Jake Lahut: Oh yeah. Watch out [inaudible 00:10:47] boys. I know that's going to be a tough one.
Zoë Schiffer: Yeah, exactly. I would love to know how the AI categorizes this, but it's kind of fascinating. I feel like there's a lot of age verification stuff going on in the United States, a lot of rules and regulations that are getting rolled out and each have their own kind of issue. But this is kind of the industry's response to that, or an attempt to try something new and see if it works. And we will be curious. Our next story deals with AI technology and it's being put to use in really, really fascinating ways in the real world. So WIRED contributor Marta Abbà reported this week that the Italian Rescue Corps relied on AI to find the body of a hiker that had been missing for nearly a year. They did it by using two drones that gathered thousands of image frames in the mountain area of Monviso where a 64-year-old, Nicola Ivaldo, I think is his name, went missing in September of last year in the Alps. So the Rescue Corps took the images that the drone had gathered and processed them with AI. So this would've taken humans many, many, many hours, days, if not weeks. The software identified pixels that were actually Nicola's helmet, and that's how the Rescue Corps knew where to go and find him. I think this is so fascinating because we're talking all the time about the amazing things AI will do, and maybe this isn't curing cancer or whatever we're hoping is going to happen in the very near future, but it also just seems like such a clear example of AI is good at this. AI can look at all of these images and really quickly identify ones that are unusual, and then humans can follow up on that. And ultimately, we can find this person and hopefully in the future we could do this in a way that would save people before they died.
Jake Lahut: Yeah, I feel like where so much of the smart money is going and we're so much of the general discussion around AI revolves around how would replacing a full-time worker's 40-hour work week look versus an AI agent? Instead, I think that these longer time horizons are way more interesting of things that humans feasibly could not have the time to get around to, that these could actually start to reveal a whole bunch of areas of life where we might actually be able to come up with solutions that just ... And it sounds very old school, Silicon Valley hopium in a lot of ways, but these are things that people didn't think they could do before or you would just give up way earlier.
Zoë Schiffer: Exactly. And I feel like it's a good example of humans and AI working together. In this example, you really do need both. You need the drones to take the images, you need the AI to process them, and then you need humans to figure out what to do with that information. So I think it's an example of AI augmenting what we can do rather than necessarily replacing us all together. Coming up after the break, we dive into our inside scoop on how OpenAI made a deal with the U.S. government to offer their services to federal employees. Stay with us. Welcome back to Uncanny Valley. I'm Zoë Schiffer. I'm joined today by WIRED's senior writer Jake Lahut to discuss OpenAI's latest partnership. The company is partnering with the U.S. government to make its models available to federal employees. In practice, this means that federal agencies can get access to OpenAI's models for $1, very nominal fee for the next year. This is the culmination of a bunch of stuff that has been happening at OpenAI. So first, Jake, as you probably saw earlier this week, they released 2 open-weight models, which is the first time the company has done so since 2019. And then yesterday it announced the release of the long-awaited new frontier model GPT-5. And since even before Trump retook the White House in January, OpenAI CEO Sam Altman and other OpenAI executives have been really trying to cozy up to the Trump administration, despite the fact that Sam Altman in the past was kind of vocally against Trump, but we're in a new era. No surprise. I reported this story alongside my colleague Will Knight, but I'm curious just to kind of get your impressions. Were you surprised to see OpenAI announce this? Is this kind of where the government is heading?