The recent world chess championship saw Magnus Carlsen defend his title against Fabiano Caruana. But it was not a contest between the two strongest chess players on the planet, only the strongest humans. Soon after Garry Kasparov, the former world chess champion, lost his re-match against IBM’s Deep Blue in 1997, the short window of human-machine chess competition was shut forever. Unlike humans, machines keep getting faster, and today a smartphone chess app can be stronger than Deep Blue.

In the late 19th century, Alfred Binet hoped that understanding why certain people stood out at chess would unlock secrets of human thought. Sixty years later, Alan Turing wondered whether this sort of ability represents an essential difference between the potentialities of the machine and the mind. Much as airplanes don’t flap their wings like birds, machines don’t generate chess moves like humans do. Early programs that attempted it were weak.

But now things are different. Based on a common game-playing algorithm(算法), AlphaZero incorporates deep learning and other AI techniques to play against itself to generate its own chess knowledge.   AlphaZerostarts   out   knowing   only   the   rules   of   chess,   with   no   inserted   human strategies(策略). In just a few hours, it plays more games against itself than have been recorded in human chess history. It teaches itself the best way to play, reevaluating such fundamental concepts as the relative values of the pieces.It quickly becomes strong enough to defeat the best chess-players in the world, winning 28, drawing 72, and losing none in a victory over Stockfish, one of the strongest chess engines.

The conventional wisdom is that machines would approach perfection with endlessstrategies, usually leading to drawn games. But AlphaZero prefers positions that look risky and aggressive, and it programs itself, which allows it to outclass the world’s top traditional program despite calculating far fewer positions per second. It’s the example of the cliché, “work smarter, not harder.”

AlphaZero shows us that machines can be the experts, not merely expert tools. It’s not going to put chess coaches out of business just yet. But the knowledge it generates is information we can all learn from. AlphaZero is surpassing us in a profound and useful way.

Machine learning systems aren’t perfect, even at a closed system like chess. There will be cases where an AI will fail to detect exceptions to their rules. Therefore, we must work together, to combine our strengths. Instead of being angry against them, it’s better if we’re all on the same side.

Archeology （考古学）isn't the dusty science it was a generation ago. New technologies that once seemed out of sci-fi are now locating buried traces of buildings and revealing the ruins of cities.

For more than a decade, Sarah Parcak and her team have been on the front line of this revolution. They use satellite images to find and explore ancient sites around the globe. Now they're about to take on a new challenge as they focus Global Xplorer citizen-science project on India.

In 2017, Parcak launched an online platform, called Global Xplorer, to crowd source （群众外包）the initial assessment of satellite images for signs of cultures from long ago. Anyone in the world with a computer and Internet access could help discover and protect remains of Peru's rich cultural heritage. The results have been surprising. About 80,000 participants from a hundred countries have identified 19,000 sites that were not in Peru's database. The platform for Peru is still running.

If all goes well, the work in India could last for years. "India has had relatively little archaeological work done," Parcak says. Also, the full extent of India's archaeological work has never been mapped completely. Parcak expects her project to make up that. ^“Wherever we end up going, the crowd's going to be able to see extraordinary things," Parcak says.

Thirty six India's cultural heritage relics are already listed as UNESCO World Heritage Sites. Parcak thinks there could be tens of thousands of as yet unknown sites mapped as part of this project. The discoveries promise to be amazing across the land that has seen a parade of cultures come and go.

In the future, she hopes other countries will contact her to launch their own satellite surveys. The possibilities are huge. Parcak estimates that there are at least 12 million potential archaeological sites yet to be discovered. That means the sky is the limit for her project now that it has gotten off the ground successfully.

These days, we wonder a lot about why social networks go bad, full of terrible behavior. But it’s equally illuminating (有启发性的) to ask about the ones that work well. These communities share one characteristic: They’re small. Generally they have only a few hundred members, or maybe a couple thousand if they’re really popular.

And smallness makes all the difference. First, these groups have a sense of cohesion (凝聚性). The members have joined specifically to talk to people with whom they share an enthusiasm. That creates a type of social glue, a context and a mutual respect that can’t exist on a highly public site like Twitter, where anyone can crash any public conversation.

More important, small groups typically have people who work to keep interactions polite. Sometimes this will be the organizer or an active, long-term participant. They’ll greet newcomers to make them feel welcome, draw out quiet people and solve conflict when they see it emerge. But what’s crucial is that these key members model good behavior, illustrating by example the community’s best standards. The internet thinkers Heather Gold, Kevin Marks and Deb Schultz put a name to this: “tummeling,” after the Yiddish “tummeler,” who keeps a party going.

None of these positive elements can exist in a massive, public social network, where millions of people can get into each other’s spaces. That’s why perhaps the single biggest problem facing social media is that our mainstream networks are too interested in scale. But scale breaks social relations.

Is there any way to approach this problem? I’ve never heard of any simple solution. Strict antitrust laws for the big networks would be useful, to encourage competition among networks. But this likely wouldn’t fully solve the problem of scale, since many users fancy scale too. Longing for massive, global audiences, they will crowd into whichever site offers the hugest. Many of the well-known solutions to social media might help, but all ignore the biggest problem of all: Bigness itself.