Altman om AI-hot mot jobben: "Inte det minsta rädd"
Open AI:s vd Sam Altman räds inte konsekvenserna på arbetsmarknaden som AI-revolutionen leder till. Det sa han på Wall Street Journals årliga konferens Tech Live i veckan: – Varje teknologisk revolution får följder för arbetsmarknaden. Jag är inte det minsta rädd för det. Under mötet samlades världens techledare – från bland annat Meta och Arm – för att diskutera de snabba förändringar som AI medför för näringslivet och politiken. Tidningen listar här några höjdpunkter från samtalen. OpenAI’s Sam Altman, Meta’s Chris Cox and others spoke at WSJ’s annual event. By WSJ Staff
The Wall Street Journal, 18 October 2023 Tech leaders convened on The Wall Street Journal’s annual Tech Live conference this week, where discussions focused on the fast-paced changes wrought by artificial intelligence across business, technology and policy-making. Here are some highlights from interviews. AI has been a central topic this year, as its impact on business and society is hotly debated. Sam Altman, chief executive of OpenAI, said the two things that will matter the most over the next few decades are abundant and inexpensive intelligence, and abundant and cheap energy. OpenAI is working to make ChatGPT cheaper and faster, so that it can be more broadly accessible. “If we can get these two things done in the world then it’s almost, like, difficult to imagine how much else we could do,” he said. Mustafa Suleyman, CEO of Inflection AI, said the race to build AI chatbots is reminiscent of the rush to build websites at the dawn of the internet or apps after the advent of smartphones. “This is really just the beginning of a complete transformation in the way we interact with computers,” said Suleyman, whose company makes a ChatGPT rival called Pi, short for personal intelligence. Meta is also optimistic about the impact of AI. “One of the most profoundly impactful applications in the near term for AI is helping businesses be more effective,” said Meta Platforms Chief Product Officer Chis Cox. Meta last month unveiled its own AI chatbots based on celebrities such as Naomi Osaka and Snoop Dogg. Cox said Meta is making it clear these characters aren’t the real people. “Having products that experiment with what is possible is great, but having anything that doesn’t make clear to people what is going on is a problem,” he said. Consumers are going to gravitate to TikTok, ChatGPT and other applications powered by generative artificial intelligence, instead of using traditional search engines, said Michael Wolf, co-founder and CEO of consulting firm Activate. He predicts that domination within the $100 billion search industry is “up for grabs,” adding that the rise of open-source AI models is paving a pathway for smaller entrants to meaningfully compete with large, established companies. Professionals from physicians to writers have been fearing that AI will entirely replace some jobs. “Every technological revolution affects the job market. I’m not afraid of that at all,” said OpenAI’s Altman. “That’s the way of progress. And we’ll find new and better jobs.” Still, it’s not going to be a seamless process. “The thing that I think we do need to confront as a society is the speed at which this is going to happen,” he added. Adam Wenchel, chief executive of AI company Arthur, took a more sanguine view of the job impact from AI than some other panelists at Tech Live. “These systems are going to roll out over time, very gradually, people are going to adapt to them and it’s going to be OK,” he said. Indeed, companies are still determining how to implement new AI technology. “Even at the highest levels, we’re still trying to figure out what does all of this mean to our business model,” said Vince Marin, chief information officer of law firm Sidley Austin. Charles Sims, chief technology officer at United Talent Agency, said AI makes it more important for people to have generalized skill sets that enable them to adapt as technology replaces specific specialties. “If you’re talking to a college student today, it’s about generalization, it’s about trying to learn as many things as you can,” he said. Elise Smith, CEO of Praxis Labs, said it is critical to involve the next generation of workers in discussions about how to use technology: “They want to be brought in and brought along on the journey,” she said. “They want to be doing the innovation day, the hackathon, where they’re getting to give ideas around how AI can transform their business.” Adobe’s president of digital media business, David Wadhwani, said that despite fears, he sees artificial intelligence as a tool that will boost employment rather than put people out of jobs. Tools like Adobe’s Firefly, which can generate images and logos, allow more people to become creative professionals, he said. “We will have creative professionals being more productive than ever before and more creative professionals in the world,” he said. Arm CEO Rene Haas said the chip company is using artificial intelligence to help in some of the areas where they struggle to hire enough talent, such as with debugging and testing chips. But he said the semiconductor industry faces some challenges in its role powering artificial intelligence. He described a future when energy shortages could constrain AI advancement, and a shortage of talent could limit production of semiconductors. “The kind of people we are looking to hire are hard to find. We are looking for really expert engineers,” Haas said. Investors are weighing whether it’s too late to get into AI. “Most investments in AI today—venture investments—will lose money,” said venture capitalist Vinod Khosla. Khosla, who founded Khosla Ventures almost two decades ago, said AI investing had entered a hype cycle, and only highly disciplined investors will reap the benefits of the transformative technology. The buzzy new technology has generated significant concerns, though. “We’ve got a fierce task ahead of us to figure out what are these downsides and discover, understand them, and build the tools to mitigate them,” said OpenAI Chief Technology Officer Mira Murati. For instance, sometimes chatbots confidently espouse information that doesn’t seem to be justified by its training data. “We’ve made a ton of progress on the hallucination issue with GPT-4, but we’re not where we need to be. We’re sort of on the right track.” Murati said. OpenAI is continuing to use techniques including reinforcement learning with human feedback to reduce the number of times that its model makes up information. It is also working on technology that can help detect the provenance of an image, Murati said. Suleyman, CEO of the company behind Pi, said another problem is that Pi and other AI chatbots aren’t designed to doubt themselves, which makes it hard to know when they’re wrong. He suggested that a possible safeguard for users would be to have responses ranked by their accuracy. “This skill of uncertainty estimation is a critical part of intelligence and actually key to making them reliable,” he said. Suleyman said he and his peers are also discussing the potential risk of AI interfering with next year’s U.S. presidential elections, and he hopes to build parameters that will prevent Pi from recommending political candidates. One of the leading risks to the development of the nation’s AI sector is the imbalance between public and private sector investment in what will soon be a technology as ubiquitous as electricity, said Fei-Fei Li, co-director of the Stanford Institute for Human-Centered Artificial Intelligence and a former vice president at Google. U.S. government investment and incentives should at least match the U.S.’s investment in space exploration decades ago with the National Aeronautics and Space Administration. “This technology is as big or even bigger than the space technology,” Li said. “We cannot just leave it to the private sector.” Li said the Food and Drug Administration, Environmental Protection Agency and other government agencies should urgently take a role in regulating AI. “It is very hard to imagine one ring that rules them all,” Li said. Roblox CEO David Baszucki said the gaming company is treading carefully when it comes to training artificial intelligence models, and isn’t harvesting anyone’s code without permission. “That’s a big societal discussion right now,” he said. The energy costs associated with powering artificial-intelligence programs have also been a concern for climate advocates. But former Meta Chief Technology Officer Mike Schroepfer, whose new firm Gigascale Capital invests in climate-focused companies, said AI will save energy in other ways. “It will be a large consumption of power, but you also have to think of the replacement costs,” he said, referring to efficiencies that AI is expected to provide. Cryptocurrency is another area in tech rife with pitfalls. Anthony Scaramucci, founder of SkyBridge Capital, said he should have been more wary of Sam Bankman-Fried, the founder of collapsed cryptocurrency exchange FTX, who is on trial in New York facing fraud and conspiracy charges. “I took the aura of all of that too seriously, and I probably should’ve been more of a skeptic,” Scaramucci said. “He committed a crime and I believe he has to go to jail for a very long time.” Executives and advocates also highlighted the risks of social media, especially for young people. Larissa May, founder and executive director, of nonprofit #HalfTheStory, said kids are spending an average of eight hours a day on their devices. “We better be looking at the place where they’re spending more time than anywhere else in their life, including sleep,” May said. Social-media companies should think about more than how much time young people are spending on social media app—they should find ways to measure whether apps are supporting or hurting them. “It’s so much bigger than just a dollar sign,” May said. Comedian and creator Elsa Majimbo said social media can be too negative. She called X, the social-media platform formerly known as Twitter, “a soft dark web” that should have a minimum age requirement of 18. Award-winning musician John Legend, who is launching his first-ever tech startup, agreed that AI has its limits. He said computer-generated music won’t replace songwriters, in part because audiences like the artists’ stories behind their music. “There’s just something that’s still so human about music, songwriting and that interaction we have with our audience,” he said. Arnold Schwarzenegger said he is aware of AI use in Hollywood and has heard a fake version of his voice. Whether or not his likeness will be used by AI in the future is a point his children will have to negotiate, the 76-year-old actor said. “I will not be around, even though I want to live forever,” he said. Even venture capitalist Khosla has tried his hand at it. When his daughter got married earlier this year, he asked ChatGPT to turn a speech he wrote into rap lyrics and then turned those lyrics into a song through an AI startup called Splash. He blared the song over speakers. “It extended my capabilities,” he said. “It meant a lot to me.” In addition to the uncertainties of AI, technology leaders are also now dealing with critical questions regarding the impact of geopolitics on the sector. Venture capitalist Khosla said that winning the race to develop advanced AI would give the U.S. an economic and political advantage over China. “I think the world’s political system—what influences Southeast Asia, Africa, Latin America—is at stake,” he said. “Western values are at stake in this technology battle, so we should do whatever we can to win this battle and beat China at it.” Khosla also warned against making the code behind advanced AI models available to the public, which some technologists have championed as a way to bolster the technology’s development. “You don’t open-source the Manhattan Project,” he said, referencing America’s clandestine efforts to build an atomic bomb during World War II. The war between Hamas and Israel, which has been a tech hub for years, was also a focal point of TechLive this year. Palmer Luckey, founder of defense technology company Anduril Industries, said U.S. corporate chiefs should be more vocal in their support for Israel. “It reflects very poorly on our billionaire class that you aren’t seeing a whole-of-country effort to become involved and to speak up about these issues. That you are seeing hedging on the condemnation of Hamas for fear of saying the wrong thing either in the court of public opinion or because it hurts their business interests,” Luckey said. Charlie Shrem, general partner of Druid Ventures, was asked about the use of cryptocurrency by Hamas to fund its attacks in Israel. He said it is “a really sad thing to see something that we were all involved in creating early on become used in these negative ways.” When it came to domestic politics, Schwarzenegger said aging leaders of the Republican and Democratic parties should step aside and make room for a new generation. The former California governor alluded to recent instances of Senate Minority Leader Mitch McConnell freezing and falling silent, and said people in that position should “start thinking about stepping aside and letting a newer generation step in and fill the vacuum.” Sarah E. Needleman, Annie Gasparro, Berber Jin, Mengqi Sun, Georgia Wells, Sarah Krouse, Heather Somerville, Tom Dotan and Deepa Seetharaman contributed to this article.
Open AI:s vd Sam Altman räds inte konsekvenserna på arbetsmarknaden som AI-revolutionen leder till. Det sa han på Wall Street Journals årliga konferens Tech Live i veckan: – Varje teknologisk revolution får följder för arbetsmarknaden. Jag är inte det minsta rädd för det. Under mötet samlades världens techledare – från bland annat Meta och Arm – för att diskutera de snabba förändringar som AI medför för näringslivet och politiken. Tidningen listar här några höjdpunkter från samtalen. OpenAI’s Sam Altman, Meta’s Chris Cox and others spoke at WSJ’s annual event. By WSJ Staff
The Wall Street Journal, 18 October 2023 Tech leaders convened on The Wall Street Journal’s annual Tech Live conference this week, where discussions focused on the fast-paced changes wrought by artificial intelligence across business, technology and policy-making. Here are some highlights from interviews. AI has been a central topic this year, as its impact on business and society is hotly debated. Sam Altman, chief executive of OpenAI, said the two things that will matter the most over the next few decades are abundant and inexpensive intelligence, and abundant and cheap energy. OpenAI is working to make ChatGPT cheaper and faster, so that it can be more broadly accessible. “If we can get these two things done in the world then it’s almost, like, difficult to imagine how much else we could do,” he said. Mustafa Suleyman, CEO of Inflection AI, said the race to build AI chatbots is reminiscent of the rush to build websites at the dawn of the internet or apps after the advent of smartphones. “This is really just the beginning of a complete transformation in the way we interact with computers,” said Suleyman, whose company makes a ChatGPT rival called Pi, short for personal intelligence. Meta is also optimistic about the impact of AI. “One of the most profoundly impactful applications in the near term for AI is helping businesses be more effective,” said Meta Platforms Chief Product Officer Chis Cox. Meta last month unveiled its own AI chatbots based on celebrities such as Naomi Osaka and Snoop Dogg. Cox said Meta is making it clear these characters aren’t the real people. “Having products that experiment with what is possible is great, but having anything that doesn’t make clear to people what is going on is a problem,” he said. Consumers are going to gravitate to TikTok, ChatGPT and other applications powered by generative artificial intelligence, instead of using traditional search engines, said Michael Wolf, co-founder and CEO of consulting firm Activate. He predicts that domination within the $100 billion search industry is “up for grabs,” adding that the rise of open-source AI models is paving a pathway for smaller entrants to meaningfully compete with large, established companies. Professionals from physicians to writers have been fearing that AI will entirely replace some jobs. “Every technological revolution affects the job market. I’m not afraid of that at all,” said OpenAI’s Altman. “That’s the way of progress. And we’ll find new and better jobs.” Still, it’s not going to be a seamless process. “The thing that I think we do need to confront as a society is the speed at which this is going to happen,” he added. Adam Wenchel, chief executive of AI company Arthur, took a more sanguine view of the job impact from AI than some other panelists at Tech Live. “These systems are going to roll out over time, very gradually, people are going to adapt to them and it’s going to be OK,” he said. Indeed, companies are still determining how to implement new AI technology. “Even at the highest levels, we’re still trying to figure out what does all of this mean to our business model,” said Vince Marin, chief information officer of law firm Sidley Austin. Charles Sims, chief technology officer at United Talent Agency, said AI makes it more important for people to have generalized skill sets that enable them to adapt as technology replaces specific specialties. “If you’re talking to a college student today, it’s about generalization, it’s about trying to learn as many things as you can,” he said. Elise Smith, CEO of Praxis Labs, said it is critical to involve the next generation of workers in discussions about how to use technology: “They want to be brought in and brought along on the journey,” she said. “They want to be doing the innovation day, the hackathon, where they’re getting to give ideas around how AI can transform their business.” Adobe’s president of digital media business, David Wadhwani, said that despite fears, he sees artificial intelligence as a tool that will boost employment rather than put people out of jobs. Tools like Adobe’s Firefly, which can generate images and logos, allow more people to become creative professionals, he said. “We will have creative professionals being more productive than ever before and more creative professionals in the world,” he said. Arm CEO Rene Haas said the chip company is using artificial intelligence to help in some of the areas where they struggle to hire enough talent, such as with debugging and testing chips. But he said the semiconductor industry faces some challenges in its role powering artificial intelligence. He described a future when energy shortages could constrain AI advancement, and a shortage of talent could limit production of semiconductors. “The kind of people we are looking to hire are hard to find. We are looking for really expert engineers,” Haas said. Investors are weighing whether it’s too late to get into AI. “Most investments in AI today—venture investments—will lose money,” said venture capitalist Vinod Khosla. Khosla, who founded Khosla Ventures almost two decades ago, said AI investing had entered a hype cycle, and only highly disciplined investors will reap the benefits of the transformative technology. The buzzy new technology has generated significant concerns, though. “We’ve got a fierce task ahead of us to figure out what are these downsides and discover, understand them, and build the tools to mitigate them,” said OpenAI Chief Technology Officer Mira Murati. For instance, sometimes chatbots confidently espouse information that doesn’t seem to be justified by its training data. “We’ve made a ton of progress on the hallucination issue with GPT-4, but we’re not where we need to be. We’re sort of on the right track.” Murati said. OpenAI is continuing to use techniques including reinforcement learning with human feedback to reduce the number of times that its model makes up information. It is also working on technology that can help detect the provenance of an image, Murati said. Suleyman, CEO of the company behind Pi, said another problem is that Pi and other AI chatbots aren’t designed to doubt themselves, which makes it hard to know when they’re wrong. He suggested that a possible safeguard for users would be to have responses ranked by their accuracy. “This skill of uncertainty estimation is a critical part of intelligence and actually key to making them reliable,” he said. Suleyman said he and his peers are also discussing the potential risk of AI interfering with next year’s U.S. presidential elections, and he hopes to build parameters that will prevent Pi from recommending political candidates. One of the leading risks to the development of the nation’s AI sector is the imbalance between public and private sector investment in what will soon be a technology as ubiquitous as electricity, said Fei-Fei Li, co-director of the Stanford Institute for Human-Centered Artificial Intelligence and a former vice president at Google. U.S. government investment and incentives should at least match the U.S.’s investment in space exploration decades ago with the National Aeronautics and Space Administration. “This technology is as big or even bigger than the space technology,” Li said. “We cannot just leave it to the private sector.” Li said the Food and Drug Administration, Environmental Protection Agency and other government agencies should urgently take a role in regulating AI. “It is very hard to imagine one ring that rules them all,” Li said. Roblox CEO David Baszucki said the gaming company is treading carefully when it comes to training artificial intelligence models, and isn’t harvesting anyone’s code without permission. “That’s a big societal discussion right now,” he said. The energy costs associated with powering artificial-intelligence programs have also been a concern for climate advocates. But former Meta Chief Technology Officer Mike Schroepfer, whose new firm Gigascale Capital invests in climate-focused companies, said AI will save energy in other ways. “It will be a large consumption of power, but you also have to think of the replacement costs,” he said, referring to efficiencies that AI is expected to provide. Cryptocurrency is another area in tech rife with pitfalls. Anthony Scaramucci, founder of SkyBridge Capital, said he should have been more wary of Sam Bankman-Fried, the founder of collapsed cryptocurrency exchange FTX, who is on trial in New York facing fraud and conspiracy charges. “I took the aura of all of that too seriously, and I probably should’ve been more of a skeptic,” Scaramucci said. “He committed a crime and I believe he has to go to jail for a very long time.” Executives and advocates also highlighted the risks of social media, especially for young people. Larissa May, founder and executive director, of nonprofit #HalfTheStory, said kids are spending an average of eight hours a day on their devices. “We better be looking at the place where they’re spending more time than anywhere else in their life, including sleep,” May said. Social-media companies should think about more than how much time young people are spending on social media app—they should find ways to measure whether apps are supporting or hurting them. “It’s so much bigger than just a dollar sign,” May said. Comedian and creator Elsa Majimbo said social media can be too negative. She called X, the social-media platform formerly known as Twitter, “a soft dark web” that should have a minimum age requirement of 18. Award-winning musician John Legend, who is launching his first-ever tech startup, agreed that AI has its limits. He said computer-generated music won’t replace songwriters, in part because audiences like the artists’ stories behind their music. “There’s just something that’s still so human about music, songwriting and that interaction we have with our audience,” he said. Arnold Schwarzenegger said he is aware of AI use in Hollywood and has heard a fake version of his voice. Whether or not his likeness will be used by AI in the future is a point his children will have to negotiate, the 76-year-old actor said. “I will not be around, even though I want to live forever,” he said. Even venture capitalist Khosla has tried his hand at it. When his daughter got married earlier this year, he asked ChatGPT to turn a speech he wrote into rap lyrics and then turned those lyrics into a song through an AI startup called Splash. He blared the song over speakers. “It extended my capabilities,” he said. “It meant a lot to me.” In addition to the uncertainties of AI, technology leaders are also now dealing with critical questions regarding the impact of geopolitics on the sector. Venture capitalist Khosla said that winning the race to develop advanced AI would give the U.S. an economic and political advantage over China. “I think the world’s political system—what influences Southeast Asia, Africa, Latin America—is at stake,” he said. “Western values are at stake in this technology battle, so we should do whatever we can to win this battle and beat China at it.” Khosla also warned against making the code behind advanced AI models available to the public, which some technologists have championed as a way to bolster the technology’s development. “You don’t open-source the Manhattan Project,” he said, referencing America’s clandestine efforts to build an atomic bomb during World War II. The war between Hamas and Israel, which has been a tech hub for years, was also a focal point of TechLive this year. Palmer Luckey, founder of defense technology company Anduril Industries, said U.S. corporate chiefs should be more vocal in their support for Israel. “It reflects very poorly on our billionaire class that you aren’t seeing a whole-of-country effort to become involved and to speak up about these issues. That you are seeing hedging on the condemnation of Hamas for fear of saying the wrong thing either in the court of public opinion or because it hurts their business interests,” Luckey said. Charlie Shrem, general partner of Druid Ventures, was asked about the use of cryptocurrency by Hamas to fund its attacks in Israel. He said it is “a really sad thing to see something that we were all involved in creating early on become used in these negative ways.” When it came to domestic politics, Schwarzenegger said aging leaders of the Republican and Democratic parties should step aside and make room for a new generation. The former California governor alluded to recent instances of Senate Minority Leader Mitch McConnell freezing and falling silent, and said people in that position should “start thinking about stepping aside and letting a newer generation step in and fill the vacuum.” Sarah E. Needleman, Annie Gasparro, Berber Jin, Mengqi Sun, Georgia Wells, Sarah Krouse, Heather Somerville, Tom Dotan and Deepa Seetharaman contributed to this article.