The Vergecast - How to use Microsoft’s AI Copilot, laptop microphones, and Amazon Sidewalk

Episode Date: April 5, 2023

Today on the flagship podcast of peer-to-peer wireless networking:  02:31 - David Pierce talks with smart home reviewer Jennifer Pattison Tuohy about Amazon’s network of smart home devices called Sidewalk and the state of Matter, the promised smart home standard.  Amazon just opened up its Sidewalk network for anyone to build connected gadgets on 32:16- Monica Chin brings six laptops to Times Square in New York City to test out the microphones. 48:47 - Tom Warren joins the show to explain how AI is being integrated into Microsoft’s products, which may be more promising than Bing’s chatbot. Microsoft’s new Copilot will change Office documents forever Microsoft Security Copilot is a new GPT-4 AI assistant for cybersecurity Email us at [e-mail] or call us at 866-VERGE11, we'd love to hear from you. Vote for us in the People’s Choice Webby Awards for Best Technology Podcast! Learn more about your ad choices. Visit podcastchoices.com/adchoices

Welcome to the Vergecast, the flagship podcast of peer-to-peer wireless networking. I'm your friend David Pierce, and I am currently in my basement, which is now technically also my home office, and I am doing some cable organizing because it's April, it's spring cleaning time, and I have reached the point where I now have four large bins full of cables. Most of them are USBC cables, so I tell myself that I need them all because USBC is the future and you can never have too many. but I also have some that are like printer cables from the 90s that I don't think I ever used for anything. So the point is, it's time to go through some stuff. Anyway, once I finish, we have a great show coming up for you today. We're going to dig into two big things about the future of the smart home, Amazon Sidewalk Network and the Matter Standard,

and we're going to try to figure out where both of those things stand now. Then we're going to head out into New York City and test some laptop microphones to see if the mic inside of your laptop can hold up to call and video chats. And lastly, we're going to talk about Microsoft's many adventures in AI. And what Microsoft is up to says about the future of GPT4, office, windows, and like what it means to use a computer. All of that's coming up in just a second. But first, I have to find my headphone cables. So wish me luck on that. This is the Vergecast. See in a sec. Support for the show comes from Retool. Too many companies run critical operations on duct taped spreadsheets, Slack workflows, and whatever else they could cobble together. Not because they want to,

but because building internal tools means weeks of waiting on someone else's backlog. That's where Retool comes in. Build custom internal tools just by describing what you need. Proms something like build me a revenue dashboard on our Salesforce data. And Retool actually builds it on your company's data and your cloud with enterprise security built in. Go to retool.com slash Verchcast. We all need to retool how we build software. What's up, y'all. I'm Skylar Diggins, seven-time WMBA All-Star, Olympic gold medalist, and mom. And I'm Cassidy Hubbard, host and reporter for nearly 20 years covering the biggest names and stories in sports and mom.

And this is Am Mom, a community for athletes, game changers, and moms of all kinds. Dropping May 14th. Tap in with us. Welcome back. Last week, a sort of strange thing happened. Amazon told millions of people that, surprise, we just turned on a nationwide network of smart home devices. To which millions of people responded, I'm sorry, what? It turns out that sidewalk, or something like it, could be a really big deal for the smart home and the smart city and this big connected world everybody's been telling us about for years.

It might even be as big a deal as matter, which is the universal smart home standard that actually were overdue for a check-in on. So here to dig into both of those things is the verges Jen Patizantui. Hi, Jen. Hello, so happy to be back so soon. We promised the people we were going to talk about matter. I know, I bet there were a lot of angry people when we just glossed over it there. We're going to do that second because I'd just like to make the people wait for the matter talk. Fair enough.

But first I want to talk about, you wrote this piece about Sidewalk, Amazon's network, which I was shocked at how many people have read this story. It's like one of our most popular story this week. people paid a lot of attention. It sparked a lot of thoughts and feelings about all this stuff. We should probably start at the very beginning just because part of the story is that a lot of people didn't know what this thing is or that it existed or that they were part of it at all. So let's just rewind all the way to the very beginning here. What is Amazon sidewalk and how did it come into the world without anybody really noticing? Okay, yes. So this does go back to 2019. Actually, it began even earlier than that, which is something I only just discovered.

Amazon developed a local connectivity protocol for its own devices. So very early on, it realized that ring devices, like doorbells and cameras, were outside your house. Your Wi-Fi router was inside your house. And sometimes the connection between the two was very bad. And no matter what they tried, they really struggled to get some people's cameras to work the way they promised them. So they actually developed internally a system that would send low, Bandwidth, low power messages between the router and the device so that you'd still get a notification if someone pressed your doorbell, even if your Wi-Fi wasn't working.

Because, you know, if you have a doorbell and you press it and it doesn't work, you get mad. That is kind of a one-job situation. Yeah. Exactly. So that's how Amazon sidewalks started as a way for Amazon to fix a problem it had for its customers. But what it's developed into is something way bigger. It's essentially a low bandwidth, long-range wireless network developed to support low-power devices. And this is perfect for the smart home because the smart home is full of sensors. So basically what Amazon did was they developed this network. They just said, okay, we're going to go turn it on.

And it's actually inside all your devices in your home, all your Amazon devices and some of your ring devices have this secret network that we've been developing. And we're just going to switch it on. And everyone was like, uh, what? What are you doing here? Why are you doing this? And the thing that really upset a lot of people completely justifiably is it actually uses some of your internet bandwidth. That's how sidewalk works. It uses these low power radios, including Laura and Bluetooth low energy.

And Laura is like a 900 megahertz frequency radio. But in order to communicate to the internet, it actually needs the internet. So it uses a little bit of your bandwidth. So when Amazon first turned on sidewalk, basically everyone was opted in. And that's never a good move. I mean, tech companies, come on. Yeah. You know, just look at the history.

Opt-in is never the solution. The option has got to be for the customer. You can't just automatically opt people in. And that's where it all went a little sideways. But they did walk it back. And now, whenever you buy a ring device or a Amazon device, which is sidewalk-enabled, which is not all of them, but a large majority of them, you do get to choose to turn it on.

But that backlash is still strong and people are still kind of mad about it because Sidewalk does use your bandwidth. So basically it's kind of a toss-up here between are the benefits of what Sidewalk is offering, something that you're interested in, and that you think is going to be good for you slash your community,

or do you think this is Amazon just kind of making a power grab for your internet and you just want to switch it off? Okay, so there's sort of two things going on here that I think are interesting. One is that I feel like From that, like, first use case you described, right, where it's like, I need a way for my doorbell to connect to my device when the Wi-Fi is not working.

It's like a very common thing. These are the sorts of things we're trying to fix. There's also this bigger sort of community angle that I have a slightly harder time wrapping my head around, that it's like, one of the big announcement this week was Amazon is like, we're rolling this at work out completely. It covers 90% of U.S. citizens. And to me, it's like, who cares? Like, why I don't need my doorbell to access my neighbor's doorbell?

I just wanted to access my phone. So what's the, like, big network case here? Because part of me thinks, if you want to go sort of full conspiracy brain, this is like Amazon sort of Trojan horsing its way into essentially building like a national network on which it can do anything. And it's now an ISP and it doesn't have to play anybody else's games. And it's like, you can sort of spin out in this if you want to. Is there a smaller version of?

of what Amazon is trying to do here? Like, what's their argument for why this needs to be a gigantic nationwide network instead of just existing inside of your home? Their argument is that there are already billions of connected devices, but the future is hundreds of billions of connected devices. That is a pretty likely outcome. You know, the Internet of Things is not getting smaller. There are sensors everywhere.

Everything that anyone wants to connect to the Internet needs some form of a sensor. The problem right now is if you are outsized. side of Wi-Fi. So if you're outside your home or outside of business, the only reliable connectivity you get is from cellular. Cellular data, be it LTE, 4G, 5G, whatever comes next, is expensive. And most of these sensors don't need the amount of data that a cellular network provides. They only need really small packets of data to send their signals, minuscule amounts of data, and very low power. Cellular radios require more power. So their argument is we need to connect more devices to the internet. We know there are more devices coming online. We need an infrastructure.

And they're not wrong. We do need an infrastructure that will connect devices outside of our homes and outside of our businesses. Because ultimately, if the internet of things is going to bring everything it's promised from, you know, decentralized power stations to helping us find our lost dog to tracking packages, to keeping tabs on food waste. I mean, there's so many things that adding a sensor to products or devices or people or pets, there's so much that can be done, but you don't want to spend $15 a month to track. For example, I've had a whistle dog tracker for about five years to help keep track of my dog. And it costs $10 a month. That is a lot of money. And there are so many use cases where we really need connectivity. It's so embedded into our daily life now that when you

leave your house, you know, you lose that connectivity with your home unless you have a cell phone. So that's the argument. Their argument is when you say, well, why are you doing this? Because this is free. Let's not forget. Sidewalk is a free to connect network. So unlike AT&T and Verizon, if you have a device that connects to a sidewalk network, the user's not paying for it and neither is the developer. It's a free network. So I ask Dave Limp, who is the VP of services and devices at Amazon, well, why are you doing this? Because it's kind of a, you know, a bit of a, is this altruistic? And they said, well, we need it. Dave said, you know, Amazon needs this. I mean, who has packages all over the country at any one minute? And so, and also what's in it for them is he explained that

In theory, some of the customers that are going to use the network, so like developers who build devices that are used on the network, potentially will use their cloud services, AWS. So there's potential revenue stream there. Amazon Sidewalk is not the only option. Obviously, there are other power, wide area networks out there. People mention Apple Air Tags. That ecosystem is one example. And functions basically exactly the same way, right? It's kind of a sneaky device to device.

thing that you don't necessarily know. And if you think about is sort of sketchy, but then if you think more about it, it's really not that bad. Like, I feel like that's the path. A lot of these are going down. The really key thing here is use cases. Once it becomes obvious how this benefits people, that's where you kind of reach tipping point.

And whether or not this is a good thing, that's when we will know. And that's why Amazon announced this because they want developers to build things. Once there are devices on the network, then we can see, is this worth? what we are sacrificing. I mean, it's not entirely different from using Gmail. Sure. You get a great service, but you're giving your data. You know, nothing's free on the internet, as we know.

So you have to decide, is it worth it to you to be able to track your dog? Or one example is care band is on sidewalk. And so that's a wearable tracker for elderly patients. And it's sort of the idea is used for someone with dementia, leaves their home and wanders off. You're able to track them down. rather than having to pay for one of these cellular devices that would cost a lot of money in the long run. So use cases. And this is where the fall down has happened with previous systems similar to sidewalk.

There's one called helium. Helium is exactly this. And it is. It has very similar use cases. It's very similar network. But where it's sort of failed is very few devices use it. So when you don't have the use cases, then things start to fall apart. So that's why Amazon's pushing this so aggressively to say.

start, you know, they're giving it away for free, they're giving away these little test kits for free. I had one and I drove around my neighbourhood to see what the signal strength was. And I live in the southeast in a corner of Charleston, which is not a rural area, but there are rural areas near me. And I was actually very impressed at how strong the coverage was, considering, you know, I am in the east, which is much more populated. If you look at the map that they released, once you start getting out west, there's definitely large gaps in coverage. The thing I've really been trying to sort through my feelings on this is how I feel about the fact that it's Amazon.

Because on the one hand, you can sort of talk yourself into like Amazon's as good as steward of it as anybody, right? Like, I'm trusting Verizon with all my data in one way. I'm trusting Comcast with it in another. Like, why not Amazon? All the internet goes through AWS anyway. So like my life runs on Amazon's networks, whether I want it to or not, regardless. So it's like I end up in that sort of position of like, yeah, Amazon's as reasonable a sewer of

data as anybody and has made, as far as I can tell, pretty strong promises and assurances about like the security of this network that seems to have made most people feel pretty good, right? Yes, yes, it has. It's published a white paper. It has, you know, three layers of encryption. None of the network traffic is visible to Amazon. Devices that actually connect on via sidewalk don't have internet access themselves, so it's not like they're coming into your network. There are a lot of privacy and security layers here. But as with anything connected to the internet in any way, shape or form, there's always potential for some kind of workaround or a hack.

But in terms of trusting Amazon, from what they've presented and what we know, and obviously we haven't been able to test yet, because this hasn't been deployed at mass scale, whether you aren't giving your data to Amazon by using Sidewalk, unless you are using a ring device to an Amazon Echo device, then you're just all in the Amazon ecosystem anyway. Right. You know what you've signed up for at that point. Right.

So if you decide to enable Sidewalk on any of your devices, you can actually, you have the choice to show location or not. So a device trying to connect to you can know where you are or not. So there are tools that the user can control. But as we know, from going back to when this first started, no one really knows about sidewalk. So, you know, people don't know to go and turn this off if they're not comfortable with it. But now going forward and has been for about a year, maybe two years now, if you buy an echo on a ring device, it does make you opt in. It's not on by default.

Right. But looking at that map, a lot of people have opted in. I mean, there is a large coverage across the country. 90% of the US population, they say. And it's impressive. And it is a little scary, though. Like, we've just turned this on and, hey, look. Yeah.

We're everywhere. So, yes, I'm with you. I'm conflicted. I really want the connectivity. I have no cell service in my house. So when my Wi-Fi goes down, nothing works. And it's frustrating. And, you know, if you live in big cities and you have great cellular service, it's not so much of an issue.

But the rest of the country, it's an issue. But yes, having one company, whether it's Amazon or Apple or anyone else that came up with this, is concerning not because necessarily there's nefarious intent, but because they can just change it. Like, it's free now, but in a year's time, they might decide to charge. And now you've been depending on this. And, oh, now what do I do? Right. Yeah. I mean, like, welcome to the problem of how we've set all of this stuff up in so many ways.

But the other part of it, I wonder, is the fact that it's Amazon, this is actually probably a good segue. to the like platform wars we're about to talk about and the like interoperating standards that we all wish would exist. It seems to me that since sidewalk exists, say Google is less likely to support sidewalk than it is to build something like sidewalk out of its own devices. Apple already has the find my ecosystem that it could pretty easily sort of flip into the way that that its HomeKit stuff works. Z wave exists.

Thread exists. Like you could, I'm imagining a world where instead of having one thing sipping off a little piece of my network for the common good. I have 12 things sipping off of pieces of my network for the common good. And suddenly half of my band-with is used by my neighbor to make their doorbell work. Are we headed down the road of like instead of one good version of this network? We're going to have 50 just because Amazon built the first one? Or is there a version of this that actually kind of interoperates for everybody? So, and I think I said this in my article, Amazon feels like right now the only company that could have done it.

Because as much as we like to think they're not a monopoly, they have so many devices out there. They gave away those echoes for free, basically. But there are more iPhones than there are echoes, no? But an iPhone isn't going to work as an always-on-powered little radio in your home, right? That's true. That's fair. That's the thing. This is like a little mini cell tower in your home.

And if you have two or three of them, or if you have a ring floodlight cameras on the side of your house, I don't see a way that Apple is going to be able to do this. There aren't enough HomePod minis out there or Apple TVs. Definitely not. Not even close. No. And the same with Google Nest, minis and hubs. I mean, they gave them away for free.

But, you know, I think the orders of magnitude between what Apple's and Google have out there versus Amazon is significantly different. So, yes, I think there is a concern there, though, that if Google and Apple don't want to jump on the Amazon bandwagon, that we could end up with competing. We do have competing networks. I mean, we mentioned helium and then we've got the Fine Mai. Samsung has a similar sort of tracking protocol for its tags, and then its little bridge is a hub for that in your home.

So that we could have multiple ones. I don't know that it's necessarily worth it for the other companies at this point, because there's a big difference. And this is the difference, I think, between, say, sidewalk and Wi-Fi or Matter, is sidewalk is about, and the way I see it, it's bridging the smart home with like the smart city. It's taking your home, connecting the gap between the two, whereas most of these other protocols like Z wave, thread, are designed for the smart home itself. And what we need is just that gap between once we step outside our home and get into our business or wherever we're going, we need something to cover that. And that's why they call it sidewalk.

That's kind of the purpose. the way it's been deployed initially, it does seem more of a competitor to things like thread and Z wave because at launch they have like smart locks and smart sensors in your home. So there is crossover there, which is what kind of muddied the water a little bit for me. I'm like, why should this work for smart locks? I get that it's going to work for like a mailbox sensor or for a dog tracker. But I think as we said at the beginning of this conversation, Amazon just wants anyone and anything to start using this network so that we can start seeing the benefits of helping,

you know, of helping our smart homes. Because a lot of these networks that are already exist that we've discussed briefly are more industrially IoT. They're more about reading your smart meters or they're being used for like wildfire tracking, for inventory management, package tracking, a lot of industrial use cases. Sidewalk is much more squarely focused at consumer use cases. But we just don't, we haven't seen a lot of them yet. So we're waiting. waiting to see what developers come up now that they've sort of opened this up. Okay, fair enough. And that's a good segue into the second part of this I want to talk about, which is why nobody

is building cool smart home stuff. And that's not actually what it is. It feels like that sometimes. But that's not actually what I want to talk about. What I want to talk about is Matter, because it feels like we were here, I don't know, six months ago. And it was like, Matter is here. This is Matter's moment.

And then I went on parental leave and forgot about Matter for a while. And then you went to CES and everybody was. talking about matter and matter was the thing and matter was like the story of c es and it felt like momentum was really good and then i feel like and maybe this is just me over the last i don't know three or four weeks it feels like there have been cracks in the the matter world that that it feels like it's losing some of the momentum that some of the big players are starting to say well maybe maybe either now is not the time or this is not the thing or some combination thereof so am i reading the vibes correctly there? What are the vibes around Matter right now?

So, as my esteemed former colleague at, I used to work for a website called The Ambient that covered the smart home, very succinctly put it. He posted this week, all the major Matter players are singing the same song, but in their own key at different times and with vastly different words. And that's essentially what he says is so true. Everyone is on board. Everyone is still working towards a solution that was promised, which is buy a smart home device, plug it in, it just works. Right, that was the big promise. Yep. But we are still a really long way until we're in the Albert Hall with a concert and everyone's singing a beautiful song together.

It is a bit of a mess right now. Nothing is working. To go back to originally we talked about it, how matter is infrastructure. It's the plumbing, right? It's the roads for the smart home. It's something we can build on. The problem is plumbing doesn't work when you don't have a toilet. And the highways fall apart when there's no bridges.

And that's where we're at right now. We have all these pieces and parts, but then crucial bits are missing. So the connections are not working. For example, Amazon doesn't support thread. So thread devices will only work with Google, Samsung and Apple. So if you go buy thread devices that work with matter and you use Amazon smart home system, It's not going to work. That's a big missing chunk. We've got no support for Google Home or Amazon Alexa on iPhones right now. So you can use, obviously, you can use Alexa app and the Google Home app on iOS, but not with Matter devices. So again, these major pieces of the infrastructure are missing. Once they are there, things will, should proceed as promised. I think essentially to boil it all down, they should have waited another six months to launch matter. Matter should have come out this summer. But it also seems like it needs like a flagship thing that it doesn't have.

Because one of the things that really jumped out to me was the Hugh announcement. And basically like the Hugh lights are, I feel like one of the great success stories of the smart home. Like they're very good. They're very successful. Lots of people have them. And Hugh is basically, if I'm remembering the announcement correctly, basically like we're still probably going to do this, but not right now. And we're suddenly not feeling the need to be in a giant hurry to turn this on because there's really nothing that people gain from it. And I read that and it's like, well, you're kind of right.

Like there's nothing that I as a Hugh user seem to be standing to gain from them doing all of this work to support Matter right this minute. So you're either relying on somebody just like out of the goodness of their heart and the belief in the future to build in a bunch of stuff that doesn't really work yet. Or there needs to be a thing. Like somebody needs to make something that is matter specific and it just doesn't seem like that thing is coming. No, I don't think it is. The biggest issue with Matter right now, is what is the benefit? That is the problem. The consumer can't see the benefit. The manufacturers are struggling to see the benefit other than this is for the greater good and we need to have a

great strong infrastructure to build on. But the problem is the smart home is about experiences. And there are many experiences that companies like Hugh and Nanoleaf and Doorlock manufacturers like Yale and August offer to their users like auto unlock when you approach your door. That's a neat smart home use case. Having your lights automatically change with the sun during the day, that's a great smart home use case. Matter doesn't support any of those features. Matter is just very basic.

And this is intentionally the way it was created because companies, manufacturers want to be able to distinguish their products. And this was where Belkin comes in. If you recall a few weeks ago, Belkin, who produces Weimo products, said, we're not going to make Matter stuff. anymore. We said we would. Now we've changed our mind. And essentially, they said we changed our mind because our products don't have anything that necessarily makes them stand out in the new matter landscape. Right. So this is where there's a big tension right now. We're trying to, you know,

because we were promised we weren't going to have to use like seven different apps or have five different bridges when matter arrived. And right now we do. Matter doesn't change, doesn't fix those problems. So this is where the companies, you know, I still need to work together and figure out how they're going to make these experiences work. The way it's likely to happen is it's going to come down to the platforms again, which means, again, you're going to still have to be stuck in ecosystems, like in little ward gardens if you want certain features. But matter will mean that you can easily switch between them. So just like you have to choose a cell phone carrier, you have to choose an ecosystem, but

it should be pretty easy to switch. Yeah, and I think the Belkin thing is so funny to me because I read that whole announcement and it just comes off like the most epic self-own of all time where it's just Belkin being like, well, we can't figure out how to make good, interesting products that are better than anybody else. So we're just going to stop and we're going to force you to buy our products because we have like pretty good distribution and we sell our stuff in Apple stores. So we're just going to con you into buying our crap that doesn't work with anybody else because if we did the thing everybody else is doing, you would just buy cheaper versions of it on Amazon.

I still think it's mostly a cell phone, but I do see at least a little bit of Belkin's point here. I mean, the initial reaction was, oh, bad Belkin, but it was kind of brave. And kind of like, okay, yeah, we're kind of in a rock and a hard place and we don't really know what to do, so we're just going to stop. These companies in the smart home are scared of commodification. That was always going to be an issue with matter, like from the beginning. but commodification is actually what we as consumers want. Yeah, it's great news. It's good, yes.

I would like Belkin to produce Wemo products that work with everyone and only cost $5. Yes. But just turn on and off and just give me the basic features that I need. I mean, the smart home is a natural evolution of the home. So eventually, when you go into a Home Depot or you go to Amazon to buy your light bulb, you don't need to see whether it works with one voice assistant or this protocol. You just buy it like you would any light bulb and it will be connected to your home. That's where we need to go.

One undercurrent of the stories you've been writing about this stuff is that Apple seems to be, if not the holdout, then at least like one of the more sort of complicated parts of this strategy that Apple obviously has this big focus on security that has made it harder for companies to work with other stuff and also HomeKit over time. And Apple also, I think, has a long history and vested interest in things not being super interoperable. What is Apple's role in kind of this moment as we're making this transition? Like, is it the stick in the mud? Is it moving stuff along?

How is Apple doing here? I think it's been very interesting. Apple is probably the most vocal supporter of matter of all of the ecosystems. Okay. But vocal in Apple's way in that they never actually say anything. Right. So that's why it's been kind of hard to pass. But from talking to people who sort of are in the discussions and sort of know what's going on behind the scenes, basically HomeKit was not really a success for Apple.

It was complicated. It was hard. So to them, Matter is the solution for the smart home for them. That's why they're sort of phasing out HomeKit. They've rebranded themselves. They're now Apple Home. And they want everyone to work with Matter so that they can work with Apple. They don't want to have to deal with this separate ecosystem anymore.

But there are stick in the mud in that they keep their cards close to their chest and matter requires openness so that developers can figure out how to connect. And this is from what I understand, one of the holdups with the Google Home and the Amazon Alexa apps coming to iOS. Apple has changed things every time they release a new iOS update. And then there's also some struggles, as we've seen, with 60s. 16.4 that just came out, which is the first time they brought back the new home kit architecture. Right. There was the new architecture and then the old architecture and now it's the

new architecture. Now we're back to the new one because the rollout of the first one was kind of a mess. That's a hold-up. So I don't think Apple is stonewalling anything here. Certainly not on purpose. I think it's a complicated process for them to kind of untangle themselves from HomeKit and reopen their platform to matter. And that's what they're trying to do. And again, this is why all of this is taking so long, crucial parts of the infrastructure just have not connected. Again, six months, they launched in November, Matter. They probably should have launched it this summer or next November. It is going to be a long time until I come back on the Vergecast and say, hey, matter just works. It's going to be a great day. That will be a great, great day on the Vergecast. All right, we've got to move on.

Jen, thank you. Appreciate it, as always. We'll see you again to yell about Matter six months. Sounds good. We're going to take a really quick break, and then we are going to come back, and we're going to go into New York City and test some laptop mics with Monica Gin. We'll be right back. Support for the show comes from Framer. Framer is an enterprise-grade, no-code website builder used by teams at companies like Perplexity and Muro to move faster. With real-time collaboration and a robust CMS, with everything you need for great SEO,

not to mention advanced analytics that include integrated AB testing, Your designers and marketers are empowered to build and maximize your dot com from day one. So whether you want to launch a new site, test a few landing pages, or migrate your full.com, Framer has programs for startups, scaleups, and large enterprises to make going from idea to live site as easy and fast as possible. Learn how you can get more out of your.com from a Framer specialist or get started building for free today at Framers. Framer.com slash verge for 30% off a Framer Pro annual plan. That's Framer.com slash verge for 30% off. Framer.com slash verge. Rules and restrictions may apply.

Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts, but time and resources are limited. Finding, connecting with, and screening the right candidates takes up valuable time you could be giving to your customers. That's where LinkedIn Hiring Pro comes in. It's built to be your hiring partner, helping you find the right candidates faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process from drafting your job to shortlisting candidates and conducting AI-powered interviews for initial screenings. Its updated conversational interface lets you describe what you need in plain language.

Nearly 60% of hirers find a candidate to interview within a week. With Hiring Pro, you spend less time searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward. Join the 2.7 million small businesses using LinkedIn to hire. Get started by posting your job for free at LinkedIn.com slash track. Terms and conditions apply. Welcome back. Every now and then on this show, we'd like to do a microphone test. Because we spend a lot of time worrying about our headphones and how things sound to us,

but not enough time testing our microphones and how we sound to others. Last year, we stress tested the microphones on a bunch of flagship wireless earbuds earbuds to see which sounded the best for voice calls. You should go listen. It's all over the Vergecast last year. But the short version is this. The Sony LinkBuds, Apple AirPods, and Google PixelBuds Pro are all pretty good. The $20 Apple Earpods, the ones that actually plug in, are amazing. Everything else we've tested is less good.

But the thing is, not everyone uses earbuds for Zoom and Meet and Teams, or whatever else you use for calls and video chats. So for this test, we're checking out how the microphones on a bunch of laptops sound, to figure out which is the best for taking calls, meetings, and even just recording audio with the built-in mic. You might remember that when we tested earbud microphones with Chris Welch last year, We sampled what the laptop mics sounded like, too, and it, well, it sounded like trash. You hear everything.

The whole world. There's a truck. But things changed fast. So we had to see. Can laptops handle a voice call? Can you get away with no extra gear whatsoever the next time you get on Zoom or meet? Monica Chin is the Verge's laptop reviewer, and she helped us stress test a few of these laptop microphones. So while I sit comfortably at my desk at home in Virginia, we sent Monica with our producer,

Andrew Marino to the single most chaotic spot we could think of in New York City. So we are walking through the middle of Times Square, which is probably the loudest place in New York City that you can go besides a concert. What you're hearing there is the 2021 MacBook Pro. It's the 16 inch one, the one with the M Pro chip. This is like the laptop that our video team really loves to record stuff on. The mics are pretty good, I would say. How does it sound? I cannot believe how good you sound. This is blowing my mind right now. Really? Well, I'm used to these microphones being really good, but they catch everything, right?

So you get tons of background noise, you get tons of everything. We could just have a video call with you doing this. We could, except that I can barely hear you. There's that. An important detail I should mention here is that Monica did not wear headphones for this test. We found that a lot of people don't have wired headphones with them when they're out of their house. So you're going to hear a lot of my voice. being picked up in the laptop's microphone,

which, spoiler alert, is actually one of the biggest takeaways here. Mikes are getting better, but echo cancellation mostly still sucks and causes a lot of problems. For the mics themselves, though, the biggest problem we came across with basically all laptop microphones is when the wind picks up. Really no surprise there. Okay, so we're getting to the most iconic part,

which is the big TKTS booth where everyone hangs out. Look, this is probably the least realistic test we've ever done. Not many of you are walking through Times Square holding your laptop in front of you. If you are, I have like a lot of follow-up questions for you and feel like we would be very good friends. But I am sure that you've tried to take a call in a crowded coffee shop or at a park in a city or in your backyard with airplanes flying over. Not that long ago, I would have told you that doing that and trying to talk into your laptop mic was a horrible idea and you're going to hurt the ears of whoever you're talking to. But I kind of couldn't believe how easy it was for me to hear Monica so

clearly. It's not amazing, but it works. Okay, now tell me what you had for breakfast this morning. This morning, I had a cake pop from Starbucks on the way to work. Part of this balanced breakfast. Yeah, exactly. I'm not a breakfast eater really, usually. So that was like a breakfast. It's like my dessert from the night before, basically. Okay, so that's the MacBook. Next up, Monica pulled out the Dell XPS 13, which is long. than our go-to Windows laptop recommendation. And it held up pretty well here, too. This is one of the best Windows laptops you can get,

but it's obviously a good bit smaller than the MacBook Pro 16. We got an Intel processor inside. We got a teeny little webcam on the top. We got downward-firing speakers on the bottom. It's a much smaller, much smaller little deal. Yeah, so I'm hearing more and more detailed background noise. I can sort of pick out individual people talking instead of just like the wooching background sound

that was in the MacBook. But even still, your voice is coming through pretty well. I can pick you up fine. It's definitely more processed than the MacBook. Interesting. But even still, I would rank this in like mid-range headphones to be using in a crowd of people. I'm like blown away by how good this is, in all cases.

One of the big things with Intel's 12th generation, which is what's in this and the 13th generation as well that just came out is like they make a big deal. out of AI noise processing. So it makes sense that it would sound more processed, I think. Yeah, it's definitely working hard to do it, but it's doing a pretty good job. Okay, prepare yourself for this one. The next laptop Monica pulled out, she didn't tell me about before we sent her to Times Square.

And when she joined our video call, it just had a sketchy name, like, surprise laptop. It definitely surprised me. Okay, that was literally you just caused physical pain in my headphones when you did that. It's awful. I hear everything. Okay, so this is the Gateway 14. It is a $400 laptop that I picked up at Walmart because I was curious if it would work at all. That was the first one in all of these tests that we've done that was like actually physically painful.

As soon as Monica entered our chat here and I started hearing sound, I like literally did the thing where you have to go like, ah, and just pull your headphones off. because that's what happened. That was awful. Before we go any further, I should say that the way we recorded this was the same as our other tests. We use an app called Riverside FM,

and we like it because it records your local audio and then uploads it. So you don't get any of that internet artifacting or echoes. It does mean these tests are going to sound slightly different from what you'll sound like on Zoom or Meet or Teams because those apps all apply their own kind of filtering and noise canceling. This is much closer to like the pure, audio coming out of the microphones. That said, that gateway laptop was worse than just the microphones.

In addition to sounding like a metal band playing inside a scratchy tin can while a lawnmower goes full blast right next to my head, the computer itself could barely even keep up with having that one Chrome tab open right out of the box, and it initially lost the recording because the computer literally couldn't keep up. The point is, it's not a very good computer. I wouldn't use it for this or anything if you have another option. Okay, let's get back to better sounding things. The next laptop we tested out was a Chromebook, the Lenovo IdeaPad 5I gaming Chromebook to be specific. Lots of people use Chromebooks, especially in education, and we live in a remote world, and lots of people take classes online, and this actually seems like something you'd hope Chromebooks specifically would be good at.

All right, we are here. This is so much better. We are on a 15-inch gaming Chromebook. Where does this fall in, like, the range of Chromebooks? I mean, it's not terrible. It's got a little bit of a plasticy build, but, you know, it's got a big screen. It's got a nice little numpad here on the right side of the keyboard. The audio is way too quiet to be using a Times Square.

But, you know, if you're going to be using headphones all the time, then maybe that doesn't matter so much to you. This is a good middle ground of, like, the good ones that we heard at the beginning and the horrible garbage one that we heard before. This one is like, I can hear a bunch of background noise. There's definitely more wind and stuff coming into the mic than before. But I can still hear you. And if we had a call like this, I would not hate. your guts, whereas like three minutes of that last one, and like, I don't want to be friends with you anymore. Could you hear there was a man rapping in front of me just now? But it sounds like it wasn't,

that wasn't quite coming through. It does seem like I'm picking up a lot more behind you than behind the computer, which is interesting. I guess, yeah, the microphones are facing that way. So this is actually a good example, because as I'm looking at you, a crowd of people just walked up behind you, and I could hear all of them talking. But then when somebody walked right in front of you behind the computer, I didn't hear that at all. That's very interesting. Interesting. So I guess if I'm doing a call in Times Square, I want to be sitting at the back of Times Square. Yeah, like face your back to a building or something. Yeah. Laptop number five also falls into the probably ought to be good at this category.

It's the Lenovo ThinkPad X-13. It's kind of a classic business laptop. But this one is actually built with mobility in mind, which is all the more reason that the mic should be good. This is a Lenovo ThinkPad. It's super super light. And it actually, they see, Snapdragon chip inside. It's powered by the Snapdragon 8 CX Gen 3. This is one of those computers that's like trying to kind of be a tablet and it's like very mobile on purpose.

Correct. So it should be good at this. Super, super thin. It's like, it feels like it weighs nothing. But I don't know how good the speakers are. It's very, very thin. So there's not a ton of room for great, great microphones necessarily. It's doing less processing than most. So like your voice

actually sounds really good. Oh, okay. I'm also picking up a lot more background noise. Can you hear there's like a party happening down and a bunch of people cheering. I think, oh, people are screaming at Mickey Mouse down the block. Can you hear that? No, spin me around. Let's see if we can hear. Oh, I think Mickey Mouse, there he is.

Yeah, Mickey Mouse and Sonic are walking into the levee store. And people are very excited about it. But I can hear you much better as well, which is a big relief. Okay, let's wrap this up. One more test subject. Let's go back to Apple. That first one we tested, we kind of know. knew was going to be the best mic of the group. But this one's probably the most common laptop

you're going to see out there. This is the newest MacBook Air with the M2 chip. It's the laptop I just tell everyone to buy it because it's great and you'll like it fine. But it turns out having a better mic array in the pro really does make a difference. We are here with the M2 MacBook Air. Tiny little, tiny little guy. And it is about to get a lot louder because we're going to walk to the subway station. How are we sounding on this one? This is, I would put like right in line with the think pad in Dell. Like that first MacBook is clearly better than everything else. Oh, interesting.

Because that's one of the things that, you know, sometimes you wonder like how much of the microphone performance is the physical hardware and how much of it is the software. And presumably these are using the same software. So it sounds like the hardware does make a difference in this case. It must because that one is like a one order of magnitude better than this. Like, this is pretty good. It's kind of like the think pad we were just on in the sense that it's like not a ton of processing, which means I'm getting a little more background than usual.

Yeah, this one is definitely harder to hear than the 16 as well. The 16, the MacoPro 16 was the one where I really didn't have to lean in at all. It is impressive to me being in an environment that's so loud and knowing that these tiny things are filtering out all that noise around me. Like, I would not expect it's technology, man. We should say as a general rule, don't do conference calls from the middle of Times Square. But I guess it's encouraging to know that this is a thing you can pull off if you absolutely need to. It sounds like if you're on the 16, you can probably do it.

Just don't do them on something that only cost $400 that you picked up at Walmart and is blue. And it's called Gateway, which is a brand I legitimately did not know still existed. Definitely don't do them when you're walking through Times Square to the subway like I am because Andrew has had to shepherd me across the crosswalk several times. All right, well, you guys go. Don't, don't die. Don't become the new spectacle in Times Square. But thank you for doing this.

This is great. Thank you. This was really fun. Okay, three thoughts on this before we go. First, I'm still kind of amazed at how good these sound. I've typically told people to avoid their built-in mics, except in really quiet spaces. Because even when they're decent mics,

they're made to pick up audio from a much larger area, so they tend to get much more background noise than your average headphones, because those can just point their mics right at your mouth because they're always situated in the same place. That's still true. A good pair of headphones will sound better than any of the laptops we just tested, but it's much less true than it used to be, and most of these at least fall into the totally usable in a pinch category.

Second thing, my rule is that you should always, always, always, always, wear headphones on a call, and that rule still holds. Without echo cancellation, if you're not wearing headphones, you're going to hear your own voice a ton. With it, the computer has to do a lot more processing, so audio gets worse, and you get that thing where you can only talk one at a time. It's just a lot worse.

So even if you're using the built-in mic, you should still use headphones. Also, for like basic human etiquette reasons, you don't want the world to listen to your phone calls. Another reason to wear headphones, by the way, is that one of the most interesting things about our testing was that Monica could barely ever hear me. I could hear her okay through the mic, but even with the volume all the way up, she was always having to hold her head right down by the speakers, and even then she could barely hear me.

So again, for so many reasons, wear headphones whenever humanly possible. And third, the hardware really does matter here. It's not an accident that the Macs, which have three mics and are set up to do noise cancellation, sounded better than the other ones, which mostly have two or in some cases even just one mic. But there is also a fair amount of software work here, which you can hear in some of the audio processing, especially on the Dell and the Lenovo. And that's really good news, actually. Over the last couple of years, it looks like laptop manufacturers started to care about the quality of their microphones for basically the first time ever.

So things have gotten a lot better really quickly. Maybe someday I'll ditch this fancy mic and just make a podcast by yelling at my computer. But here's what that sounds like right now. So yeah, maybe not yet. Anyway, we've got to take a break, and then we're going to come back and talk about Microsoft's many adventures in AI. We'll be right back.

Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts, but time and resources are limited. Finding, connecting with, and screening the right candidates takes up valuable time you could be giving to your customers. That's where LinkedIn hiring pro

comes in. It's built to be your hiring partner, helping you find the right candidates faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process from drafting your job to shortlisting candidates and conducting AI-powered interviews for initial screenings. Its updated conversational interface lets you describe what you need in plain language. Nearly 60% of hirers find a candidate to interview within a week. With Hiring Pro, you spend less time searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward. Join the 2.7 million small businesses using LinkedIn to hire.

Get started by posting your job for free at LinkedIn.com slash track. Terms and conditions apply. Support for the show comes from Anthropic. Not every question has an easy answer. And the ones that are really worth asking usually come with a healthy mix of inspiration and backpedaling, aha moments, and quiet meditation. When you're working through one of those problems, you want a partner to bounce ideas off of and figure out where the deeper issue lies.

That's where Claude can help. Claude is the AI for minds that don't stop at good enough. It's the collaborator that actually understands your entire workflow and thinks with you, Whether you're debugging code at midnight or strategizing your next business move, Claude extends your thinking to tackle the problems that matter. Plus, Claude's research capabilities go deeper than basic search. It can have comprehensive, reliable analysis with proper citations, turning hours of research into minutes.

Ready to tackle bigger problems? Get started with Claude today at cloud. cloud. a. a. slash vergecast. That's clod.a.ai slash vergecast. And check out Claude pro, which includes access to all the features mentioned in today's episode. Claude.a.ai slash vergecast. Welcome back. One of the questions I'm personally most interested in in tech right now is what is AI actually good for? We've talked a ton about GPT4 and chat GPT and Bing and Bard, but a lot of the things that we talk about are neat novelties or their glimpses at like a far-out future for when these products get a lot better. But the truth is, AI is much more than chatbots. And Microsoft is being much more aggressive than most in figuring out how exactly to put AI into all of its products.

The Virges Tom Warren has been writing about and testing as much of it as he can get his hands on over the last several months. So he's here to tell us how AI is going to change the way we do everything, at least the way Microsoft sees it. Hi, Tom. Hello. Welcome back. It's been a while. Yeah, it has. There's a lot happening right now. This is why I wanted to do this, because it feels like Bing has gotten tons of coverage, and we've talked about it a lot, and it's like really interesting. But if the question is, what is AI actually good for? I feel like it's all the other stuff that Microsoft has done, which we have not talked enough about. Yeah, that's the promising stuff. The search engine stuff is cool, but having it everywhere is cooler. Okay, so let's start with co-pilot, which I think is probably the sort of biggest version of

what Microsoft is thinking about here. And I was reading a bunch of stories and realized I have a hard trouble figuring out exactly what co-pilot is. It's kind of a chatbot, but it's not totally a chat bot. It's different things in different apps, but it's kind of one thing. Like, frame co-pilot for me. What is this thing? Yeah, so it's essentially, think of like Clippy or like all the promises of Cortana.

It's like basically an assistant, so a virtual assistant. So yes, it can spring up and be a chat bar and you can, you know, converse with it and stuff. But it can also just sort of like magically appear almost with AI magic. In documents, say you highlight a paragraph and you're like, I want to rewrite that or reword it or whatever, then it will just sort of like hint in a way that you know, you know when you goof up and you get a typo and word comes along and underlines it, that sort of thing.

It will just sort of gently appear and be like, I can rewrite this for you. So that's kind of the interesting way that co-pilot works. So you can say, you know, create a document, create a PowerPoint, create an Excel based on a bunch of prompts, or you can drag a file into a PowerPoint or a word file, and then based on that, all that text, it will generate a bunch of slides. So you can do all that sort of stuff. But I think the interesting part is where it just shows up or, like, in Excel, let's say it does something, like you say, here's a bunch of data.

Here's, you know, financial earnings for his company the past six months, put it into a table or pivot table or whatever, or something you don't really know how to do. It will then tell you how it did it and show you how it did it. So it's like training you as well. So it's like it's an assistant, but it's also a tutor, I guess, as well. So it's like there's a bunch to it, really. That's actually really clever because I think one of the things I was looking at as I was reading your story is like the sort of famous office problem over the years is that office is

this like hideous mishmash of menus because there's so much that all these apps can do that it's impossible to figure out how they work. And Microsoft is like, we made the ribbon better. And everybody is like, all my buttons are gone. And like it's just, it's an impossible problem. And it feels like if they can get this right, like there's somebody at Microsoft being like, we can finally get rid of the ribbon, right?

Like that has to be what's happening here. It's the complexity of the interface, right? And just Excel's the worst for this because it's just like you can open it up and it looks relatively simple, but the stuff that you can do with it, like the difference between me using Excel and like some power user is like, you know, massive.

And it's just reducing that complexity and making it more available to people and just easier to use. I think that's kind of like half the story behind and copilot and the other half is like, yeah, you can chat to this thing, have it back and forth, get it to do stuff. And you can say to it, like you wouldn't chat GPT really and Bing and stuff, you can say, like, write this and you know, you could prompt it afterwards to tweak it and make it better

or make it into a list and all that sort of basic stuff. But it's really interesting where it takes that complexity out of office apps and then, you know, just surfaces it in really interesting ways. And I think they're really at the beginning of that. So there's not like a ton of examples yet. So I was trying to like kind of get that out of them like, you know, where is this heading? But I think for me it's like it sounds like it's just going to be the virtual system they've always promised, or the beginnings of it at least.

Because they tried, they said that Cortana would do this stuff, right? Well, that's part of what I think is so interesting here. Because there's, I mean, you go back and like you mentioned Clippy and there's Cortana. And there's like one of two things is happening here. Either Microsoft has been right all along and finally the technology is starting to catch up to kind of this thing that they want build. Or Microsoft has been wrong all along and just refuses to acknowledge that like this sucks and nobody wants it. And Clippy was either a great idea that was too early, Cortana was either a great idea that was too early or a total disaster of an idea. And this feels like the same,

like why can't Microsoft let go of this idea? They just love as an assistant, don't they? I guess so. It's kind of interesting because obviously Clippy was, it was interesting at the time, but then it got annoying because it would pop up when you didn't want it to and that was obviously the meme and then Cortana was like supposed to be this thing that would do everything for you and it just failed over and I think Satya Nadella recently said it was dumb as a rock right and all of them like you know Alexa and the rest so they obviously and they went they went down the route of going for bots as well chatbots initially a few years ago um obviously before open AI models were ready so there's always been this promise and they've always been obsessed with it but it does kind of feel like

okay now's the time where it might actually work it's just is they have to be careful. They don't go down the sort of clippy route where it pops up annoyingly. It just kind of gets in the way. And it's more like actually a useful assistant that you call upon when you need stuff. But it's also like, what is the interaction model there? Like when do you call on the assistant or the assistant appears rather you just click a button to bold something? Or, you know, like the user interface is still there.

Like how do they lean back from that basic user interface into this, into the complexity? So there's a bunch of different challenges. and they like customize it per app. So it doesn't do everything you'd expect in each app. That's another kind of like learning curve to it almost is that you might think it's going to do something in this app but it's obviously been customized heavily for like PowerPoint because PowerPoint's completely different to Word.

So yeah, so they have to manage a bunch of a bunch of that as well. But from what they showed me, and I have played around with it on my own PC in Word, but it's very basic now. But from what I can see, what they've shown is it does. It is pretty impressive. I was kind of impressed with it. And I'm old now.

And, you know, I don't get excited about things. Yeah. So, like, this stuff is like, oh, okay, this is actually interesting. So it's not like, it's not one of the city future vision things. It's like, oh, it's actually working. Yeah, how does it give me a sense of kind of what those demos are like? Because I think one of the things you mentioned that jumps out to me the most is the thing where it's like, it pops up when you write a paragraph and it's like you did this thing wrong.

And I think, honestly, even tools like grammarly go too hard in a, a space like that where it's like I'm writing an email that I'm like deliberately writing like a person because I'm a person and it's like you should potentially use more formal language in this email and I'm like no grammarly like shut up leave me alone and I can imagine with all of this stuff tuning that thing between like how do you sort of get involved and be helpful and how do you avoid being clippy just popping up every two seconds being like look at this thing that you're doing can I help that's a super hard line to walk and Microsoft is going ham doing all of this stuff I can imagine a world in which they would kind of way overshoot and have to pull it back over time.

But what have you seen so far? Yeah. So the basic stuff I've seen so far is that like you highlight a paragraph and then it kind of like it does like a purple sort of outline to the paragraph. And then you kind of go to that purple outline and then it's like co-pilot sort of like gradually appears almost. Okay. It appears like as like a prompter you can click. And then it obviously comes into the sidebar from there and you know, you interact with it and tell it how you want to rewrite the paragraph.

Or it might even, actually I think in that particular instance, it comes up and says like here's 12 examples or whatever I've rewritten this paragraph for you. And you can just flick through them and you can just see them, you know, the text. So that's like okay because you have to take action on it. It's there like it's always kind of in the background. So it's not like trying to jump out. It's not trying to suggest, you know, text to you without you, you know, prompting it to do that. So I think that's probably the way they will go about it. that's the main one that I saw of that sort of use case that they showed.

There's some other stuff like PowerPoint where you can add a file and then it will say, you know, do you want to match the colors across the slides and stuff like that, which is kind of useful. But I think right now it's like it's less trying to jump out at you and trying to interact with you and more like you call upon it, you know. For example, if you wanted to, in PowerPoint, you want to create a file based on that word document you were working earlier, it will know what you have. recently so it'll kind of suggest like oh do you want to pick this file and you go yeah okay and then it

will take a few seconds to do that and then you and then it will sort of have you know like in bing where it says suggested responses after you've you started talking it does that it says oh do you want to you know customize this based on that and the way that they're doing it right now seems to be very you know baby steps step by step being careful with it so i think they're obviously cautious of the whole clipping stuff so yeah because did they mention clipy on the on the stream i don't think they did, but I definitely asked them about that. And they said, you know, they're proud of the fact that they've tried this stuff of, like, helping people in the past. But, yeah, to me, it looks like they're being cautious. Yeah. And it does seem like if I'm Microsoft,

this is a tough spot to be in. Because on the one hand, I did get the sense reading your stories that they think this is the moment, right? That, like, all the stuff we've been trying to do this idea of, like, how do we bake in these sort of helpful additional tools? Like, they're not shy about how sort of big a moment they feel like this is. I mean, obviously, you don't invest $10 billion in Open AI if you don't think this is the moment, right? But then on the flip side, like the long history of Microsoft says, most people don't want to change their workflows. People who do these things all day, every day don't want help.

And what they want is for you to like leave them alone so they can do your job. And that has making it really hard to, you know, change the interface, to change how things work, to even add things like collaborative features has been tough because it just screws with the way that people have been doing things for 30 years. is Microsoft wary of that or is this just kind of like we're just going to push through the pain because this is the future moment? I think they are because the way that it shows up right now is it's literally a button in the ribbon. So it's not like trying to replace that ribbon. It's just another command that you kind of summon almost for now anyway.

And it comes up in a cyber and there's been a cyber in office for years. So none of that is completely different to what you'd expect in office. Whereas if they'd gone, you know, pulled out the ribbon and said, this is how you're going to use it. like they do with the Xbox One with Connect. They were like, you've got to talk to your Xbox. And then they had to like walk that all back and make it work with a controller. Yeah. So like they haven't gone that drastic with it.

So I think they're definitely aware of that. But I think like you said, it is the moment for them. You get a sense that they feel like they're ahead of Google, right? Ahead of competitors that are trying to do similar stuff in this space. And that they're poised to sort of, you know, take on that across everywhere. They've done a copilot in, they've had it in GitHub for like a year or something. owner, which obviously wasn't quite as powerful as what they showed, but then they, you know, a week later, they were like, here's GitHub Kai Pilot X or whatever. So they're going fast

everywhere. And I'm sure they announced a security copilot as well. Well, so this is what I'm trying to figure out. There's the, there's the security co-pilot, there's copilot on all the apps, co-pilot exists in teams. There's like business chat, which is kind of co-pilot, kind of not. Is like, is the big vision that there's this like one assistant that you sort of interact with everywhere or is it like lots of different little AI things across your different apps? Like how does Microsoft see it? Yeah, I did ask that and I think that

the basic sort of gist of the co-pilots is that they're tailored to each sort of instance. Like in Word, for example, it's obviously customized for writing and, you know, in certain imagery and stuff. And PowerPoint's obviously a bunch of commands to make your slide decks look nicer and you can pull in sort of

they're going to pull in like Dali images and stuff into there and Excel is obviously very command driven. So it's very different per app. And then obviously GitHub, you don't want like, they don't have like the GPT4 chatbot sort of interface where you're in line in your code because that's slow. You know, the latency there would be too slow for them to do those automatic suggestions that they do with the GitHub copilot there.

So they're tweaking it and customising it where it makes sense, which I think does make sense right now because we're still in the early phases. This stuff is still a little bit slow for, you know, that rapid. respond stuff that you might want in certain scenarios. But who knows long term, will it become this sort of Cortano sort of assistant? I think it probably will. But I think that's probably a ways off from the initial stuff that they're adding over the next couple of years.

But yeah, it makes sense. Because like you said, the business chat stuff, that is co-pilot, but it's only in teams. So you go to that and it sort of scours across your whole M365 stuff, so all your documents. So it kind of can take actions. all those individual co-pilot actions in that interface. Let's say you said, oh, I want to find this word document. Can we make it into a PowerPoint? You could do that there.

Whereas if you said that in the word co-pilot, it doesn't have that broad view because it's very tailored. That's kind of a hint of where they will go. And I think there probably will be, you know, 10 years there'll be a co-pilot that can do all your personal stuff and help you out, all that sort of stuff. And it just switches over to work mode and does all that sort of stuff. So, I mean, they haven't quite laid it out as a vision, but it seems like that would be the way, right?

Yeah, I wonder even if that's why it's called co-pilot and not like, you know, John the co-pilot. Like, it seems very obvious to me that Microsoft has been, like, deeply burned by having too many sort of silly characters that everybody grows to hate. And it's like, whether it was Cortano or Clippy or, like, TAY, it's like, Microsoft has done this thing

where it's, like, interact with the machine, like it's a person. And everybody's like, this person sucks and I hate them, get them out of my face. And co-pilot is just much more of like a, it's a set of tools rather than, your new best friend. And I feel like that's probably a much saner way to go about the future. Yeah. And it's also, then it doesn't also feel like it's replacing you, right?

True, yeah. As a Cortana. So I think that's definitely part of it. And the co-pilot name kind lends to what it's, you know, the way they've designed it. So it's there alongside you to help you rather than, again, replace you. So yeah, like, I think they've been very careful with that. And apparently Satya loves the name co-pilot, so. Which obviously GitHub kind of started using initially. Yeah, so you mentioned business chat, which I think if I were to pick like one thing that feels like the sort of big picture, like if I'm, if I rewind, you know, a few years and it's like this is when Microsoft really got interested in what opening eye is up to you. Like business chat feels like that thing, right? Where it's like plug all of your company data in and use this model to essentially like query and talk to your company's information. And for Microsoft, it's like it's been, it's experimented with, you know, a million different ways of. storing company information, SharePoint, OneDrive and Yammer, and there's just a billion ways to communicate and share stuff and never quite gotten it right. But it feels like business chat,

especially with coming with this like redesign of Teams that's coming out now and it seems like Teams is very clearly working. Is that combination as central to the future of Microsoft as it seems like it is to me? I don't know if it's necessarily like essential to the future, but definitely with Teams, they've always viewed that as like the Work Hub, like where you're, where you you sort of start your day essentially and then you go into your office. So it kind of makes sense while they've done business chat there. But yeah, the redesigned teams,

they're definitely, like they've gone head on into like fully being just web on that. Like it's just a web. It was electron anyway. But now, like they could have gone to a native app. You know, they haven't.

So you can tell that they're going to fill that with loads of AI experiments. And they already is a bunch in there anyway. So they're definitely preparing that as the platform that perhaps they'll push more of this stuff too. Because if everyone's spending their time more in that and then jumping in and out of these office apps, that would probably be the hub where they can experiment more with this AI stuff. Yeah, that makes sense. That definitely feels like it. And probably the safest place to do that just because fewer people, like we were talking about before, have nobody has a 20-year history of how to use teams in the way that people are like really married to how Excel works.

So you can kind of, you can have your cake and eat it too a little bit there. Yeah, you can play around with it there. And they have done a bunch of that. Like they've done a bunch of features in teams over the years. that they've added in pretty rapidly, especially when since the pandemic kicked off, they really started adding stuff to teams then. They were like, Zoom can't beat us. Right.

Yeah, totally. They just went crazy with that sort of stuff. So, yeah, like if they move stuff around there, like moving people's cheese, I think that's less of an issue with teams because I think it's, that interface is pretty, pretty basic. You know, you can move the buttons around. Yeah, it feels like that app has changed every few months for like three years now. Yeah.

And it seems to be fine. Yeah, it does. People are over it. Exactly. Okay. What do you make of the security stuff around all of this? Because obviously, I think one of the interesting trends of the last few years for Microsoft has been it's doing like the work analytics and it's just showing people sort of more of the information it has about them and all of their information and making use of all their documents and all this stuff. And now to say, you know, pour all of your most sensitive data into this large language model that will spit information back to you that is like sometimes true and sometimes a lie.

And I think the idea of like telling chat GPT to make sense of my company's financials is like a great idea in theory, but it's going to be wrong half the time. And it's like, are we just going to tank the stock market by letting chat GPT look at people's Excel spreadsheets? I know Microsoft's answer to this seems to be like, oh, we're taking it slow, whatever. But what is your sense of how they're thinking this stuff through right now? Yeah, like my personal opinion on this is that I think the Bing stuff is more dangerous. Because you go to Bing, you go to a search engine, you want to find answers that accurate, right? Like you don't have that feedback loop there available for it to go wrong. Like it needs to be right most of the time, right?

If that's wrong, you're putting misinformation out into the world, right? Whereas if I generate a document at work, I use this copilot thing. If it's wrong, that's on me. You know, that's not like I'm putting the essentially the misinformation and it's my name on that document when I submitted to my boss or whatever. So it's up to me to fact check it. And it's less of an issue, I think, in that context. but financial data, if it screws that up, there's some dangerous scenarios where this could go very wrong.

So they have to be obviously very careful there. And I did ask them, like, is Excel like a case where you're, you know, trading carefully, considering it powers the world's economy. But they gave the general, you know, we're being careful. Shets and preview. That sort of thing. Which is like, yeah. But then they must be right.

They must be going careful there with the features that they go in certain apps. And they're doing the thing that Google is doing too, right? which is just like have constant banners that are like, this thing will lie to you. Do not trust anything this thing says. Exactly, yeah. And it's like, yeah, this is your confidential information going into this. Make sure you fact check it. And those warnings only go so far, right?

Like, you see them and then you ignore them after a while. So they're kind of, those warnings are kind of irrelevant. But it's definitely dangerous, I think. I think the speed, because now we've got Microsoft going up against Google, those two button heads over this stuff, it just keeps pushing. until there will be like some disaster scenario that will happen where it screws up like some big bangs data or something like that

and then you know that will be a lesson to be learned sort of thing where they need to slow down but I think it's very different to the search stuff like the search stuff I think I'm personally not happy that they're going so fast with that stuff because that stuff can do real world harm outside of your organisation

yeah like it's really down to people using the co-pilot stuff alongside there it's like it should be a tool, right? You shouldn't rely on it to write your whole document for you because that's, that ain't going to be accurate. Right. Yeah, and I think the question of kind of where one ends and the other begins is for Microsoft and for everybody going to be really interesting, right?

Because I think like take this document and turn it into a PowerPoint. Or like you use the example in one of your articles of like, I have blue titles on every slide in my deck and I would like them all to be orange. It's like, that's the kind of thing that takes a long time and is like a lot of people essentially do that for a lot of. living. And then you can make that just a simple command. That I think is like universal victory. Like let's let's solve the interface by asking questions of technology. Right. Like into that. Love that. Sold. I don't know that it's great at all that yet, but I'm very excited about the path

that all that is on. Yeah. But then this question of like how much information can it give me, how much should I rely on it and how much should I allow it to make stuff for me is just a mess. There's the lowest stakes stuff there is all, I think, fine. The highest stakes stuff. there, like, if anyone ever reports AI-generated earnings from their company, like, were host. But somewhere in the middle is that cutoff. And I wonder if there's going to have to be things that it's like Microsoft is sort of artificially, you know, dumbing down some of the stuff these systems can do until they're

much, much better than they currently are. Or if we're just going to get into a place where it's like, the data lies sometimes do with it what you will. And in that case, I think you're right that we're in trouble. Yeah, I think it's also whether we get to the point where it like, makes so little mistakes where it's almost like making human mistakes like when it gets to that level which is far away but then do we accept that because we can't

you can't expect the machine to be 100% like you wouldn't expect a human to be so it's like when does it get to that level as well which I think is some interesting questions around it but it's definitely like that blurred line between what is acceptable and in what context and what app and wherever you're using it and how you're using that data as well Because some of this stuff could be a bit freaky as well, where it starts, like, you know, like, all the AI tools we have right now, sometimes they're a bit, like, alarming. You can look yourself up on Bing, and it's like, it pulls out so much data that you can kind of forget that's out there, right? Bing thinks I'm four different people, which I enjoy very much. This is the delightful thing about having an incredibly common name, is it thinks I'm, I'm the Texas, the University of Texas baseball coach. I'm an actor who was on Frazier. I think I'm like an attorney in Phoenix, and I'm the editor at Large at the Verge.

Like a Marvel superhero or something. It's different backstories. That's what I don't like about this Bing stuff. It's just, it pulls out so much information. Like, I put up my girlfriend's name on it, and it's like, it just lifts off so much stuff. And even, like, gets to the point where it's about to give out her, like, mobile number. Oh, wow. Because obviously she's had her CV out and, like, cashes that stuff.

But it's just like, I think for normal people, like, she's not, you know, in the public eye, for having normal people's data out on it is, yeah. It's going to get strange. Yeah, agreed. So last thing, before I let you go here, is in the stuff you've been writing about the last few months, there have been like little sort of glimmers of what all of this might look like on Windows. And there's the new Bing button in the task bar. But I had this moment of reading the story of like, is the future of Windows that you just hit the Windows key

and just up pops a chatbot? And that's like how you use your computer from now on? Like, is that where we're headed? What's your sense? Yeah, I think so because I think they have the search stuff. right like you just mentioned, but it's just basically a link that just throws you into edge and into Bing. Yeah, like opens edge. Yeah, that's not that. Yeah. So it's not like even really integrate into the search experience. So I think obviously the logical next step is to fully integrate that

into the search experience and into file explorer and all those sort of like touchpoints that you use daily, the start menu, which has some AI stuff already. But yeah, like it wouldn't be surprised me if like, if they did something like Alfred for Windows, but it was just like the chat bot and, you know, pulled out to everywhere. Oh man, that's the dream. That just made me very excited. That would be awesome. Yeah. That would be awesome. Like an actual search interface that was universal. What a world. And you just, yeah, you don't have to worry like where your favorites are or like your app links. Any of that stuff, all that user interface stuff just goes out the window. You just have this search for prompts. That's another one where I would assume Microsoft is going to

have to tread pretty carefully. Although I don't know, they've fully redesigned the start menu like 16 times in the last decade. So like, what's one more? Yeah, they could do a Windows 8, couldn't they again? Just have a full screen chat in that face. That went so great the last time. Like, well, what could go wrong? As long as they make it mouse and keyboard friendly, it could work. This is what I'm saying. Microsoft, you heard it here first. Just forget Windows. It just boots to a chat bot. It's like going back in time, like a command prompt.

Exactly. Yeah. This is, it's everything from the 1970s is new again. It's just like the screens are better now. That's all that's changed. Awesome. Tom, thank you. I appreciate it. This is not over anytime soon, so we'll have you back shortly. Okay. Thanks for having me. All right. That's it for the Vergecast today. Thanks to Monica, Tom, and Jen for being on the show. And thank you, as always, for listening. There's a whole lot more from all of these conversations at theverge.com.

There are a bunch links in the show notes, but also lots more on the website. It's a good website. And if you have thoughts, feedback, feelings, or Taylor Swift tickets you want to sell me, you can always email us at Vergecast at theverge.com. Or call the hotline, 866, Verge1,000. Send us all of your questions, all of your thoughts. We love hearing from you. This show is produced by Andrew Marino and Liam James.

Brooke Minters is our editorial director of audio. The Vergecast is a Verge production and part of the Vox Media Podcast Network. We'll be back on Friday with Alex and Neli to talk about Elon Musk, Twitter's chaos, and everything else happening in tech. See you then. Rock and roll.