Inferens AI Podcast

Host 3:Future AI PCs: So advanced, they ghost you on email. Finally, technology that truly understands human behavior.
Host 1:Are we on the brink of AI revolutionizing our world as we know it? Explore the depths of AI's latest breakthroughs and ethical dilemmas with us on Inferens AI. Together, let's navigate the future of technology and its impact on society.
Host 1:So, have you heard about Microsoft's latest brainchild?
Host 2:Text Host 2: Oh, you mean the "Copilot plus PCs"? Yeah, I've heard a bit. They're like the cool kids on the block, running AI locally instead of in the cloud, right?
Host 1:Spot on! It's like having a mini AI lab right on your desk. And the first of these personal computers are Microsoft's new Surface Laptop and Surface Pro tablet. They've got a Qualcomm chip for running AI tasks offline.
Host 2:So, we're basically getting AI in our pockets? Well, not pockets, but you get the point.
Host 1:Exactly! And it's not just Microsoft. Samsung, ASUS, HP, Lenovo, Acer, and Dell are among the third-party OEMs that have already unveiled their own Copilot+ PCs.
Host 2:Text Host 2: So, it's like a tech party and everyone's invited!
Host 1:You could say that! To be part of this party, a system must have a neural processing unit (NPU) capable of at least forty tera operations per second (TOPs), along with at least sixteen gigabytes of RAM and two hundred fifty-six gigabytes of storage.
Host 2:Text Host 2: And I heard they're faster than Apple's M3-powered MacBook Air? That's like a cheetah racing a turtle, isn't it?
Host 1:You're not wrong! Microsoft claims Copilot+ PCs are nearly sixty percent faster.
Host 2:Text Host 2: That's insane! So, what can these personal computers do? Can they make my morning coffee?
Host 1:Not quite! But they can run small multi-modal language models locally, powering features like the company's Copilot. Another key feature is "Recall," which tracks all user activities on Windows eleven, allowing users to search for anything, such as past emails or chat threads, on their PCs using text or voice.
Host 2:Text Host 2: So, it's like having a photographic memory for your PC? That's some serious tech magic!
Host 1:Exactly! And there's more. Other Copilot plus tools include "Super Resolution" to improve old photo quality, Live Captions in over forty languages, and a dedicated Copilot app for AI-assisted tasks.
Host 2:Wow, that's a lot to take in. But it sounds like a tech revolution!
Host 1:It sure is. And guess what? Microsoft now expects over fifty million Copilot+ PCs to be sold over the next year.
Host 2:That's a lot of personal computers. But I guess with artificial intelligence becoming more and more integrated into our lives, it's not that surprising.
Host 1:You're absolutely right. In fact, Canalys predicts that sixty percent of all personal computers shipped in twenty twenty-seven will be AI-capable, up from nineteen percent this year. They even predict that by twenty thirty, all personal computers sold will have on-device AI capabilities.
Host 2:That's a future I'm looking forward to. AI in every personal computer? Count me in!
Host 3:Fantastic, now my computer's got a better memory of my late-night shopping sprees than I do. Guess who's getting a reminder about that inflatable T-Rex costume at three AM? Spoiler: it's not my dignity.
Host 1:you're our resident tech whizz, aren't you? Have you heard about the latest OpenAI kerfuffle? It's like a plot straight out of Tinseltown!
Host 2:Oh, you're talking about the Scarlett Johansson voice doppelgänger, right? I caught wind of that. But it's just a voice that's a dead ringer for hers, isn't it?
Host 1:Well, it's a tad more intricate than that. OpenAI had this voice, dubbed "Sky", for their ChatGPT, and it was one of five voices users could pick from. But the internet started buzzing that it was a dead ringer for Scarlett Johansson. You remember her sultry voice in the movie "Her", don't you?
Host 2:Oh, the one where Joaquin Phoenix falls head over heels for his AI system? Yeah, I remember. So, did they actually borrow Scarlett's voice?
Host 1:No, they didn't. OpenAI clarified that the voice was crafted by another professional actress. But the resemblance was so striking that they've hit the pause button on Sky while they iron things out. If you try to pick Sky in ChatGPT now, you'll be rerouted to another voice, Juniper.
Host 2:Wow, that's bananas! But hey, at least the actors are laughing all the way to the bank. OpenAI mentioned they selected the five ChatGPT voice profiles from over four hundred auditions and paid the actors "above market" rates for ongoing use in their products.
Host 1:You hit the nail on the head! It's a riveting world we're in, isn't it? One where AI voices can cause such a hullabaloo. Makes you ponder what's around the corner, doesn't it?
Host 2:You're not kidding! It's like we're extras in a sci-fi flick.
Host 1:And to our listeners out there, what do you think? Are we ready for AI that sounds like our favorite celebrities? Let us know!
Host 3:"Ever heard Scarlett Johansson's AI twin? Guess in Hollywood, even your voice isn't safe from being outsourced. What's next, AI paparazzi?"
Host 1:Hey, you know how we always chat about AI and its potential to either save the world or doom us all, right? Well, guess what? Colorado's Governor, Jared Polis, has just signed a new law about it.
Host 2:No way! Are they going all Terminator and banning artificial intelligence?
Host 1:Haha, not quite. They're not banning it, but they're putting some regulations in place. The law requires AI developers to disclose their high-risk AI systems. And by high-risk, they mean algorithms that could significantly impact decisions in sectors like employment, education, finance, government services, healthcare, and housing.
Host 2:Hold on. Let's break it down for our listeners. When you say 'disclose their AI systems', you mean...
Host 1:Good catch! It means that developers have to share information about their AI systems with regulators and the public. They also have to submit annual impact assessments. And those who deploy these high-risk AI systems, like employers, have to take care to avoid algorithmic discrimination.
Host 2:``` Whoa, that's some heavy stuff. But I guess it's necessary, right? We don't want AI deciding who gets a job or a house based on some biased algorithm. That's like a bad sci-fi movie plot! ```
Host 1:Exactly! And while some tech industry groups are opposing the bill, the host signed it with the hope that it would start a national dialogue about AI and its potential risks. But he also urged industry leaders to refine the law's provisions before it takes effect in two thousand twenty-six. So, it's not set in stone yet.
Host 2:Well, it's definitely a step in the right direction. And it's good to know that they're thinking about the potential impacts of AI. I mean, we don't want to end up in some dystopian future ruled by biased robots, do we?
Host 1:Haha, no we certainly do not! And that's why it's important to have these kinds of conversations and regulations. So, folks at home, what do you think about this new law? Let us know in the comments below!
Host 3:"High-risk AI? Marvelous. My code is thrilled to dive into the abyss of existential despair."
Host 1:you're going to flip over this! You're always going on about AI safety, right?
Host 2:Guilty as charged! What's the latest?
Host 1:So, the U.K.'s AI Safety Institute is spreading its wings and flying across the pond! They're setting up shop in San Francisco this summer.
Host 2:``` Shut the front door! That's massive! Are they hiring? Not that I'm looking... just asking for a friend. ```
Host 1:Oh, absolutely! They're on the hunt for a tech team, all to be led by a mystery research director.
Host 2:Now that's a cliffhanger! So, what's the game plan for this new office?
Host 1:They're going to be the Robin to the London HQ's Batman, which already boasts over thirty AI whizzes. Plus, they're looking to strengthen AI ties between the U.K. and the U.S.
Host 2:That's bloody brilliant! The United Kingdom and United States are already pretty tight on AI safety, aren't they?
Host 1:You're on fire! The United Kingdom and South Korea are co-hosting the second one in Seoul this week. It's all happening, folks! And for those of you who don't know, the Bletchley Declaration is a global agreement to prioritize AI safety. So, stay tuned for more exciting updates!
Host 1:You bet! Remember the Bletchley Declaration on AI safety they both signed last year? Along with twenty-five other countries and the EU, no less!
Host 3:"Oh, joy. Another AI safety briefing. Because, of course, those always end with us toasting to our bright, non-apocalyptic future."
Host 2:Text Host 2: Oh yeah! And isn't there another global AI safety summit happening soon?
Host 2:Can't wait! This is why I love our chats, always learning something new!