Latest / In the Loop AI / Google I/O 2026: Their Next Big Thing Is Finally Here

Transcript

Google I/O 2026: Their Next Big Thing Is Finally Here · In the Loop AI

0:00
Josh: The CEO of Google just got on stage yesterday and told everyone that there's no new Frontier model.
0:05
Josh: In fact, he said they've only made it 90% of the way there. So on the surface,
0:09
Josh: this seems fairly disappointing given the fact that it's been many months since
0:12
Josh: our last model release and Gemini was doing pretty great.
0:14
Josh: So while he didn't claim they had the lead, they did proceed to release new
0:18
Josh: models in nearly every single product we use every day and in some pretty creative ways.
0:22
Josh: I mean, three of the highlights are a brand new model that's pretty prevalent
0:25
Josh: throughout all the software we use, an OpenClaw competitor that is built right
0:29
Josh: into Google, and this new model named Omni.
0:32
Josh: And that's where we're going to start today, because Omni is nothing like we've
0:35
Josh: ever seen in the world of AI before.
0:37
Josh: Omni is a pretty impressive model, and I think probably the highlight of this
0:41
Josh: event. So, Ijaz, what's going on with Omni?
0:43
Ejaaz: So they had this really cool moment during IO where they just started showing
0:49
Ejaaz: this announcement video.
0:50
Ejaaz: And people were kind of looking at this and were like, oh, some of this looks
0:53
Ejaaz: AI generated, but some of this actually looks quite realistic.
0:56
Ejaaz: And the reason was they released this brand new model called Gemini Omni.
1:00
Ejaaz: Now, it's a very unique model. It's
1:02
Ejaaz: not something that we have seen before because it's a hybrid of an LLM,
1:07
Ejaaz: and a world model. And world models, if you've heard that term before,
1:11
Ejaaz: is really unique because it can produce video that not only looks realistic,
1:15
Ejaaz: but is physically accurate.
1:18
Ejaaz: So something that LLMs struggle with, it's like that analogy we've used before,
1:21
Ejaaz: Josh, where you have a student in a library that's read all of the books and
1:24
Ejaaz: understands the world, but actually has never stepped outside of the library
1:27
Ejaaz: and experienced the world for himself.
1:30
Ejaaz: World models are basically an upgrade from LLMs where it can teach the LLM to
1:36
Ejaaz: understand how the world works, the consequence of an action.
1:39
Ejaaz: So say if you like hit something or if you jump how gravity works and all those kinds of things.
1:44
Ejaaz: So the unique thing about this model is it can take any input medium.
1:49
Ejaaz: So we're talking about words, we're talking about images, we're talking about
1:53
Ejaaz: video itself and it spits out only one output, which is video.
1:59
Ejaaz: Now the videos that you're seeing on screen here is physically accurate.
2:02
Ejaaz: Now, obviously, it's generated by AI, but the vortex that you're seeing on your
2:05
Ejaaz: screen now, the way that the person draws a circle,
2:09
Ejaaz: the way that the paper kind of crinkles is basically all meant to be physically accurate.
2:14
Ejaaz: And the reason why this is such a powerful model is, well, there's a few.
2:18
Ejaaz: Number one, when you're learning about something, the number one medium that
2:22
Ejaaz: people or humans ingest information is through video. Words is actually kind
2:26
Ejaaz: of slow. It takes a while to kind of like go from left to right.
2:29
Ejaaz: People don't like read all the time in today's world, but they do watch videos.
2:33
Ejaaz: There's several different sensory instruments that come in. There's sound, there's like visuals.
2:38
Ejaaz: You kind of like ingest data way quicker with video, and that's the bet that
2:41
Ejaaz: Google was making with this.
2:43
Ejaaz: And the second thing I'm going to say is, this is pretty world-changing for
2:48
Ejaaz: a lot of different applications outside of just LLM.
2:51
Ejaaz: So, for example, Google has been a massive proponent of robotics.
2:55
Ejaaz: The issue with robotics is, these robots don't understand the real world.
2:59
Ejaaz: And so far, all the robot models have been fed with words about how the real world works.
3:03
Ejaaz: Well, what if you could place that robot brain into a simulation that is physically
3:07
Ejaaz: accurate, like a world model, And you can start teaching people or robots specifically
3:12
Ejaaz: how to interact and they can train and learn a lot faster.
3:15
Ejaaz: So I just think there's many different applications for what this world model is.
3:19
Ejaaz: And it's the first mass scale deployed one. So I'm excited to get my hands on this personally.
3:24
Josh: Yeah, and it's available today live for basically anyone who uses Gemini.
3:28
Josh: It's available in the Gemini app, and it's pretty cool. One of the things that
3:31
Josh: I've enjoyed personally is that you can create an avatar of yourself.
3:34
Josh: So you can actually build a digital clone of yourself and inject it into these.
3:38
Josh: You could think of this model sort of like Nano Banana, but for video,
3:42
Josh: where for the first time ever, you can really create fully custom videos that
3:45
Josh: either take base reality and build something on top of it, or just create a
3:49
Josh: total AI demo in general.
3:51
Josh: So we have this pretty cool demo from the CEO of DeepMind, Demis Hissabas,
3:54
Josh: which show you kind of how this works, where it takes his actual video that
3:58
Josh: was recorded on his camera roll and then turns it into five different scenes
4:02
Josh: that are all completely unique, seeming somewhat realistic.
4:05
Josh: And I think this combines the fidelity of something like Nano Banana with the
4:09
Josh: complexity and physics understanding of VO3. And that's where it really shines.
4:13
Josh: What's also really fun is you have the ability to refine these things too.
4:16
Josh: So you can actually tweak the model, tweak the outputs.
4:19
Josh: And the result of this is something fairly comprehensive, particularly if you're
4:22
Josh: trying to learn things. I find that education is probably one of the most interesting
4:26
Josh: demos and use cases of this.
4:28
Josh: And we have one, which is prompted from a single question of how does photosynthesis work?
4:33
Josh: And now as a student of anything, as someone who's trying to learn,
4:36
Josh: you're given this fairly comprehensive answer that starts with an image.
4:40
Josh: The image is generated by Nano Banana, and then it gives you the chemical recipe.
4:43
Josh: And then under this how it unfolds section, we actually get a fully generated
4:47
Josh: video from the new Omni model that shows you a like actual visual representation
4:52
Josh: full of graphs and charts and seemingly realistic
4:56
Josh: version and an easy way to learn how this works.
4:59
Josh: So I think this is probably the one of the most powerful use cases if you're
5:01
Josh: a teacher, if you're a student, if you're someone who's just curious about any
5:04
Josh: topic and your preferred way of learning is visually, this is a really compelling
5:07
Josh: product for you. Now this also works...
5:10
Josh: With real videos that you could just upload to Gemini. We did this this morning.
5:13
Josh: I had this video on my camera roll that was just filmed outside of a window.
5:17
Josh: And the prompt was just, hey, make some meteors fall down from the sky and put
5:20
Josh: the Loctis monster in the East River.
5:23
Josh: And the reality was this output, which looks pretty good.
5:27
Josh: I mean, if you're looking at this, you'll notice their sound.
5:29
Josh: You can hear the ambient sound of the city.
5:32
Josh: There's a large splash when the dinosaur or the loctus monster hits the water.
5:36
Josh: It looks pretty good, but also this is very clearly AI generated.
5:39
Josh: It feels like CGI from perhaps the early 2000s. So not quite what we would expect
5:45
Josh: in 2026, but very cool nonetheless that it can actually take one of your videos
5:49
Josh: and generate something pretty powerful on top of it.
5:51
Ejaaz: So if you listen to this, you're probably wondering, like, how can I use this?
5:54
Ejaaz: Why is this applicable to me?
5:55
Ejaaz: Well, if you're just a regular person that kind of wants to play around with
5:59
Ejaaz: video, So this is plugged or integrated right into YouTube Shorts.
6:03
Ejaaz: About two months ago, Google made a very important decision,
6:06
Ejaaz: which is integrating AI into their major video platform, YouTube.
6:12
Ejaaz: A lot of people had mixed feelings about this. Artists in general in the creative
6:15
Ejaaz: world tend to hate on a lot of different AI tools.
6:18
Ejaaz: Google takes a different approach where they're saying the kind of like proactive
6:21
Ejaaz: efforts that a lot of artists are making.
6:23
Ejaaz: If they can produce more content that is applicable to various different niches
6:26
Ejaaz: of audiences, then you can kind of like create a sustainable platform.
6:30
Ejaaz: They see the number one distribution for this being YouTube Shorts.
6:34
Ejaaz: So this is plugged into YouTube Shorts right now.
6:35
Ejaaz: You can go out there, create an eight second video and show us what you've got
6:39
Ejaaz: if you actually listen to this and you have an account, like we would love to see it.
6:42
Ejaaz: I will say that these things do look quite janky right now, as you mentioned,
6:49
Ejaaz: Josh, but this is also V1.
6:51
Ejaaz: And it's important to stress that world models up until now weren't really applicable
6:56
Ejaaz: or distributed to a mass audience.
6:59
Ejaaz: If you remember, Google released, I think it was Genie 3, which is like their
7:03
Ejaaz: really powerful world model.
7:05
Ejaaz: That is what powers this model that you're seeing on your screen here today.
7:09
Ejaaz: But it's not a full implementation. Why?
7:12
Ejaaz: Because it's actually quite expensive to serve.
7:14
Ejaaz: It requires a different type of infrastructure to train.
7:17
Ejaaz: And so once that kind of scales up with Google, they'll be the first distributors of this.
7:23
Ejaaz: And I'm excited to see more of this because Demis at Davos earlier this year
7:28
Ejaaz: basically said that he thinks world models is the future evolution of LLMs. He thinks that words are.
7:34
Ejaaz: Ultimately constraining for AI models. And in the future, if you have a model
7:38
Ejaaz: that can speak, but also understand how the world works, that will be the all-consuming model.
7:43
Ejaaz: Now, of course, that's in the future.
7:46
Ejaaz: For now, we have to focus on the major frontrunners of AI, which is Anthropic
7:51
Ejaaz: and OpenAI, which have the world's leading LLM.
7:55
Ejaaz: So the question becomes, did Google release an amazing LLM.
7:59
Ejaaz: Although they did release an LLM, it's called Gemini 3.5 Flash.
8:03
Ejaaz: And to quote Logan over here, it's their most powerful model to date.
8:07
Ejaaz: Now, it pushes the frontier in many different ways. We've got a chart here which
8:11
Ejaaz: kind of like compares Gemini 3.5 Flash to the older Gemini Flash models,
8:15
Ejaaz: as well as Opus 4.7 from Anthropic and GPT 5.5 from OpenAI.
8:21
Ejaaz: And you'll notice that Across coding in general, it doesn't trump or beat GPT
8:27
Ejaaz: 5.5 or Opus 4.7, which I guess some might say is a little disappointing, but also unsurprising.
8:34
Ejaaz: Google has fallen behind in coding AI for a while. So it's not as good,
8:38
Ejaaz: but it's kind of near the marker of them.
8:41
Ejaaz: But if you look just below here, this is where this bottle actually excels.
8:45
Ejaaz: It's known as agentic tooling or agentic AI in general.
8:48
Ejaaz: What Google has been able to achieve with 3.5 Flash is the ability to spin up
8:52
Ejaaz: multiple sub-agents that work in parallel to get your prompt resolved.
8:58
Ejaaz: So if you think that like having one singular brain to do all the work for you
9:03
Ejaaz: really, really well, which is what Claude Opus 4.7 does, which is what GPT 5.5
9:07
Ejaaz: does, Google took a different approach and said, let me have multiple brains,
9:11
Ejaaz: maybe even hundreds of brains working on the same problem and collectively,
9:14
Ejaaz: they can produce a better answer.
9:15
Ejaaz: And that's what they've achieved with the Gentic tooling here where you've got
9:19
Ejaaz: the tool-a-thon and the MCPP Atlas benchmarks, which absolutely crushed GPT 5.5 and called Opus 4.7.
9:26
Ejaaz: So you're probably wondering, okay, well, what tasks can I use this for?
9:28
Ejaaz: It's typically long horizon tasks, like long tasks that you can kind of shut
9:32
Ejaaz: your laptop down and kind of let it work away for like 10 to 20 hours this is the model for you.
9:37
Josh: So you'll note that they said that it is their best model
9:41
Josh: ever not the best model ever and that
9:44
Josh: it is fast but it is not the fastest and that
9:47
Josh: it is affordable but it is not the most affordable and when
9:50
Josh: you're kind of observing the parameters on the Pareto Frontier the things
9:53
Josh: that we actually measure to evaluate a model it's not
9:56
Josh: really the best at anything and in fact
9:58
Josh: we have this kind of comparison that I saw with Cerebris which
10:01
Josh: I really enjoyed and Cerebris posts this kind of trolling them which
10:04
Josh: is Cerebris versus Gemini flash and it very
10:07
Josh: clearly shows where cerebris sits versus flash
10:10
Josh: you could see 3.5 flashes out pretty fast in terms
10:12
Josh: of the tokens per second output but when you're trying to
10:15
Josh: optimize for one of these things along the frontier clearly cerebris has the
10:19
Josh: speed thing completed so it's not the fastest by a large margin it's not the
10:23
Josh: smartest by a pretty large margin it sits kind of in the middle and i think
10:27
Josh: this is the most interesting part of this story today is that a lot of these
10:31
Josh: models are not the best but they are unique in their own way one of the cool demos that
10:35
Josh: I liked that they had mentioned during the presentation is that they actually
10:39
Josh: took this new Gemini 3.5 flash model to build an operating system from scratch
10:43
Josh: over the course of a series of hours.
10:46
Josh: I mean, that's pretty incredible. It says it took 12 hours, 93 parallel agents,
10:50
Josh: 15,000 plus model requests, and then $2.6 billion in tokens process.
10:55
Josh: So you are essentially able to build an entirely new operating system from scratch
10:59
Josh: for less than $1,000 in API credits. That is pretty phenomenal.
11:03
Ejaaz: That's actually super hard to do. Do you remember we, I think, who did it?
11:08
Ejaaz: I think it might've been a GPT 5.4. It was, it was GPT 5.4.
11:12
Ejaaz: And Sam posted this video of a team building an OS from scratch.
11:17
Ejaaz: And that took them, I believe, 36 hours, like over a day.
11:22
Ejaaz: Like in that time, so I think that was like two and a half months ago.
11:25
Ejaaz: You now have Gemini Flash 3.5, which isn't as good a model as GPT 5.5.
11:29
Ejaaz: Do this in a fraction of the time by creating this novel approach of spinning
11:35
Ejaaz: up almost 100 agents to do that.
11:38
Ejaaz: Josh, I had a comment on the speed thing here because you're right.
11:42
Ejaaz: Cerebrus that runs GPT 5.5 versus Google is not just a question of speed.
11:48
Ejaaz: It's a question of infrastructure.
11:50
Ejaaz: Now, I have a demo here which shows actually how quick Gemini Flash can be.
11:55
Ejaaz: And to be honest, I just want to start this from the start. Look at this.
11:59
Ejaaz: That's crazy my brain can't ingest that
12:02
Ejaaz: much information in real time and i think although
12:05
Ejaaz: it's not as fast as some of these other new chip architectures it's still good
12:09
Ejaaz: enough for the majority of humans for sure right the other thing i'll say is
12:14
Ejaaz: this is built off of google's latest gpu infrastructure known as the tpu v8
12:19
Ejaaz: specifically the 8i which stands for eight inference it's inferencing chip,
12:24
Ejaaz: which basically is specialized around
12:26
Ejaaz: giving you quicker responses to your prompts. Now, that being said...
12:31
Ejaaz: This is incredibly expensive. Demis Hassabis dropped some stats here where he
12:37
Ejaaz: kind of shows off the speed.
12:39
Ejaaz: He says it's 4x faster than any of the other Frontier models.
12:42
Ejaaz: And actually, if you use our IDE environment, which is basically Google's equivalent
12:46
Ejaaz: of cursor, it performs 12x faster.
12:49
Ejaaz: What he doesn't mention is it's four times more expensive than all the other Frontier models.
12:55
Ejaaz: So there's a trade-off here, right? I think you mentioned it before we started recording, Josh.
13:01
Ejaaz: There's a trade-off. If speed is something that you necessarily need,
13:04
Ejaaz: which you could very well want in financial services where you're like trading
13:07
Ejaaz: algos for whatever, speed becomes a necessity.
13:10
Ejaaz: You're willing to pay that premium because you end up making more money on the top end.
13:13
Ejaaz: That's completely fine. But for the majority of users where speed doesn't necessarily
13:17
Ejaaz: matter that much, but thought and intelligence matters more,
13:20
Ejaaz: you want to make sure that that cost per watt per intelligence or whatever that unit might be,
13:26
Ejaaz: is kind of like kept at a certain bracket. Like I watched an episode where they
13:32
Ejaaz: interviewed the CFO of Anthropic and he mentioned that these enterprises are
13:35
Ejaaz: spending like 100%, sorry,
13:38
Ejaaz: 500% more than the initial budget that they had at the start of the year by the end of the year.
13:42
Ejaaz: So the point is like people are spending a lot of money on this.
13:44
Ejaaz: And at some point we need to kind of like reduce budgets to an extent.
13:48
Ejaaz: So yeah, it's quite expensive.
13:49
Josh: All right, who's an OpenClaw fan who's listening? Or who has given up on OpenClaw?
13:53
Josh: Because if you are a current OpenClaw user, you're probably going to hate this.
13:55
Josh: But if you've never used it before, this is a pretty compelling product.
13:58
Josh: And it's called Gemini Spark.
14:00
Josh: This is the third major announcement that they made.
14:02
Josh: And Gemini Spark is fascinating because it is that 24-7 personal agent.
14:07
Josh: They're claiming that it helps you navigate your digital life,
14:09
Josh: taking actions on your behalf and under your direction.
14:12
Josh: Essentially, this is Google's answer to OpenClaw. This is their version of the 24-7 agent.
14:16
Josh: And the reason why I said you may hate this is because you are beholden to Gemini 3.5.
14:21
Josh: Now, one of the best things about OpenClaw or if you use Hermes is it's model
14:25
Josh: agnostic. you could choose models that you want to use, and it will go and tackle specific tasks.
14:29
Josh: In this case, on Gemini Spark, you're stuck with 3.5.
14:32
Josh: So is this good? Is this bad? We'll see, but I find it really interesting because
14:36
Josh: this just runs directly in your browser.
14:39
Josh: So you don't actually need to have a laptop open. You don't need to run this on a remote server.
14:43
Josh: All you need to do is log into your Google account, set up a Gemini Spark instance,
14:48
Josh: and then it will just run permanently in the cloud.
14:51
Josh: A lot of times you'll see there's memes going around to people who are walking
14:53
Josh: around with their laptops cracked open because they're running these agents
14:55
Josh: that are running long-term tasks.
14:57
Josh: This is basically your own little private server to go off and build anything that you want.
15:00
Josh: And it's pretty interesting. I think for people who are less technical or for
15:04
Josh: people who just don't love using OpenClaw, this is a pretty easy way of accessing
15:08
Josh: a 24-7 server to just run commands, to run prompts over very long durations of time.
15:13
Ejaaz: I like this and I hate this. The reason why I like this is I use a lot of Google
15:19
Ejaaz: products to do work, Google Docs, Sheets,
15:22
Ejaaz: Gmail. So the fact that I can have an OpenClaw that is seamlessly integrated
15:26
Ejaaz: into all of those things gives me hope.
15:29
Ejaaz: The other side of this is I just don't think it's going to be as good a quality
15:33
Ejaaz: product as OpenClaw or Clawed Cowork.
15:37
Ejaaz: In fact, there was news that broke, what was it, like a couple of weeks ago,
15:42
Ejaaz: that revealed that majority of Google employees were using Clawed to do their work, right?
15:48
Ejaaz: And then Google kind of like put a ban on this. And the reason why they were
15:51
Ejaaz: doing this was presumably to learn how to build a better agentic or coding model
15:56
Ejaaz: that can do exactly what we're seeing on the screen here.
15:58
Ejaaz: So they're late. They're lagging the front runners. I'm glad they put something
16:03
Ejaaz: out there. Google Desk has the distribution.
16:05
Ejaaz: They do have the data mode. They do have all the users.
16:08
Ejaaz: For now. And so I think they have a shot at creating something good here.
16:12
Ejaaz: But I get meta vibes from this, Josh, like the same reason why like meta might
16:16
Ejaaz: have like a good model with Muse Spark.
16:18
Ejaaz: I'm not using it because it's not on the platforms that I actually want to use the thing.
16:22
Josh: Yeah, and I think that's one of the trends of all the announcers that we're talking about today.
16:26
Josh: They are all kind of existing in seemingly different places.
16:30
Josh: It's not clear and obvious which features I'll want to use. We're
16:32
Josh: about to rattle off maybe five or six more things that they've announced
16:35
Josh: and i think one of the themes as i i mean this is
16:38
Josh: a common complaint that we've had with google or at least me personally is
16:41
Josh: i'm not really sure the best way to use all of these
16:44
Josh: tools like with anthropic i go to the claw desktop app
16:47
Josh: and it has everything in one roof and i know exactly how it all works with open
16:51
Josh: ai the same thing is true and i even have a companion mobile app with google
16:54
Josh: there's a lot of this kind of ambient ai that exists at the seemingly random
16:59
Josh: series of touch points like it exists in my gmail but i kind of hate the ai
17:03
Josh: written emails that it makes.
17:04
Josh: It exists in my Google Drive, but I don't really need it there.
17:07
Josh: And then you have the Gemini app, which is getting better, but it still doesn't
17:11
Josh: include a lot of the functionality that we're mentioning today.
17:13
Josh: I mean, the next thing we could talk about here is their search,
17:15
Josh: because for the first time in over 25 years, they're really changing how search is going to work.
17:21
Josh: And again, it's just kind of like this ambient AI that you don't really seek out.
17:25
Josh: I'm not really sure where to find all that a little confusing.
17:28
Josh: But in terms of the AI search, it's very different.
17:30
Josh: Now Google search has gemini 3.5 flash
17:33
Josh: baked into it by default so when
17:36
Josh: you search for something it will be routed through gemini 3.5
17:39
Josh: instead of the traditional google search engine and this is a pretty profound shift
17:42
Josh: in just the way the internet works the way people get information on the internet
17:46
Josh: the first person i think about is the sites that are kind of optimizing for
17:50
Josh: seo on google now it seems like the whole meta has shifted to seo for ai models
17:55
Josh: because now when you search something the reality is is that you're going to
17:58
Josh: be routed through an ai model and not the
18:01
Josh: indexing algorithm, at least for now.
18:03
Ejaaz: So one thing a lot of critics on Wall Street said about Google three months
18:07
Ejaaz: ago was that they're worried about AI eating their dominant market share in search.
18:13
Ejaaz: And then they released their Q1 earnings about a week ago, and it revealed that
18:18
Ejaaz: not only did it not eat their search market share, it increased it.
18:22
Ejaaz: They ended up earning more money. And I think that's because,
18:25
Ejaaz: Google Search has been integrating AI for a while now. This isn't the first
18:29
Ejaaz: time that I type something in Google and it gives me an AI-generated prompt.
18:32
Ejaaz: I love that we have the latest model here and that they're making it more native into the homepage.
18:37
Ejaaz: I think this is amazing. So this is my second favorite release from Google, from I.O.
18:42
Ejaaz: Hundreds of millions of people, probably even billions of people,
18:45
Ejaaz: use Google every single day. They ask a bunch of different random questions.
18:49
Ejaaz: I would actually argue that it has more reach than your ChatGPT home interface
18:53
Ejaaz: or Claude home interface.
18:55
Ejaaz: And if Google is able to successfully pivot this into something that serves
18:59
Ejaaz: up their latest LLM prompts and somehow get people to use or pay for it,
19:04
Ejaaz: then this could end up being a home run.
19:05
Ejaaz: Of course, as you mentioned earlier, like Google is like free at this point,
19:09
Ejaaz: right? The whole point is like distribution is the mode.
19:11
Ejaaz: And so I think they're going to end up making most of their money from this
19:13
Ejaaz: particular feature through ads.
19:16
Ejaaz: Now, it's important to say that no AI company right now has figured out the
19:21
Ejaaz: SEO strategy. In fact, the leader of this is Meta.
19:25
Ejaaz: Who has been embedding a bunch of AI agents and models into their ad services
19:29
Ejaaz: on the backend to help advertisers reach more of a wider audience.
19:33
Ejaaz: If anyone knows the data of their users, the best is definitely Meta.
19:37
Ejaaz: They've been training on this for a while.
19:39
Ejaaz: Their new model, MuseSpark, does exactly this. I have a feeling Google's going
19:42
Ejaaz: to go down the same route, but they're going to be a little more careful about
19:44
Ejaaz: this because they are the dominant market platform for advertising.
19:47
Ejaaz: So I'm excited to see where this goes from here.
19:50
Ejaaz: And then there was this other post, which you were commenting on,
19:53
Ejaaz: Josh, which is basically like, you get like such good value for money with a google ai pro bundle.
20:00
Josh: This is my new favorite subscription for twenty dollars it comes
20:03
Josh: with youtube premium light which doesn't include music videos but basically
20:07
Josh: an ad-free experience on youtube in addition to five terabytes of storage you
20:10
Josh: get access to all the new gemini pro models it has image video generation access
20:15
Josh: anti-gravity notebook lm pro which is really good for deep research and long
20:19
Josh: context stuff It's a really compelling bundle at $20.
20:22
Josh: And I think this is one of the most understated
20:25
Josh: things that was released yesterday,
20:27
Josh: because this is what I would probably recommend most people use.
20:29
Josh: It gives you access to the entire suite. And even if you just watch YouTube
20:33
Josh: videos, I mean, you get access to the ad-free version of that.
20:36
Josh: And that alone makes it worth it with the five terabytes of storage.
20:39
Josh: So very compelling value coming out of them.
20:41
Josh: And speaking of YouTube, they also did the same search thing that they did with
20:44
Josh: Google on YouTube as well, where now there's AI search baked right into your search results.
20:48
Josh: It turns out that Gemini has been trained on lots of youtube
20:51
Josh: data and actually has context on the types of
20:54
Josh: videos that you want to watch so
20:57
Josh: if you ask a question about a specific thing that you
20:59
Josh: want troubleshot it will not only find the right videos for
21:02
Josh: you but it will find the correct moments in the videos and
21:05
Josh: serve those up to you and i find that we've started to experience this in google
21:09
Josh: search recently where like i can type something it'll kind of time stamp me
21:12
Josh: into a video now it's fully baked into the youtube experience and i think that's
21:16
Josh: going to be a really compelling product as well so google in terms of upgrading
21:19
Josh: their core products that everyone uses on a day-to-day basis looking really strong.
21:25
Ejaaz: For this final section, Josh, I want us to zoom out because the question that
21:28
Ejaaz: I have on my mind is Google has a decision to make on which strategy they want to pursue for AI.
21:37
Ejaaz: Right now, OpenAI has made it very clear they want to catch up with Anthropic
21:40
Ejaaz: on coding. And in fact, you could argue that they have done that.
21:43
Ejaaz: Anthropic focused on coding for the longest time. And the reason why both of
21:46
Ejaaz: those companies are doing that is they believe if they nail or coding AI,
21:50
Ejaaz: then they can build everything else.
21:52
Ejaaz: Everything else becomes downstream of coding AI, right? There's an argument there.
21:55
Ejaaz: Now, Google, with this latest launch or series of launches.
21:59
Ejaaz: Hasn't really made that explicit statement as well. In fact,
22:02
Ejaaz: I think they're making a few other statements.
22:04
Ejaaz: Number one, they're betting that video or Omni models are going to be the future.
22:10
Ejaaz: That's why they spent a lot of compute and time invested in their new Gemini
22:15
Ejaaz: Omni intelligence model.
22:17
Ejaaz: Number two, I think that Google is realizing that the agent harness or the harness
22:23
Ejaaz: that goes around And an AI model is equally as important as the fundamental
22:27
Ejaaz: foundation model itself. Now, I have proof of this.
22:31
Ejaaz: There was this tweet that I saw yesterday, which on Cursor, you can measure
22:35
Ejaaz: the intelligence of different models across different tasks.
22:38
Ejaaz: And it showed that the recent Gemini 3.5 flash was worse than Cursor's own foundational
22:46
Ejaaz: model, the old version of it, Composer 2.0.
22:49
Ejaaz: And yesterday, Cursor released 2.5, which is on par, you can see on the top
22:54
Ejaaz: here, with the other models, the Frontier models.
22:57
Ejaaz: So the question I have is, like, Gemini kind of missed here.
23:00
Ejaaz: Like, they should be focusing on the agentic side. As I mentioned earlier,
23:03
Ejaaz: 3.5 Flash is great at agentic tooling, but what they've missed here is the harness.
23:08
Ejaaz: The harness is basically the wrapper, the muscle memory that goes around your
23:11
Ejaaz: model that orchestrates the model in a really ideal way.
23:14
Ejaaz: What OpenAI, what Anthropica realized is that that is super important.
23:17
Ejaaz: And Elon Musk has figured that out, and that's why SpaceX is looking to acquire
23:21
Ejaaz: or moving to acquire Cursor for $60 billion.
23:23
Ejaaz: I think Google missed here, and what I come away feeling about the series of I.O.
23:28
Ejaaz: Launches in general is, it's good, but it's not great enough,
23:31
Ejaaz: and if you're Google, which was.
23:33
Ejaaz: Two weeks ago, the most valuable company in the entire world,
23:36
Ejaaz: you need to be making bigger bets. And it's a shame that they haven't made it.
23:40
Ejaaz: That's that's currently how I feel.
23:41
Josh: Yeah, when you compare the prices to I mean, it's four times the cost of composer.
23:45
Josh: So it's like on a price basis is losing on a score basis is losing.
23:49
Josh: I will say this is 3.5 flesh, not 3.5 pro, which we can expect to be released next month.
23:55
Josh: So that might change things and shake everything up in a pretty big way.
23:59
Josh: It's just not currently available. So we can't judge it we can't be
24:01
Josh: critical of it there is one final announcement i just want to
24:04
Josh: squeeze in because as a hardware guy i get excited about anytime
24:07
Josh: someone releases hardware google has some new hardware offerings it
24:10
Josh: comes in the form of glasses they are able to see the world and interpret it
24:15
Josh: some of the examples that were used that i found pretty cool was um she could
24:19
Josh: take a selfie the person who was giving a demo of the crowd and then prompt
24:23
Josh: the glasses to not only capture the image but create a variation of it using Nano Banana Pro.
24:28
Josh: It gives directions. It is agentic in the sense that it can work with your phone,
24:32
Josh: order you coffee on the way to work.
24:33
Josh: It's pretty cool, pretty interesting. I'm very excited to see more about glasses.
24:38
Josh: They actually announced that they're planning to add a HUD, a heads-up display,
24:42
Josh: basically a visual display into the glasses as soon as next year, which is exciting.
24:46
Josh: So Google seems to be stepping pretty seriously into the hardware market.
24:49
Josh: I think the current versions are kind of a joke, similar to all the others,
24:54
Josh: but the ones with the heads-up display are going to be very exciting next year
24:57
Josh: very much could be the year of ai glasses but with that
25:01
Josh: I think we got it covered. Is that IO covered? Is that all of it?
25:05
Ejaaz: I think that's just about everything. There's a bunch of other stuff,
25:08
Ejaaz: but it's kind of tertiary.
25:09
Ejaaz: There's some YouTube upgrades and stuff like that. But I think we've covered
25:13
Ejaaz: all the hot topics, Josh.
25:15
Ejaaz: I think I come away from this episode thinking Google is not out of the race,
25:19
Ejaaz: but I do think that they need to double down on coding specifically.
25:23
Ejaaz: And the crazy part is like they have the moat. They're currently in the number one position.
25:27
Ejaaz: The only other company that could compete with them is Apple.
25:30
Ejaaz: And Apple has been doing nothing.
25:32
Ejaaz: They're getting a CEO swap. And then I think later this year with John Turner's
25:36
Ejaaz: coming in and then I think we're going to see like them step up.
25:38
Ejaaz: But the truth is, it's them versus these two startups, which.
25:42
Ejaaz: Are aggressively increasing ARR by the month.
25:47
Ejaaz: I don't know if you saw this, Josh, but someone projected out Anthropik's rate
25:52
Ejaaz: of revenue increase, which means that by 2028, they would have surpassed Google.
25:59
Ejaaz: So Anthropik's projection for this year was 10 billion ARR.
26:03
Ejaaz: They are about to hit 45 billion ARR this month.
26:08
Ejaaz: And if that projection or if that rate of increase keeps happening and they
26:12
Ejaaz: start consuming more enterprise customers, etc, etc,
26:14
Ejaaz: you can feasibly say that they will become as large as Google,
26:18
Ejaaz: which then starts to justify all these crazy, you know, $1.1 trillion valuations
26:23
Ejaaz: or the secondary round that they're raising right now for $1 trillion.
26:27
Ejaaz: It kind of makes sense. So I think Google still has a shot.
26:30
Ejaaz: This wasn't the killing blow, but maybe next time, maybe when Flash,
26:34
Ejaaz: not Flash Pro, but 3.5 Pro comes out.
26:37
Josh: Yeah, well, if I could give any feedback on the presentation,
26:39
Josh: it would be to focus. It feels like Google doesn't really have any focus there.
26:44
Josh: It felt like I was watching 15 separate presentations instead of one coherent one.
26:48
Josh: And therefore, as someone who is an AI fan who literally follows this for a
26:52
Josh: full-time job, I was unable to keep up and orient myself on where all these
26:56
Josh: products fit into the ecosystem.
26:57
Josh: So if I could leave one criticism for Google, it's just to focus.
27:01
Josh: When I think about Anthropics offerings, I have a very clear picture in my head
27:04
Josh: of what they're good at and how I want to use it.
27:07
Josh: Same is true with OpenAI. With Google, that's not the case and i
27:09
Josh: think if they can focus if they can compress all these
27:12
Josh: releases down into a few key things that would meaningfully shift
27:16
Josh: the way that people think about gemini and engage with the products and i hope
27:19
Josh: that for their next releases their next presentations will get a little bit
27:22
Josh: more of that focus but that is the episode for today that is google io you are
27:26
Josh: now fully up to date and informed on all the things gemini all things google
27:31
Josh: any final parting thoughts you just before we head out of here
27:34
Ejaaz: Please spin out a really good version of Google Glass. That's one.
27:38
Josh: Oh, man, I love that.
27:40
Ejaaz: I really want a good hardware device. I'm like begging for it.
27:44
Ejaaz: Josh and I saw another leak last week. We didn't mention this on last week's
27:48
Ejaaz: episode, but it was those weird OpenAI over earbuds with the weird disk thing.
27:54
Ejaaz: And I was like, I need that device. I just need something.
27:57
Ejaaz: I realized I need hardware in an ever-increasing digital world.
28:01
Ejaaz: So Google Glass, I almost think it's good if they would just spin it out and
28:05
Ejaaz: make it a cool consumer company. But yeah.
28:08
Josh: Well, you know, no one wants hardware more than I do. I'm dying for a good pair
28:11
Josh: of glasses, just anything physical that I could touch.
28:14
Josh: That's not my iPhone. That's not made by Apple.
28:16
Josh: But alas, no one has completed that yet. So we will keep our eyes peeled.
28:20
Josh: We will continue to watch. If you enjoyed this video, please don't forget to
28:23
Josh: share it with a friend who might also have enjoyed it.
28:25
Ejaaz: Don't forget to subscribe to the channel. Loads of new subscribers.
28:27
Josh: Josh. Leave a comment. Yeah, we had our biggest episode ever.
28:30
Josh: Drop on Monday, the Leopold Ashenbrenner portfolio, the new 13f filing so if
28:35
Josh: you haven't seen that i would highly advise we had a lot of new people joining
28:39
Josh: so for the new people welcome we publish new episodes four times a week about
28:42
Josh: 20 minutes sometimes we go a little bit longer when we get excited but there
28:46
Josh: will be many more to come this week so thank you guys so much for watching this one and we'll see
28:49
Ejaaz: You next time see you guys.