This is a publication of The Atlantic Daily, a newsletter that guides you through the best stories of the day, helps you discover new ideas and recommends the best in culture. Register here.
Derek Thompson told me in December that recent advances in generative AI, like the DALL-E image generator and ChatGPT's extended language model, are "potentially like the launch of the iPhone in 2007 or the invention of desktop computers". . Here are the latest AI developments to watch in the weeks and months ahead.
But first, three new stories from The Atlantic .
Hit cars
Investors invest in artificial intelligence.
According to market data firm Pitchbook, investors poured at least $1.37 billion into generative AI companies through 78 deals last year, almost as much as they invested in the past few years. last five years.
Microsoft in particular has taken a big leap forward. In 2019, the company invested $3 billion in OpenAI, which developed DALL-E and ChatGPT, and is reportedly in talks to invest another $10 billion. Microsoft has acquired an exclusive license for part of its OpenAI technology and is working on a new version of its Bing search engine that combines OpenAI with a tool like ChatGPT.
Schools are concerned about academic integrity.
How will these tools change our lives? Like Derek told me the other day. "We don't know. The architects of these technologies know very little. But it's so exciting to play and the technology is improving so fast that we have to take it very seriously, as if it's something that cannot be avoided."
Some universities are modifying their courses to minimize the risk of students submitting essays created by an AI tool. And most likely they will be working on more powerful tools soon; OpenAI plans to release GPT-4, which can generate text better than current versions. A 22-year-old computer science student has now developed a program that recognizes if a text has been written by a bot.
Maybe it's time to start thinking about deepfakes again .
You may remember back in 2018 as prime minister, when media and disinformation pundits panicked over the rise of fake and fake videos. (In a famous example from BuzzFeed , Barack Obama said, "President Trump is a complete idiot.")
While this panic remains a breakthrough in generative AI, "experts fear a deep pseudo-apocalypse is looming on the horizon," our editor Matteo Wong reported last month. As AI-powered media becomes more sophisticated, these experts say the internet will be flooded with fake video and audio messages conveying false information in years to come.
Tools like ChatGPT may not be as smart as they seem...
Last week, them Atlantic Associate Writer Ian Bogost brought some skepticism to the AI debate. "ChatGPT knows nothing , but instead creates compositions that simulate knowledge through a probabilistic framework," Bogost wrote. "The news of this surprise makes it increasingly clear that ChatGPT is less a magic wish-granting machine and more a feedback exchange partner." Could all this investment in technology be the result of a bad idea, you ask?
But don't expect the hype to die down anytime soon.
Some have asked if we see Crypto 2.0. a fancy new technology attracts media attention and investor money, only for some of the top companies built around it to crash dramatically. But cryptocurrency is not a good model for thinking about AI, Derek told me. “Crypto used to be worthless money,” he said, while tools like ChatGPT are now “a free utility.” Generative AI, he said, "is clearly a thing , even if it wants to pretend to be a toy for now."
Moreover, AI is already succeeding in ways that cryptocurrency never could, Derek noted. Although you'll hear some people use AI as a generic term, the technology that's emerging right now is generative . Tools that can create new content such as text or images. We have all been living with artificial intelligence for years. "Go to Instagram. Why do some stories or posts stand out from others? Because of artificial intelligence, Derek says, if you use the most popular social networking apps, you live in a world created by intelligence artificial.
With reference to:
News of the day
- The number of deaths in China exceeded the number of births for the first time in six decades, the government said.
- Swedish climate activist Greta Thunberg has been arrested by German police while protesting against the expansion of a coal mine.
- Tampa Bay Buccaneers wide receiver Russell Gage was hospitalized after suffering a concussion in Monday's playoff game against the Dallas Cowboys.
Deliveries:
See all our newsletters here.
read at night
American religion is not dead yet
By Wendy Cage and Elan Babchuk
Walk down Main Street in almost any major city in the country and you'll pass a church with more homes for sale and a sluggish housing market. This phenomenon will not change any time soon. According to the author of the 2021 report on the future of religion in America, 30% of churches are unlikely to survive in the next 20 years. Add to that declining participation and membership, and you might conclude that American religion is doomed to extinction.
But the old measures of success – attendance and membership – or colloquially “butts, budgets and buildings” – no longer reflect the state of American religion. While participation in traditional religious institutions (churches, synagogues, mosques, schools, etc.) has declined, signs of life are visible elsewhere: in conversations with religious leaders, in communities started online, but also in social justice groups, they start private lives. Linked and rooted in a common faith.
Read the full article.
Over the Atlantic
A cultural break
Read it. Janet Malcolm's posthumous memoir, Still Pictures , criticizes the idea of memoirs.
The appearance. Our reviewer writes that the Netflix adaptation of Elena Ferrante's The Fake Life is incredibly slow at times, but it's also surprising that nothing happened after Mad Men .
Play our daily crossword game.
PS:
If you want to dive into the AI rabbit hole, I recommend the Read Max newsletter from technical author Max Read. Read is leading a project to understand how we should think about AI, and last week he listed seven thought-provoking and thought-provoking questions he uses to guide his research, including "Why AI?". previous technological advances have not yielded as much. To mix together? and "Is artificial intelligence dumb?"
-Isabelle