first, it kind of was first to play after the Recovery Act of 2009. And in fact, early in its life, it gave a very important loan to Tesla. It gave this almost bridge loan to Tesla that helped Tesla build up manufacturing capacity, and it got Tesla to where it is today.

EZRA KLEIN: It’s because one of the questions I have about that office and that you see in some of the coverage of them is they’re very afraid of having another Solyndra.

Now, depending on other numbers, including the D.O.E., it’s potentially as high as $100 billion, but that’s because the whole thing about the I.R.A. is it’s meant to encourage the build-out of this hydrogen infrastructure.

EZRA KLEIN: I’m never that excited when I see a government loans program turning a profit because I think that tends to mean they’re not making risky enough loans. The point of the government should be to bear quite a bit of risk —

And to some degree, Ford now has to compete, and US automakers are trying to catch up with Chinese EV automakers. And its firms have EV battery technology especially, but just have kind of comprehensive understanding of the EV supply chain that no other countries’ companies have

ROBINSON MEYER: You’re absolutely right that this is the key question. They gave this $9.2 billion loan to Ford to build these EV battery plants in Kentucky and Tennessee. It’s the largest loan in the office’s history. It actually means that the investment in these factories is going to be entirely covered by the government, which is great for Ford and great for our build-out of EVs

And to some degree, I should say, one of the roles of L.P.O. and one of the roles of any kind of State development bank, right, is to loan to these big factory projects that, yes, may eventually be profitable, may, in fact, assuredly be profitable, but just aren’t there yet or need financing that the private market can’t provide. That being said, they have moved very slowly, I think.

And they feel like they’re moving quickly. They just got out new guidelines that are supposed to streamline a lot of this. Their core programs, they just redefined and streamlined in the name of speeding them up

However, so far, L.P.O. has been quite slow in getting out new loans

I want to say that the pressure they’re under is very real. Solyndra was a disaster for the Department of Energy. Whether that was fair or not fair, there’s a real fear that if you make a couple bad loans that go bad in a big way, you will destroy the political support for this program, and the money will be clawed back, a future Republican administration will wreck the office, whatever it might be. So this is not an easy call.

when you tell me they just made the biggest loan in their history to Ford, I’m not saying you shouldn’t lend any money to Ford, but when I think of what is the kind of company that cannot raise money on the capital markets, the one that comes to mind is not Ford

They have made loans to a number of more risky companies than Ford, but in addition to speed, do you think they are taking bets on the kinds of companies that need bets? It’s a little bit hard for me to believe that it would have been impossible for Ford to figure out how to finance factorie

ROBINSON MEYER: Now, I guess what I would say about that is that Ford is — let’s go back to why Solyndra failed, right? Solyndra failed because Chinese solar deluged the market. Now, why did Chinese solar deluge the market? Because there’s such support of Chinese financing from the state for massive solar factories and massive scale.

EZRA KLEIN: — the private market can’t. So that’s the meta question I’m asking here. In your view, because you’re tracking this much closer than I am, are they too much under the shadow of Solyndra? Are they being too cautious? Are they getting money out fast enough?

ROBINSON MEYER: I think that’s right; that basically, if we think the US should stay competitive and stay as close as it can and not even stay competitive, but catch up with Chinese companies, it is going to require large-scale state support of manufacturing.

EZRA KLEIN: OK, that’s fair. I will say, in general, there’s a constant thing you find reporting on government that people in government feel like they are moving very quickly

EZRA KLEIN: — given the procedural work they have to go through. And they often are moving very quickly compared to what has been done in that respect before, compared to what they have to get over. They are working weekends, they are working nights, and they are still not actually moving that quickly compared to what a VC firm can do or an investment bank or someone else who doesn’t have the weight of congressional oversight committees potentially calling you in and government procurement rules and all the rest of it.

ROBINSON MEYER: I think that’s a theme across the government’s implementation of the I.R.A. right now, is that generally the government feels like it’s moving as fast as it can. And if you look at the Department of Treasury, they feel like we are publishing — basically, the way that most of the I.R.A. subsidies work is that they will eventually be administered by the I.R.S., but first the Department of the Treasury has to write the guidebook for all these subsidies, right?

the law says there’s a very general kind of “here’s thousands of dollars for EVs under this circumstance.” Someone still has to go in and write all the fine print. The Department of Treasury is doing that right now for each tax credit, and they have to do that before anyone can claim that tax credit to the I.R.S. Treasury feels like it’s moving extremely quickly. It basically feels like it’s completely at capacity with these, and it’s sequenced these so it feels like it’s getting out the most important tax credits first.

Private industry feels like we need certainty. It’s almost a year since the law passed, and you haven’t gotten us the domestic content bonus. You haven’t gotten us the community solar bonus. You haven’t gotten us all these things yet.

a theme across the government right now is that the I.R.A. passed. Agencies have to write the regulations for all these tax credits. They feel like they’re moving very quickly, and yet companies feel like they’re not moving fast enough.

that’s how we get to this point where we’re 311 days out from the I.R.A. passing, and you’re like, well, has it made a big difference? And I’m like, well, frankly, wind and solar developers broadly don’t feel like they have the full understanding of all the subsidies they need yet to begin making the massive investments

I think it’s fair to say maybe the biggest bet on that is green hydrogen, if you’re looking in the bill.

We think it’s going to be an important tool in industry. It may be an important tool for storing energy in the power grid. It may be an important tool for anything that needs combustion.

ROBINSON MEYER: Yeah, absolutely. So green hydrogen — and let’s just actually talk about hydrogen broadly as this potential tool in the decarbonization tool kit.

It’s a molecule. It is a very light element, and you can burn it, but it’s not a fossil fuel. And a lot of the importance of hydrogen kind of comes back to that attribute of it.

So when we look at sectors of the economy that are going to be quite hard to decarbonize — and that’s because there is something about fossil fuels chemically that is essential to how that sector works either because they provide combustion heat and steelmaking or because fossil fuels are actually a chemical feedstock where the molecules in the fossil fuel are going into the product or because fossil fuels are so energy dense that you can carry a lot of energy while actually not carrying that much mass — any of those places, that’s where we look at hydrogen as going.

green hydrogen is something new, and the size of the bet is huge. So can you talk about first just what is green hydrogen? Because my understanding of it is spotty.

The I.R.A. is extremely generous — like extremely, extremely generous — in its hydrogen subsidies

The first is for what’s called blue hydrogen, which is hydrogen made from natural gas, where we then capture the carbon dioxide that was released from that process and pump it back into the ground. That’s one thing that’s subsidized. It’s basically subsidized as part of this broader set of packages targeted at carbon capture

green hydrogen, which is where we take water, use electrolyzers on it, basically zap it apart, take the hydrogen from the water, and then use that as a fue

The I.R.A. subsidies for green hydrogen specifically, which is the one with water and electricity, are so generous that relatively immediately, it’s going to have a negative cost to make green hydrogen. It will cost less than $0 to make green hydrogen. The government’s going to fully cover the cost of producing it.

That is intentional because what needs to happen now is that green hydrogen moves into places where we’re using natural gas, other places in the industrial economy, and it needs to be price competitive with those things, with natural gas, for instance. And so as it kind of is transported, it’s going to cost money

As you make the investment to replace the technology, it’s going to cost money. And so as the hydrogen moves through the system, it’s going to wind up being price competitive with natural gas, but the subsidies in the bill are so generous that hydrogen will cost less than $0 to make a kilogram of it

There seems to be a sense that hydrogen, green hydrogen, is something we sort of know how to make, but we don’t know how to make it cost competitive yet. We don’t know how to infuse it into all the processes that we need to be infused into. And so a place where the I.R.A. is trying to create a reality that does not yet exist is a reality where green hydrogen is widely used, we have to know how to use it, et cetera.

And they just seem to think we don’t. And so you need all these factories. You need all this innovation. Like, they have to create a whole innovation and supply chain almost from scratch. Is that right?

ROBINSON MEYER: That’s exactly right. There’s a great Department of Energy report that I would actually recommend anyone interested in this read called “The Liftoff Report for Clean Hydrogen.” They made it for a few other technologies. It’s a hundred-page book that’s basically how the D.O.E. believes we’re going to build out a clean hydrogen economy.

And, of course, that is policy in its own right because the D.O.E. is saying, here is the years we’re going to invest to have certain infrastructure come online. Here’s what we think we need. That’s kind of a signal to industry that everyone should plan around those years as well.

It’s a great book. It’s like the best piece of industrial policy I’ve actually seen from the government at all. But one of the points it makes is that you’re going to make green hydrogen. You’re then going to need to move it. You’re going to need to move it in a pipeline or maybe a truck or maybe in storage tanks that you then cart around.

Once it gets to a facility that uses green hydrogen, you’re going to need to store some green hydrogen there in storage tanks on site because you basically need kind of a backup supply in case your main supply fails. All of those things are going to add cost to hydrogen. And not only are they going to add cost, we don’t really know how to do them. We have very few pipelines that are hydrogen ready.

All of that investment needs to happen as a result to make the green hydrogen economy come alive. And why it’s so lavishly subsidized is to kind of fund all that downstream investment that’s eventually going to make the economy come true.

But a lot of what has to happen here, including once the money is given out, is that things we do know how to build get built, and they get built really fast, and they get built at this crazy scale.

So I’ve been reading this paper on what they call “The Greens’ Dilemma” by J.B. Ruhl and James Salzman, who also wrote this paper called “Old Green Laws, New Green Deal,” or something like that. And I think they get at the scale problem here really well.

“The largest solar facility currently online in the US is capable of generating 585 megawatts. To meet even a middle-road renewable energy scenario would require bringing online two new 400-megawatt solar power facilities, each taking up at least 2,000 acres of land every week for the next 30 years.”

And that’s just solar. We’re not talking wind there. We’re not talking any of the other stuff we’ve discussed here, transmission lines. Can we do that? Do we have that capacity?

ROBINSON MEYER: No, we do not. We absolutely do not. I think we’re going to build a ton of wind and solar. We do not right now have the system set up to use that much land to build that much new solar and wind by the time that we need to build it. I think it is partially because of permitting laws, and I think it’s also partially because right now there is no master plan

There’s no overarching strategic entity in the government that’s saying, how do we get from all these subsidies in the I.R.A. to net zero? What is our actual plan to get from where we are right now to where we’re emitting zero carbon as an economy? And without that function, no project is essential. No activity that we do absolutely needs to happen, and so therefore everything just kind of proceeds along at a convenient pace.

given the scale of what’s being attempted here, you might think that something the I.R.A. does is to have some entity in the government, as you’re saying, say, OK, we need this many solar farms. This is where we think we should put them. Let’s find some people to build them, or let’s build them ourselves.

what it actually does is there’s an office somewhere waiting for private companies to send in an application for a tax credit for solar that they say they’re going to build, and then we hope they build it

it’s an almost entirely passive process on the part of the government. Entirely would be going too far because I do think they talk to people, and they’re having conversations

the builder applies, not the government plans. Is that accurate?

ROBINSON MEYER: That’s correct. Yes.

ROBINSON MEYER: I think here’s what I would say, and this gets back to what do we want the I.R.A. to do and what are our expectations for the I.R.A

If the I.R.A. exists to build out a ton of green capacity and shift the political economy of the country toward being less dominated by fossil fuels and more dominated by the clean energy industry, frankly, then it is working

If the I.R.A. is meant to get us all the way to net zero, then it is not capable of that.

in 2022, right, we had no way to see how we were going to reduce emissions. We did not know if we were going to get a climate bill at all. Now, we have this really aggressive climate bill, and we’re like, oh, is this going to get us to net zero?

But getting to net zero was not even a possibility in 2022.

The issue is that the I.R.A. requires, ultimately, private actors to come forward and do these things. And as more and more renewables get onto the grid, almost mechanically, there’s going to be less interest in bringing the final pieces of decarbonized electricity infrastructure onto the grid as well.

EZRA KLEIN: Because the first things that get applied for are the ones that are more obviously profitable

The issue is when you talk to solar developers, they don’t see it like, “Am I going to make a ton of money, yes or no?” They see it like they have a capital stack, and they have certain incentives and certain ways to make money based off certain things they can do. And as more and more solar gets on the grid, building solar at all becomes less profitable

also, just generally, there’s less people willing to buy the solar.

as we get closer to a zero-carbon grid, there is this risk that basically less and less gets built because it will become less and less profitable

EZRA KLEIN: Let’s call that the last 20 percent risk

EZRA KLEIN: — or the last 40 percent. I mean, you can probably attach different numbers to that

ROBINSON MEYER: Permitting is the primary thing that is going to hold back any construction basically, especially out West,