AI arms race to build ‘digital god’

In today’s episode of decoder, we will try to find a “digital god.” I felt like we’ve been doing this for a long time, let’s just follow it. Can we create an artificial intelligence so powerful that it changes the world and answers all our questions? The AI ​​industry has decided that the answer is yes.

In September, Sam Altman of OpenAI published a blog post claiming that we will have intelligent AI in “a few thousand days.” And earlier this month, Dario Amodei, CEO of OpenAI competitor Anthropic, published a 14,000-word document laying out exactly what he thinks such a system will be able to do when it arrives, which he says could be as soon as 2026.

What’s striking is that the ideas put forward in both spaces are very similar – both promise AI of incredible intelligence that will bring about huge improvements in work, science and healthcare, as well as in democracy and prosperity. Digital god, baby.

But while the ideas are similar, the companies are, in many ways, clearly at odds: Anthropic is the first story of the OpenAI revolution. Dario and a team of fellow researchers left OpenAI in 2021 after concerns about its commercial growth and security approach, and formed Anthropic to become a safe, slow-moving AI company. And the emphasis was actually on safety until recently; just last year, great New York Times A company profile called it a “white hot center for AI doomerism.”

But the launch of ChatGPT, and subsequent AI production, set off a massive technological arms race, and now, Anthropic is in the game just like everyone else. It took billions in funding, mostly from Amazon, and built Claude, a chatbot with a language model that rivals OpenAI’s GPT-4. Now, Dario writes long blog posts about democratizing AI.

So what’s going on here? Why is the head of Anthropic suddenly speaking so optimistically about AI, when it was previously known to be safer, less expensive than OpenAI at all costs? Is this just AI hype to court investors? And if AGI is really close, how do we measure what it means to be safe?

To break it all, I brought The Verge Senior AI reporter Kylie Robison discusses what it means, what’s happening in the industry, and whether we can trust these AI leaders to tell us what they really think.

If you’d like to read more about some of the stories and topics we discussed in this episode, check out the links below:

• Machines of Loving Kindness | Dario Amodei
• The Intelligence Age | Sam Altman
• Anthropic CEO thinks AI will lead to utopia | The Verge
• AI manifestos flood the tech scene | Axios
• OpenAI just raised $6.6 billion to build ever-growing AI models | The Verge
• OpenAI was a research lab – now it’s another technology company | The Verge
• Anthropic’s latest AI update can use the computer itself | The Verge
• Agents are the promising AI companies of the future – and they are in great demand | The Verge
• California governor blocks major AI security bill | The Verge
• Inside the white hot center of AI doomerism | The NYT
• Microsoft and OpenAI’s close relationship shows signs of deterioration | The NYT
• The 14 billion question that separates OpenAI from Microsoft | The WSJ
• Anthropic floats $40 billion valuation in funding talks | Information