Summary:
In this episode of the Sidecar Sync, Amith Nagarajan and Mallory Mejias unpack Google I/O 2026 and what it signals for the future of AI-powered work, search, and member engagement. They explore Google’s push toward proactive, agentic AI across Gemini, Workspace, Search, and new infrastructure like Antigravity and TPU chips, while digging into what these changes mean for associations trying to protect their content, improve digital experiences, and stay relevant as members increasingly expect voice, multimodal interaction, intelligent search, and personalized service. The conversation also covers AI’s impact on career advice, leadership, web traffic, SEO, smart glasses, privacy, and why associations may need to double down on trust, niche expertise, and human connection in an increasingly agent-driven world.
Timestamps:
00:00 - Graduation, Career Advice, and the AI Labor Market
👥Provide comprehensive AI education for your team
https://learn.sidecar.ai/teams
📅 Register for digitalNow 2026:
https://digitalnow.sidecar.ai/digitalnow
🤖 Join the AI Mastermind:
https://sidecar.ai/association-ai-mas...
🎀 Use code AIPOD50 for $50 off your Association AI Professional (AAiP) certification
📕 Download ‘Ascend 3rd Edition: Unlocking the Power of AI for Associations’ for FREE
🛠 AI Tools and Resources Mentioned in This Episode:
Eric Schmidt Booed During Graduation Speech About AI - https://shorturl.at/bm7OZ
Google Glass ➔ https://en.wikipedia.org/wiki/Google_Glass
Google Gemini ➔ https://gemini.google.com/
Google AI Mode ➔ https://search.google/ways-to-search/ai-mode/
Google Workspace with Gemini ➔ https://workspace.google.com/
Google Antigravity ➔ https://antigravity.google/
Google AI / Gemini Omni ➔ https://ai.google/
Google DeepMind Gemini Models ➔ https://deepmind.google/models/gemini/
OpenAI Codex ➔ https://developers.openai.com/codex/
Claude Code ➔ https://code.claude.com/
MemberJunction ➔ https://memberjunction.org/
https://www.linkedin.com/company/sidecar-global
https://twitter.com/sidecarglobal
https://www.youtube.com/@SidecarSync
⚙️ Other Resources from Sidecar:
More about Your Hosts:
Amith Nagarajan is the Chairman of Blue Cypress 🔗 https://BlueCypress.io, a family of purpose-driven companies and proud practitioners of Conscious Capitalism. The Blue Cypress companies focus on helping associations, non-profits, and other purpose-driven organizations achieve long-term success. Amith is also an active early-stage investor in B2B SaaS companies. He’s had the good fortune of nearly three decades of success as an entrepreneur and enjoys helping others in their journey.
📣 Follow Amith on LinkedIn:
https://linkedin.com/amithnagarajan
Mallory Mejias is passionate about creating opportunities for association professionals to learn, grow, and better serve their members using artificial intelligence. She enjoys blending creativity and innovation to produce fresh, meaningful content for the association space.
📣 Follow Mallory on Linkedin:
https://linkedin.com/mallorymejias
🤖 Please note this transcript was generated using (you guessed it) AI, so please excuse any errors 🤖
[00:00:00:14 - 00:00:09:17]
Amith
Welcome to the Sidecar Sync Podcast, your home for all things innovation, artificial intelligence and associations.
[00:00:09:17 - 00:00:23:09]
Amith
My name is Amith Nagarajan.
[00:00:23:09 - 00:00:25:03]
Mallory
And my name is Mallory Mejias.
[00:00:25:03 - 00:00:38:18]
Amith
And we are your hosts and we have a lot to talk about today. We've got an action packed episode. It'll probably be a little bit longer one. So buckle up and let's get rolling with that. Mallory, how are you doing today?
[00:00:38:18 - 00:00:54:22]
Mallory
Amith, I'm doing very well. We are recording this May 20th right now. And I feel like the month of May is always a good month right before summer. You and I were talking about graduation season. I know you have a son that's graduating from high school. How are you feeling about that?
[00:00:54:22 - 00:02:09:01]
Amith
Feeling fantastic. You know, he's a good kid. He's off to college on the west coast and he'll be doing great. So I'm excited to see him launch and it'll be interesting to see what happens. You know, we talk about a lot of different things. But one of the subjects actually is, surprise, surprise, AI and how that will affect the labor market by the time he graduates from college in hopefully four years. And so we'll see. But I've tried to give him sage counsel, which I have no idea if there is such a thing in this day and age and certainly with a four year time horizon. But I've told him to meet a lot of different kinds of people, learn a lot of different kinds of things, become good at communication and try to learn a couple of hard skills as well, whatever they are. And we'll see how that works out for him. So he's going to be doing well, I'm sure. But it's just fun to see changes happening. Graduation is kind of that part of the rhythm, right? Where people are having a phase change from one aspect or one phase of life to another. And I think that's an apt analogy in some ways to how associations might need to think about themselves graduating from legacy business models to an AI native business model and thinking about a fresh look at how the world should look to their eyes when they wake up the next morning after that.
[00:02:09:01 - 00:02:27:20]
Mallory
I like that analogy you made. My husband and I were actually talking about just this past week, what advice to give to a young person planning their career, especially in the age of AI. But I like what you said, meet different people, learn about different topics, improve your communication, maybe a couple of hard skills sprinkled in. I think that's a good way to put it.
[00:02:27:20 - 00:04:20:14]
Amith
Yeah, it's hard to know what's going to be in demand or not, but I think relationships amongst people will always be in demand. And forget about technology, forget about psychology or sociology, just think about biology. We are wired to seek community and connection. And it's one of the most beautiful things about the human race is that we work together and collectively can achieve more than any one individual. And as we scale that, that leads to potentially all sorts of amazing things. And that's actually one of the things I love about the association market is associations, just think about it at the most fundamental level of the name association. It's to associate, we want to bring people together for some common goal, some common objective, and we want to be able to collaborate, we want to be able to learn together, we want to be able to grow together. That's a beautiful thing. And so I think AI might change the work that the people are doing, but it can accelerate the mission. And so I think it's pretty cool. But ultimately, when I think about young people or older people who are rethinking life and what's going to happen because that reset, that graduation, the phase change essentially, it's going to affect all of us and perhaps many, many times in the coming years because so much is happening. That many of the things that we consider to be like, perhaps our life's work are no longer necessary, right? So for example, imagine if you're someone in the drug discovery business and perhaps AI solves, you know, all disease, that sounds fantastical and sci fi, but imagine if that's possible. And I think it is in the next 10, 20, 30 years. What are you going to go do with yourself? I'm sure there'll be something productive for people with that amazing level of skill to do. But that, you know, is that is a different problem, right? So it means that there's some degree of perhaps crisis, but also opportunity that lays in front of all of us, not just kids that are graduating from high school and college.
[00:04:22:01 - 00:04:56:18]
Mallory
I like that realistic yet optimistic outlook, Amith. But speaking of graduation, we both saw a headline from this past week that former Google CEO gave a commencement speech. Now, neither you nor I have watched it yet. I plan to. But on the stage, he was booed. So Amith, what is your take on that? I feel like we're very fortunate that the audience we speak to tend to want to hear, I hope, the things that we say about AI. But I imagine that would be kind of, I don't know, that would shake me if I were giving a commencement speech and then I was getting booed during. What about you?
[00:04:56:18 - 00:07:13:07]
Amith
Yeah, you know, I typically am not booed when I'm giving keynotes, but I have had situations and I actually kind of seek them out a little bit where people are uncomfortable. And I'm kind of ruffling their feathers. For example, I'm probably pretty well known in this market by now for saying that if you're not educating all of your employees and all of your volunteers on AI, you're committing leadership now practice. I deeply believe that to be a true statement. But it kind of upsets people because if they're not fully on board with that, they might not understand exactly where I'm coming from. I lead up to that statement usually. But I think that being provocative is an important part of doing your service as a speaker. It's not necessarily just to say the feel-good things that everyone wants to agree with and nod their heads to. In the case of Eric Schmidt, the former CEO of Google, he was giving a commencement address, I believe, at the University of Arizona. And I believe was booed almost throughout his speech. And he apparently was taking a measured approach to AI. It wasn't, you know, an AI super optimist view or an AI cautious view, but just kind of an AI realist view that, "Listen, AI's here. You're all going to need to know it. I encourage you to learn it, to embrace it, to use it." And I think that's a pretty realistic view. If you ignore the technology, it's probably not going to work out too well. So it's just an interesting kind of moment in time to hear people graduating from college feeling that way. Now, the flip side of it is employment rates for newly minted college graduates are not particularly strong right now in a number of fields. But particularly in technical fields, actually, like computer science, which not too long ago was almost a guaranteed path to having employment locked up well before graduation. There are tons of computer science majors from good universities who are not employed yet. And that was true last year as well. And this is just one field I happen to be very familiar with, but there's many others that are facing even deeper crises. So I can empathize with the graduates saying, "Hey, listen, I spent four years and probably up to my eyeballs in debt. And now you're telling me that AI is going to do everything I just learned how to do." That would basically suck. So I could see how that people would feel that way. But at the same time, I think Schmidt did a good job. And again, I haven't heard the actual speech, but the concepts of what he's telling people to think about, I think maybe unpopular but important.
[00:07:14:22 - 00:07:35:16]
Mallory
I agree with that. I also agree that you sometimes seek out the provocative messaging, but it works. And I think as a leader, at times, maybe all the time, you can't just do what's popular, what makes people nod their heads, what gathers consensus. I don't know. That's doing a lot of the same. I think to be a leader is to say something new. And sometimes that is provocative or ruffles feathers, like you said.
[00:07:35:16 - 00:09:06:21]
Amith
Let me just think about, think deeply for a minute about these different words we use kind of in passing. We talked about association to associate and leader should lead. And leading means generally going forward, going somewhere that you haven't yet been often, right? Picking the path, choosing the trail, determining which mountain top to climb. These are the things that leaders do. And so to do that, you have to be at least somewhat bold. You have to be willing to take that first step on behalf of your organization, your family, yourself, whoever it is, and whatever group it is you're talking about. So leaders often need to be in this uncomfortable space. So I think it's a moment in time, hopefully for people to reflect, not to just say, "Oh my gosh, I can't believe this billionaire, former CEO of Google was so insensitive to these kids. 30% of them are unemployed that he's saying AI is going to do all these things." I think maybe that's one point of view that's reasonable to at least reflect on for a moment. But the other point of view is he's trying to help these people recalibrate their perspective, if perhaps in an unpopular way at the moment for that group, but to do so in a way that helps them move forward in a more productive way. And so even if a very small handful of the people who listen to him speak had their mind shifted even a tiny bit, he's accomplished his mission, right? He's trying to share something that's useful to people, not just get a bunch of applause. I don't think that Eric Schmidt at this point in his life is that worried about applause. I don't know the man, so maybe he is, but my suspicion is he's trying to do a service there by appearing at this university and giving a commencement address.
[00:09:08:10 - 00:13:12:14]
Mallory
Well, Amith, it doesn't always happen this way, but I love when our chat at the top of the episode leads really well into the episode topic. Today, we're talking about Google I.O., which is Google's annual developer conference held every May in Mountain View, California. It's where Google announces its biggest product launches for the year and for the past several years, that's meant AI. Day one of I.O. 2026 was a single theme keynote to Google going all in on proactive, agenic AI across every surface, less, a little bit less about new model launches than last year and more about Gemini doing real work in the background. So today, we're going to walk through one, the Gemini app and workspace overhaul, two, the biggest search redesign in nearly 30 years, and three, what's under the hood, the infrastructure powering all of this and where it shows up next. So topic one, everyone buckle up. We've got a lot to cover. I'm going to talk about the Gemini app and workspace going proactive. So at last year's I.O., Gemini was at 400 million monthly users. Today, it's at 900 million across 230 countries. The headline of day one was a top to bottom rework of the Gemini app and workspace framed as a shift from an assistant that answers your questions to an agent that does the work. The first model in Google's next generation 3.5 family combining frontier level reasoning with very low latency is Gemini 3.5 flash Gemini 3.5 Pro is expected next month. How it stacks up against the frontier on Google's published benchmarks, the flash tier model beats pro tier competitors like Claude Opus 4.7 and GPT 5.5 on agent benchmarks like MCP Atlas and multimodal benchmarks like the MM M U Pro. It runs roughly 4x faster than Opus 4.7 and GPT 5.5 and is priced about 10x cheaper than Opus 4.7 and 3x cheaper than GPT 5.5. There are some trade offs flash loses to GPT 5.5 on hard abstract reasoning and to Claude Opus 4.7 on the toughest real world coding tasks. But Google is clearly leaning into flash as their workhorse rather than leading with a pro release, which is a notable strategic choice. We're seeing neural expressive redesign within the Gemini app. This is a new design language rolling out globally on Android, iOS and web so that responses now include images, graphics and narrated video alongside text. You also get easier switching between typing and talking via a live button so a richer Gemini live conversation experience. They also launched the daily brief a personalized morning digest that pulls from your inbox, your calendar and your tasks and organizes them into a prioritized overview with suggested next steps. Gemini Spark is a big announcement. This is a 24 seven cloud based personal AI agent built on Gemini 3.5 flash and Google's anti gravity agent harness because it runs in the cloud. It keeps working when your phone is locked or your laptop is closed. Some examples that Google demoed are scanning credit card statements for hidden subscriptions monitoring school emails for deadlines turning scattered meeting notes into polished docs and follow up emails. Spark will also be able to make purchases on your behalf in coming months. Google says that it will ask permission before high stakes actions are taken like spending money. Next thing I want to cover is the Gemini Omni model. This is a separate model family from 3.5 flash. So 3.5 flash is the reasoning language model behind spark and search. Omni is the generation model behind video image and audio creation. Omni collapses what used to be separate Google models video for video nano banana for images into a single unified system. It takes text image and audio and video as input and it can output video with image and audio output planned.
[00:13:13:17 - 00:13:59:19]
Mallory
Finally the workspace voice layer. So voice is becoming a primary input across Gmail through Gmail live docs through docs live and keep. So you can ask spoken questions of your inbox. What's my flightgate man that would have come in handy for me. I was just in the airport this weekend. It can dictate ideas that the AI structures into a doc and grounds in your drive Gmail or chat or you can brain dump and keep and get organized notes back. So I mean this is believe it or not this is just topic one of the episode Gemini 3.5 flash is now the engine running AI mode in search the Gemini app and spark. And it seems Google is leaning hard into flash rather than leading with the pro release as I mentioned. So what do you think about that.
[00:13:59:19 - 00:16:12:06]
Amith
Well flash is the workhorse and it's actually been true for some number of releases now. I don't remember the first version they released flash might even going back to one or one point five but flash is a really interesting model. It started off as the really fast model is the name flash would indicate that actually wasn't that smart. And as we've talked about Mallory for some time models at the smaller sizes keep getting smarter and smarter and smarter. While they retain their cost and speed advantages and Google certainly done tons of work here. But at the same time that the thing that to point out is 3 5 flash which is their mid level model is smarter than Opus 4.7 and GPT 5.5 in many categories. You pointed out a few it's slightly behind in a couple of categories but not terribly so in that in the hard abstract reasoning and the hardest of the hard software development questions. It is behind slightly Opus 4.7 and in one or two cases I think GPT 5 5 but it's ahead on everything else. And in addition to that if you compare GPT sorry Gemini 3 5 flash to Gemini 3.1 Pro Gemini 3.5 flash is better pretty much across the board by a pretty wide margin. And it's dramatically faster and cheaper. And remember that Gemini 3 1 Pro just came out. It just came out. So the relentless progression of AI the compression curve of cost relative to quality and speed it just keeps on going and it's even faster than any of us have been predicting. So I find it very exciting. So what does this mean as a practical matter. It means that when we talk about frontier level intelligence the smartest of smart models are now really fast. So Gemini 3 5 flash outputting at 300 plus tokens per second is a big deal. It's time to first token or TTFT is the industry acronym since we all need more acronyms in our lives. TTFT that's an important acronym to remember with model latency because it has to do with how quickly can the model start responding. So if I give Gemini 3 5 flash you know a whole book to read and needs to respond. Does it take a minute to respond or just a handful of seconds. And Gemini 3 5 flash is also really good in this category.
[00:16:13:09 - 00:18:07:05]
Amith
Whereas if you compare it to like Opus 4.7 if you give it a whole book it'll take quite a long time to start getting a response. What does this mean. It means that the intelligence frontier is also moving to modalities like voice. And as we've talked about also on this pod a number of times voice is going to become a bigger and bigger and bigger deal. And you're seeing that happen now with one of the major both consumer and B2B brands putting voice you know kind of as a first class option across their suite of products. Ultimately though coming back to the core of your question not leading with pro and leading with flash. I think is really smart because that's where the majority of the workloads are going to go. They're trying to position 3 5 flash as the everything model every day and then 3 5 pro when it comes out. I'm sure we'll have extraordinary benchmarks. I'm sure it'll beat everything at that point whatever is out there they'll make sure that's the case. But it'll be kind of like sipping fine wine. It'll be the thing that you do when you need it. It won't be something that you gulp. It'll be something that you sit. And so that's how they're positioning pro whereas a lot of times that people have been doing building agents is they've essentially said well I've got to have the smartest model all the time. Let me use Opus 4 6 4 7 Gemini 3 1 pro. And the reason for that is is the intelligence was necessary. You don't necessarily need the most intelligent model once something like flash is as smart as three months ago the very smartest model in the world. So to give you a concrete example for applications and agents that we build skip Betty Izzy these are agents that do very sophisticated work very complex work for associations work that can't be screwed up has to be correct. And until now we've only used kind of the top tier models for the hardest portions of those agents that's going to change. We're going to make skip by default use Gemini 3 5 flash we've been using 3 1 pro for the hardest coding tasks that skip does the 3 5 flash should be totally sufficient we're testing it right now actually.
[00:18:08:14 - 00:18:44:16]
Amith
But I think there's places where we'll use pro as well, but that is great because it means the cost is coming down. It means the speed is better since they're having to wait 7 to 10 minutes for an average response and building an interactive component it might take you know a third of that time to do it which is which is great you know closer we get to real time the better. The lower the cost gets the more accessible this is and then more broadly speaking when we talk about you know more complex reasoning tasks being able to do them at scale across the entire corpus of content that an association has things like that you need more power and you need to be fast and cheap, so this is this is a pretty big deal.
[00:18:44:16 - 00:19:14:16]
Mallory
I feel like in the past we've always you know simplified it for the sake of the pod as you got the smaller models that are fast but they're less smart and then you've got these huge models that are smarter but slower and more expensive but it seems like Gemini 3.5 flashes is different in a different league because it's. Smarter or as smart as the frontier models like opus 4.7 and gpt 5.5 but it's faster and cheaper so is this kind of a whole new class of models are we going to even have a use for the massive frontier ones in the future.
[00:19:14:16 - 00:21:20:04]
Amith
I think we will I think that the most complex most sophisticated models will always be pushing the frontier but the frontier will most likely be in disciplines like you know deeper forms of scientific discovery. Really novel things like that what we consider like ordinary white collar jobs reasoning across marketing and finance and even CEO work. I would argue can be done quite well with the current frontier models and as the models now become faster and cheaper you'll take what is currently frontier you'll package into something that will eventually run on your phone for no cost probably within a couple years. And you'll be very happy with that for most things because you know what's the point. What like what's the point of using a model less dramatically bigger than what you need you know, for example Mallory if you're going to fly from Atlanta. To San Francisco I don't think you're going to load up you know you're gonna say hey I need my very own Boeing 747 jumbo jet to fly me across the country just for you right that's like a massively oversized aircraft relative to your mission you need something much smaller. Or it's a big stand that even further you wouldn't want to like if you have a single letter that you want to mail to someone you probably don't want one letter to go on a jumbo jet across the country right so. You need to match the workload with a model that fits the workload it's the same thing as matching you know a semi truck with a full load of stuff to deliver versus. You know smaller vehicle with a smaller you know payload to deliver since the same kind of concept there's many many workloads across so many of the projects that we want to do in the world I don't mean we as blue cypress but we as a community. That require intelligence no further along than what we have in the current frontier so now it's going to be yes the frontier pushes along gets more amazing more intelligent. But they're the current capabilities are so outstanding so incredible that when we compress those a lot of people are gonna say this is fine and by the way. I'd rather have it be 100 times cheaper and not any smarter because it's smart enough I just want to be 100 times cheaper and 20 times faster that's kind of what we're going here with with flash.
[00:21:21:18 - 00:21:44:07]
Mallory
You touched a little bit on voice and I want to double down on that because I think the way we as humans are so accustomed to interacting with technology is through text so that is just naturally the lens that we think about. Offering services to our members is through knowledge assistance that we can chat back and forth with do you think voice is the future of the way humans interact with technology do you think it's something else that we can imagine.
[00:21:45:07 - 00:23:10:10]
Amith
I think voice is a big part of it you know I think when we interact with other people. Certainly we can share documents with each other we can write but we do tend to speak as a very natural way of communicating so for us it's a convenient and efficient way to communicate with other people and also with machines now. So I think it's part of it I think there's certain things that make more sense in text and in images, for example, if I'm gonna share with you a whole bunch of numbers probably showing them on the screen would be very helpful for you to read them. But if I then describe to you what those numbers mean and say hey Mallory here's a table of membership growth over the last 15 years let's study this table together and I start explaining to you what I see in the table that might be interesting and what I think maybe we might want to draw. As an observation from that data that would be an interesting voice over whereas the data itself perhaps would be best shown to you visually either in text or perhaps in a graphic or something. So I think the world is really going towards this multimodality concept which of course is a good segue to Omni in a minute that the idea here ultimately is yes voice is one of the modalities that's going to get a lot of growth because it's hardly used at all. I talked to some of the you know most cutting edge people in AI regularly inside and outside of the association market and I asked them a common question I asked them is hey how often do you use the audio mode on the cloud mobile app or the chat GPT mobile app. And what do you think the answer that is Mallory have you asked that question of people you run into.
[00:23:10:10 - 00:23:19:11]
Mallory
No I have not it's funny because I feel like you've been asking that question since the beginning of all this stuff a few years ago but I would imagine not many people are using it.
[00:23:19:11 - 00:24:37:19]
Amith
That's generally what I find like I if I didn't I haven't counted the answers or anything like that but I would guess that it's less than a quarter of the people that I asked that question use it regularly. And you know probably less than half of even tried it but less than a quarter are regular users now I don't use voice instead of text but I use voice when I find it convenient and there's there's two cases that I would say voice are really is really powerful one is when screens are not helpful available or safe like if I'm on the move. You know walking in New Orleans is a hazardous sport even with your paying 100% attention you most certainly do not want to be looking at a screen you'll be you'll very likely be hit by a car or or you will be absorbed by a giant pot hole if you walk in New Orleans while you're distracted by cell phone but you can certainly talk to another person or an AI. I do that all the time here another example is driving you can certainly talk to an AI just like you might talk on the phone and I find it quite helpful for that. But independent of that actually one of the things I love to do with voice even if I'm fixed in location is to brainstorm ideas so I find that if I'm thinking through something I'm not if I have like a particular task I'm like hey I want to knock out a B and C I go type it up and bam it off it goes I come back a few minutes later and I see the results and it's usually good but if I want to think through something and iterate something I love the voice option because it actually just causes me to think differently than if I'm just typing and reading.
[00:24:37:19 - 00:24:59:06]
Mallory
Yeah you know when you discuss something with a with a fellow human send you talk it out quote unquote and you really okay now I understand how I feel about this better than before I talked it out I agree with you I use it to brainstorm all the time and just when I'm lazy like if I know it's a long drawn out prompt in my mind I don't want to type that out so I'll just use my voice but I'm a big fan.
[00:24:59:06 - 00:25:08:21]
Amith
Or imagine if you're working on a project like let's say you're trying to build a new web page for launching an event for your association and you're thinking through like how should this event page be constructed.
[00:25:09:21 - 00:26:20:22]
Amith
Getting on a zoom call and sharing your screen and looking at the website together and saying well we should move this piece over here to here or you know or imagine yourself at a venue and you're trying to think through the best way to organize. The stage for a keynote or something like that being on site and physically seeing the space and working together is helpful now you're talking you're also probably taking notes by hand you might be looking at pieces of paper that talk about capacity and. You know projections for attendees and all these things so these are all multi modal experiences right there's physical space there's virtual video space audio text. So, etc, and as the AI can start to come along with us into these other modalities it started off with just text that I moved into images and now it's moving into voice and video. Soon, I think, you know, with these things like augmented reality glasses that all the major players are getting into you'll be able to take the AI with you most places you go and then the AI being able to experience. You know as much of the world as possible that you want to allow it to experience can be a richer thought partner for you in all those situations so to answer your question I think voice is a very big deal and we're. We're investing like crazy invoice sending voice needs to be in every experience you have it needs to be an option so I love what they're doing with Google live.
[00:26:22:07 - 00:26:54:01]
Mallory
And then with the Omni model the generation model behind video image and audio creation so that multi modality that you were just talking about me. Do you think that is going to become the consumer expectation or how they interact with the web and finding information and I'm asking that because from an association perspective. Should associations that maybe just launched a knowledge assistant that is text only should they be thinking about ways to incorporate video and audio into that experience now.
[00:26:54:01 - 00:27:08:02]
Amith
So now I'll digress briefly but I'll come back to that one of the reasons I'm so optimistic about. humanity is that we adapt really quickly and particularly new generations, but even people have been around for a long long time.
[00:27:09:03 - 00:28:15:20]
Amith
I'll give you an example when the iPad first came out and people started just touching the screen and like moving around. There was no training there was no need for people to like learn that other than you just like touch the screen and you made it work and you just kind of expected everything to have touch screens. And then you hear stories of like kids that are used to their iPads they walk up to the TV and they start trying to like move things around on the TV right and I think a lot of people experience that. The same thing is true for these technologies so people are going to assume that ais they interact with can talk to them can listen to them can express ideas in images and even in video. Very very soon so it's going to take you know the killer app experience and maybe this Google Omni moment will be a moment in time when people start to experience like this full multimodality concept. We're all just blends together and then ultimately it's just an app it's just AI even kind of becomes a background term we know it's AI. It's just a computer and it's just a way we interact with a computer and it's this really rich comprehensive encompassing experience that includes all the modalities and there's different tools that are dependent upon you know they're used dependent upon what makes the most sense.
[00:28:15:20 - 00:28:27:01]
Mallory
And so for the association part of that question you think because we adapt so quickly associations need to hop on that bandwagon because that's what you're going to expect okay.
[00:28:27:01 - 00:30:44:22]
Amith
Yeah the expectation is going to be a rich interactive experience if you think that your association felt old school because you didn't have a website for a long time or you didn't support. Online transactions like member sign up remember renewal for a long time and most associations have that now but there was a pretty big lag from when e-commerce became like a broadly adopted thing for consumers and when association starts allow their members to do these things online. So if we allow that lag to happen again here and in some cases it was a decade right before people have those capabilities in their association. If we allow that to happen here associations are really going to seem like a relic to most people and they won't bother because the issue is is if you can have this truly extraordinary experience in Google and from open AI and from anthropic and perhaps from other players and it's rich immersive and the quality is really really good. Just gonna use it for everything and you know no the Google engine doesn't have access to your wealth of content in your domain it doesn't have your community it doesn't have your trusted brand however. If it is so far along in terms of its experience and you're so far behind that's a problem you can't bank on your brand and your content as being the moat that's going to protect you from the pace of progress. I don't think any association thinks that at this point but it also means that we have to move this community along fairly rapidly. We've got to do that through small experiments we've got to be doing things with voice things with images things with video and see where they work well in our communities and every community is a little bit different so what works well at a specialty medical field might not work well in an engineering discipline or a trade association so that's where these constant experiments are important. I know by the way on that topic I do believe it's important to listen to your members or listen to your customers I also think it's really important to try stuff to see what happens because if you just go ask your members what do they want. They often don't know because they don't know what these experiences should be but if you go show it to them and say hey this is something we're playing with and if it's easy to do that and they go no that's totally terrible or they say oh my gosh this is amazing it's changed my workflow and I want to use this every day. Anyhow you can know you have something right so I think you have to do both you have to listen empathetically and intently to your customers or members and you have to lead with some vision you have to be want to throw some stuff out there and see what sticks.
[00:30:46:00 - 00:32:07:02]
Mallory
Alright I want to segue into topic two which is google search getting its biggest overhaul in nearly 30 years. Google is calling this the biggest change to search since the search box debuted AI mode is now powered by Gemini 3.5 flash globally and has crossed a billion monthly users. AI overviews is at 2.5 billion the pitch is that search is no longer a list of links it's an interactive agentic generative experience. So first I want to talk about the intelligent search box the classic blank rectangle is being replaced with a dynamically expanding box that accepts text images files video and even open chrome tabs as input. AI powered query suggestions go beyond auto complete by anticipating your intent shortcuts underneath the box surface AI mode talk so that search live and create and that's nano banana and Google lens. They're also information agents, so these are 24 seven background agents that can monitor the web for changes on whatever you care about and then send synthesized updates with the ability to take action. Google described this as Google alerts rebuilt with a frontier language model examples include continuously scanning apartment listings against detailed criteria monitoring stop movements with specific parameters and tracking sneaker drops if that's your thing.
[00:32:08:03 - 00:34:00:19]
Mallory
Also generative you I and many apps so generative you I builds custom widgets dynamic layouts and interactive visualizations on the fly in response to your queries. So a question about black holes might return an interactive visual not a paragraph it's going to be free for everybody to use and then many apps let's a user describe a tool in natural language. Custom fitness tracker wedding plan or whatever that may be and search will code it on the spot using anti gravity apps can be saved and shared. There's also personal intelligence the feature that connects gmail Google photos and soon calendar to AI mode is expanding to nearly 200 countries and 98 languages that's also free no subscription required. And this will this is the piece that let's search know your context of course you can opt in app by app on that. There's also a gented a gented booking that's where search can complete a transaction for you it's expanding to home repair beauty and pet care services in the US this summer. And then this is always where my mind goes I thought it was apt to mention the SEO and traffic impact so systrix is a German SEO analytics firm they analyzed over 100 million keywords and found that when an AI overview is present. Click through rate at position one drops from 27% to 11% a 59% relative decline they estimate Germany alone is losing 265 million organic clicks per month from AI overviews. There's an important nuance here the impact varies dramatically by category so informational content like health parenting how to's gets crushed transactional pages like for bookings and e commerce barely move. Gartner has projected that traditional search volume will fall 25% by the end of 2026.
[00:34:02:02 - 00:34:14:10]
Mallory
Holata unpack your meath Google's essentially telling websites we're gonna summarize you monitor you on someone's behalf and even build apps in front of you and you may not get the click so what does that mean for associations.
[00:34:16:02 - 00:34:38:08]
Amith
It's a great question I mean first of all this is probably all pretty obvious to folks now that Google was gonna adapt and figure this out. It's a lot riding on this the volume of search they have is directly court correlated to their primary revenue stream and so. Now there's there's gonna be opportunities by the way for I think some really interesting advertising products in the mix here because of the depth of context that they'll have.
[00:34:39:09 - 00:38:34:10]
Amith
On an individual user will be truly extraordinary and allow you to do some similar things as an advertiser which may be an opportunity for associations to target. New audiences as well as going deeper with existing audiences come back to your question about the displacement issue yeah it's an issue for everyone, so if you're in the information business and Google can do all these things now why would. A user on the web bother coming to your website. And you can kind of let that question hang for a second there has to be some differentiated value there has to be something about you that makes them want to take the time to come to you, they have to keep that friction as low as possible but. If your website is hard to use and it's hard to find stuff and there's good enough information on Google or on chat GPT or these other alternatives people are just going to stay there we've been saying this for a long time well before this release this announcement. So this just amplifies that so my perspective is that the association needs to double down on trust they need to double down on the depth of extremely. Specific content that they produce in their in their space associations are all narrow casted niche players right they're not focusing on boiling the ocean of all content it's typically focus on one particular very specialized lane of content. That that lane of content typically includes publications could be journals could be less formal types of things like magazines or blogs. Content in the form of conferences, which I think is an opportunity for associations to grow a ton. So to me more demand I think for people to connect face to face or online and I think associations need to really go deep on asserting themselves as the experts in their field. More so than ever and then finding a way to say look we're going to share a little bits of that content so that to the extent that there is attribution and click through from these various. AI engines, including Google will be able to get some some volume of traffic from that, but more importantly we want people who find us to say hey there's a differentiated value right when I go to this association I know I can trust. This content did not come from reddit not come from Twitter did not come from these other places but rather came from a place I trust the other thing I'll add to the mix is when it comes to. The associations ability to connect people I think that you know associations have been saying that forever that they're here to connect people and and work together towards some common goal. And we've been talked about that earlier in the episode yet the way we do it is pretty old school it's pretty generic it's not proactive it's not personalized. I think it's a big opportunity here for associations to lean hard into the relationship side of the business to focus on connecting Mallory with people Mallory would like to connect with in a trusted safe environment that has other professionals with common interests right or whatever the relationship is. It might be an enthusiast organization that is you know built around some other some other non professional interests, but the point is associations can create this space. That allows people to connect which I think people are going to be craving more and more with all the stuff going on, so I think that's an opportunity we've been talking about that forever personalized engagement connecting members personalized recommendations at conferences. And I've seen basically none of that or close to none of that handful of people we work with are doing it and quite honestly the technology to do what I'm describing is not difficult anymore it was hard five years ago it's pretty easy now. So I think there's opportunities here for associations to lean hard into what they're best at which is the people side of the business while playing nicely on the tech side you're not going to be better than Google or chat GPT. But you can you can have a good offering a low friction high value offering for people who want to go deep in your field. So that's my current working theory whether or not that's defensible I don't even know.
[00:38:34:10 - 00:38:56:05]
Mallory
No I hear what you're saying but then I'm thinking is it a waste of time for an association to dedicate tons of energy and resources on putting intelligent search on their website so that members can surface information on their website. And that's what I'm saying is that the service information they're looking for more easily if all of these changes with Google means that the member doesn't make it to the website in the first place. What do you say to that?
[00:38:56:05 - 00:39:12:07]
Amith
Well I'd say that it's really important not to hand over your content to Google or chat GPT or cloud because it's a one way door once that content is out it's it's out and the value of your historical content oftentimes in your field is very very high.
[00:39:15:02 - 00:39:38:16]
Amith
And I think that's really important because you're standards these are things that associations have that nobody else has and the world would love to have this content because if that content existed in Google then you're basically done. So I think it's really important that you hold close the most valuable content that doesn't mean you don't put anything out there in the public web you put certain excerpts and you put like summaries of things and you can do that so easily now for all of your content.
[00:39:41:02 - 00:40:13:08]
Amith
And I think that's really important because it's very important that you have the best content to connect with them. But if you are driving people back into your funnel when they're deeply interested in your subjects and you have a really compelling engaging way for them to connect with you then I think those kinds of tools are critically important because even if I do have the best content Mallory if I don't have intelligence search if I don't have a conversational agent that can help me out with what I'm looking for. I probably wouldn't bother. There's a certain limit that I would have in terms of the pain I'm willing to tolerate in looking for what I want to go to the best source of content.
[00:40:14:10 - 00:40:42:00]
Amith
I'll tolerate a little bit of friction there but I'm not going to I'm not going to work super super hard to get to it. So if it's a classical association experience people eventually just say it's not worth it and they'll move on. And that's been true actually for a long long time. It's been true pretty much since the Internet started that there's been a displacement challenge for associations. The value of what's been the alternative has been so far below what associations provide historically that it hasn't been as big of a threat. But now it's an enormous threat.
[00:40:43:03 - 00:40:50:24]
Amith
So I think you have to lead into having a good tech experience. If you don't do that people will be frustrated with you extremely quickly and just move on.
[00:40:50:24 - 00:41:01:01]
Mallory
Lean into a good tech experience as you said and then also if there's a piece of content or information that only your association has probably guard that pretty well I would say.
[00:41:01:01 - 00:41:27:01]
Amith
Definitely. Well it goes to the point we share often on this podcast which is you got to think about your data governance really critically. It doesn't mean that you shouldn't use AI by no means has have we ever said that in this pod. We have to be thoughtful about it. So what you should be thoughtful about is if you're in Claude or if you're in chat GPT or Gemini and you have the opportunity to connect it through MCP or some other mechanism to your SharePoint to your Google Drive to your vast array of content.
[00:41:28:01 - 00:42:23:17]
Amith
Think about what you want to actually connect to these AI systems. There are other ways to do this. There's things like the AI data platform we talk about with member junction. There's other ways of doing it that essentially put an intermediary in place to protect your data so that the AI system like Gemini or Claude doesn't have full unfettered access to your core content. We think that's pretty important. But in the meanwhile if you don't have those kinds of systems in place it's best to be thoughtful about putting limited amounts of your content into those tools because once it's in one of those tools it's in those tools. And again this is not AI alarmism around like you know these are bad people do that are going to do bad things. It just simply means that if you've put your content in these environments you are protected on paper legally that the systems are not going to be you know they won't be training on your content. But that doesn't mean that there won't be some bad actor who does that at some point within one of these organizations or many of these organizations. So I think you have to be really careful with this.
[00:42:25:13 - 00:43:01:01]
Mallory
I want to move to topic three for today which is under the hood and then some announcements that didn't fit anywhere else neatly in the episode. Everything in the first two topics spark information agents generative UI. Many apps is built on a common infrastructure. This last block looks at that infrastructure and at the two places these agents are showing up next perhaps on your face and in your shopping cart. So we've got to talk about chips. Google announced its eighth generation TPUs originally previewed at cloud next with a notable shift for the first time a dual chip approach with specialized architectures for training and inference.
[00:43:02:01 - 00:43:16:11]
Mallory
They launched the or announced the TPU 8T optimized for large scale pre-training nearly 3x the raw compute of the previous generation and the TPU 8I optimized for inference which basically means running the models in production.
[00:43:17:11 - 00:43:20:01]
Mallory
I also want to talk about anti gravity as the common harness.
[00:43:21:01 - 00:44:44:01]
Mallory
Antigravity is Google's agent first development platform publicly previewed in November of 2025 originally pitched as an ID or integrated development environment. The software developers used to write test and run code. It's now showing up as the underlying harness behind almost everything announced today spark runs on it. Many apps in search are built with it and Google's calling out anti gravity 2.0 features tied to the I.O. announcements. This is the foundation that makes a genteck AI work at scale across Google's products. You mentioned I think you mentioned I wear earlier in me. So I thought that would be a good topic to include Samsung Google and Qualcomm previewed Android XR glasses launching this fall. The demo that they shared you're walking past a cafe. You give a voice command and Gemini cues a door dash coffee order on your phone like that. The agentic point is that the glasses aren't just a notification surface like metas Ray bands. They can complete multi step transactions. And then another thing I thought was interesting to include is the universal cart. So a new intelligent shopping cart that works across merchants partnered with Amazon Shopify and Walmart backed by a new open standard called the universal commerce protocol or UCP. The cart works in the background after you add something finds deals tracks price history alerts when items are back in stock and it catches compatibility issues.
[00:44:45:01 - 00:44:52:06]
Mallory
I mean where do you see Google fitting into the great AI chip race.
[00:44:53:09 - 00:46:29:17]
Amith
I think they're giving Nvidia run for the money in a lot of ways. And they actually use a lot of Nvidia chips too. But what you described about this launch it's notable that it's it's exclusively running on their eight I chips. So Google's TPU tensor processing and it's their proprietary chip architecture they've been working on for quite some time. And it's extremely fast. It's extremely efficient and it's really cheap for Google to scale. So there's some really interesting storyline around all that. You know everybody else is having a challenge scaling up. Google doesn't talk about that problem because Google had that pretty well figured out already and their ability to produce the TPUs at scale really outstrips probably everyone else. So there's some really interesting things happening there. A side note related to that is they're partnering with a number of other companies to create third party data centers that will be able to run the TPU architecture and host AI workloads for other customers that are unrelated to the Google Cloud. So for about a year I think at least six months they've been renting TPU capacity to other model companies on the Google Cloud platform. But now there's opportunities here for for other third party data centers to do this. And they have partnership deals with a number of other major players as well. So as a hardware vendor I think it's quite interesting to see what they're doing. Competition tends to drive both innovation and lower cost. It's generally a really good thing. So I'm excited about it. I think Amazon's doing some interesting work with hardware as well. There's a number of chip startups and upstarts that are doing some cool things. So it's a really exciting time in semiconductors for AI. Obviously the prize is so enormous that everybody wants to go after this who's capable.
[00:46:30:17 - 00:46:35:07]
Amith
So I would say Google is right in there with Nvidia and doing some great work. It's really exciting.
[00:46:36:16 - 00:46:46:24]
Mallory
Antigravity is the harness. As I mentioned underneath everything Google is shipping right now. From your vantage point how does their agent stack compare to what OpenAI and Anthropic are building?
[00:46:46:24 - 00:47:00:01]
Amith
I'll tell you more how it compares to Cloud Code and the Cloud Stack because I'm more familiar with that. I have less personal familiarity with OpenAI. But looking at codecs and what they've done there it's pretty similar. So when we talk about an agentic harness what does that mean?
[00:47:01:01 - 00:49:08:11]
Amith
It's essentially saying we're going to take some kind of a brain which is the language model or the AI model and we're going to connect it with context which is a combination of memory and input from the user. Some memory is what it learns about you over time. Input is what you'd ask the to do. And tools which is the ability for the agent to do things. Not just to talk to you but to take action on your behalf. And then the idea of a harness is simply to be able to connect these things in a smart way where the brain is aware of its context and its tools and can achieve goals. So you can say to Antigravity, you can say to Codex, you can say to Cloud Code, do the following. And you tell it what you want as the outcome. You say I want an app that tracks my personal habits. You don't tell it how to build it. It knows that it has certain innate capabilities because it has certain intelligence. It has certain tools where it can create certain things like it can go create the app like on some website. All those things. And it knows it has certain context. It has memory about you. It has context about what you've asked. So the harness is basically a piece of software that is able to facilitate and orchestrate all of these things happening. Ultimately, it's a goal achievement machine, right? You stick a goal in one end and it does the thing for you. And so of course this is a big, big deal. So Cloud Code was really the one, it's a coding tool, but Anthropic realized very quickly, actually the Cloud Code engine is something we can use to power other stuff. So CoWork, for example, is actually Cloud Code. It is Cloud Code. It just has like a pretty user interface on top of it, but it literally is the exact same Cloud Code engine under the hood. They're also using Cloud Code for their other vertical offerings like Cloud for legal, Cloud for finance. They all use Cloud Code as the harness underneath. And the same thing essentially as what Google is saying their strategy is anti-gravity, which now includes the Gemini CLI, the CLI is the command line interface. The Gemini CLI has been incorporated into anti-gravity, but now anti-gravity is just the term they're using for their Cloud Code type of harness. It's more than that. It also has a full software development environment.
[00:49:09:23 - 00:50:52:13]
Amith
And Codex from OpenAI is very similar. So the idea behind this in my mind is every major company is going to have this type of tooling to help you build agents. And there are other options. There's other ways to do harnesses that give you a little bit more independence because each of these harnesses I just spoke of are vendor specific, not vendor agnostic. And that's an important point because if you say, hey, I want to do this stuff for my association, but I'd like to blend a little bit of cloud, a little bit of Gemini and a little bit of some open source models. Because by the way, Gemini 3.5 Flash, super smart, quite fast, but not cheap. It's cheaper than 3.1 Pro, but it's actually not that cheap. Some people have complained about that because Flash historically was about a third the cost of the new Flash. So what if you wanted to have a mixed approach where you had a little bit of Gemini, but you also maybe wanted to use something like Q1 3.7, which is an outstanding open source model. And you wanted to run it on Fireworks or you wanted to run it on Cerebrus or one of these other open cloud vendors who runs open source models. Right. So if you wanted to do that, you need to have a harness that's capable of supporting any number of different models, three number of different vendors. And you can't do that with cloud code without jailbreaking. You can't do that with anti-gravity or with codecs. CODECS is specifically OpenAI, for example. So there are ways to do this. There's a lot of tools that you can do this with. But the thought process behind it is what's important because what I think is really important is having autonomy, owning your own destiny, essentially, and having flexibility and choice rather than being locked into any one vendor. I'm a big Google fan, to be clear. I love what they're doing, but I still don't want my entire destiny to be shaped only by what anti-gravity can do.
[00:50:53:16 - 00:51:09:01]
Mallory
And so to my knowledge, you, Amith, and the team use CODECode heavily to work on products within the Blue Cypress family. I would assume Member Junction as well, which is the open source AI data platform. But you can't do all the work in cloud code, but you can do some of it. Is that what I'm hearing?
[00:51:09:01 - 00:52:04:18]
Amith
Well, no, we use, I mean, cloud code can build entire software framework. So we use cloud code as our primary tool to build software like Member Junction and all of our other software products. But we also use other tools. We do use a little bit of CODECS sometimes. There's certain things it's better at. And we certainly are big fans of Gemini. We use that for a ton of stuff. But that's essentially the point that I'm making for associations is when you think about a harness for your own agents, do you want to be dependent upon one particular vendor or would you like to have a scenario where you have flexibility to work across a number of vendors? And as that changes, you know, this week, Google is a big deal and they have the best stuff. But I assure you that before the end of this quarter, probably before the end of the month, OpenAI and anthropic are not going to let that just sit. They're going to come back with whatever they have cooked up that they haven't released. So it's just the constant cat and mouse game. And the point I would make is you don't know what's best for you until you discover it. And that's going to change over time. So you should have some optionality.
[00:52:04:18 - 00:52:11:23]
Mallory
To wrap up the episode with kind of a more fun question, I feel like most of the people I've seen with smart glasses and meath,
[00:52:13:01 - 00:52:21:23]
Mallory
this is anecdotal, but they tend to be techy people who enjoy technology. Do you have Meta Ray bands? Are you interested in the glasses movement? I've got to know.
[00:52:21:23 - 00:52:58:18]
Amith
I'm interested in wearables in general. I think glasses are interesting. The augmented reality concept is interesting. The ability to like have the glasses have awareness of where you're at and have that additional context. So you don't have to say, I'm approaching this cafe at this location. Please order something for me instead to just have it see what you see. And then to be able to have that context, I think there's some significant opportunities there. The flip side is you have to really trust the company whose wearables you choose to wear. I, for one, am not going to put on a product from Meta because of my personal views of the company.
[00:53:00:00 - 00:53:40:09]
Amith
Google, I'd consider maybe, Microsoft, I might consider a bit like, I don't know. It's one of these things that I'd look at it and go, I need to really know a lot about the privacy side. Because what's going to happen with these wearables is people are going to forget that they're on. And then they're going to go into the cafe after they ordered their latte. And then they're going to sit down with a friend and have a conversation about something that they don't want the glasses to be listening to. But yet, the glasses, unintentionally perhaps, or maybe intentionally, are still on. And so that's the kind of thing that I think people are going to be very concerned with. You know, Google was early, early, early in wearables. They actually had this project called Google Glass. I don't know if you've ever heard of that, Mallory. It goes back, I think, almost 20 years or something.
[00:53:40:09 - 00:53:41:13]
Mallory
Okay. No, I don't think so.
[00:53:41:13 - 00:54:57:00]
Amith
Look it up. We'll post a link to it. I'm sure there's some Wikipedia articles out there. But Google Glass was this really cyborg-looking thing. It was this wireframe that you'd put on, and it had this little bit of glass where you could see something. But it did have, I think, some early form of a camera or something on it that could detect your surroundings. There was a big backlash to it. The technology, apparently, was, for what it was at the time, worked. But, you know, people were getting assaulted for wearing Google Glass. I don't think that's going to happen now with these new glasses. But people are very sensitive to this, and understandably so. So I think it's an interesting thing. Another question would be, from the association point of view, how do you feel about people wearing these augmented reality glasses to your events and recording your content? What do you think about that? That potentially is an issue. Or wear it into a blockbuster film release and record the movie. These are fairly simplistic questions, I think. They're core privacy questions, but they're basic use cases. But what else are the concerns people would have? So I think it's going to happen. I think there's tremendous opportunity here. But I think we have to be very thoughtful about this. This is an area where responsible use, AI ethics, is such a critical topic for us to be thinking deeply about. But it's coming. It's going to be here. It is here. It's going to be pervasive very soon, I'm sure. What do you think about it?
[00:54:57:00 - 00:55:16:00]
Mallory
I'm just looking up the Google Glass. You're right. There is a Wikipedia article. This looks insane. I definitely don't remember that. You know, I've not considered wearables, or at least, like, glasses wearables at this point. I'm going to let you be my cue for that. When a meet starts wearing them, I'll be like, they are safe to wear.
[00:55:16:00 - 00:55:18:01]
Amith
You might have to wait a really long time.
[00:55:18:01 - 00:56:09:09]
Mallory
Okay. And that's fine. I'll do it the safe way. I think there could be a lot of value. I was at a wedding a few months ago, and someone had them on. And my husband and I were chatting about how it is a little bit off-putting to know, like, is it recording me? Or is it pulling up my, I don't know, like, my LinkedIn, matching it to my face, and thinking through things like that. So maybe as a society, we have a longer way to go and getting comfortable with something like that. But I think if you were at an association event and watching a session, and maybe as you're watching, it could be, you know, in the side, pulling up relevant content from the association's website. I think there could be value. Or, hey, if you really like this content, maybe, like, it identifies the keynote speaker across the room, but then you're right. What are the privacy implications of such a thing? And do we need to know that much as humans? You know, do we need that level of information at our fingertips?
[00:56:10:09 - 00:57:49:05]
Amith
Well, I think it's, one of the areas that I think where the information at your fingertips could be extraordinarily valuable is, you know, high-risk, dangerous occupations where effectively a heads-up display for life, which is what this is, becomes a way of reducing fatigue, increasing the right information at the right time, making people aware of safety problems, or making them more accurate in the work they're doing. For example, a surgeon, perhaps, having this on, or someone who's working at a construction site, making sure, the glasses are making sure that they're doing things per the blueprints, right? Instead of making a mistake and finding out about it later. There's a lot of applications where I think at work, there could be incredible value, both from an industrial safety perspective and other use cases. And I think that's where we're going to see a lot of adoption quite quickly, but there still are some privacy issues even there. But ultimately, you know, the value of the technology, if it's sufficiently high, will basically steamroll these concerns. And I just think it's important for people to think through these things before they put them on. And it's cool, it's novel. I think it'd be super fun to try them out. But you have to be very thoughtful about how quickly wearables just exist in your life. People have found this out the hard way about having Alexa devices all over their home and things like that. And, you know, there's been lots of stories about that. Same thing with Siri. Siri sometimes just jumps into a conversation and says, "Hey, I didn't quite get that." It's like, "Well, Siri, I wasn't talking to you." But there's going to be more and more of that. Again, you know, I don't want to be an AI alarmist. I think these are all really interesting things. And the modalities we were talking about earlier, this is a new modality, right? So it's just something we have to experiment with and learn where the value is.
[00:57:50:05 - 00:57:53:01]
Mallory
I don't think anybody would describe you as an AI alarmist, Amit, but we'll keep an eye on that.
[00:57:54:01 - 00:58:11:00]
Mallory
The through line of I.O. 2026 Day One is that Google is reorganizing every one of its products around proactive, agentic AI. And it's confident enough in that direction to redesign the search box for the first time in 30 years. The contrast with last year's I.O., which we covered actually back in episode 85,
[00:58:11:00 - 00:58:16:07]
(Music Playing)
[00:58:27:00 - 00:58:43:24]
Mallory
Thanks for tuning into the Sidecar Sync podcast. If you want to dive deeper into anything mentioned in this episode, please check out the links in our show notes. And if you're looking for more in-depth AI education for you, your entire team, or your members, head to sidecar.ai.
[00:58:43:24 - 00:58:47:05]
(Music Playing)