Who Won the Week: Google or OpenAI?

0:00Today on the AI Daily Brief, everything announced at Google I.O and a look at the battle between Google and OpenAI. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, hop on the Discord with a link in our show notes.

0:24Hello, friends. One more day of non-traditional formats. We're just going to be talking about Google I.O. and the comparison to OpenAI, as that's where a ton of conversation is right now. The other big obvious story that we will cover is Ilya formally leaving OpenAI, so keep your ears peeled for that in the next day or so. But for now, let's talk about Google I.O. Welcome back to the AI Daily Brief. Yesterday was the main keynote day of Google I.O. This is Google's big annual conference where they announce all the stuff that they're working on, and at the surprise of no one, AI was even more center stage than it was last year.

0:56Now, the timing of this event was no coincidence, or more specifically, the timing of OpenAI's event just before this was no coincidence. It was pretty clear then and is definitely clear now that OpenAI was trying to front-run Google and capture excitement and attention before I.O. happened. What we're going to do today is talk about the difference between these two announcements, who the AI Twittersphere thinks won, but we have to start with just everything that Google actually announced. This is in no particular order, other than broadly speaking the biggest, most discussed things being up top.

1:25First of all, we got a new version of Gemini. This is called Gemini 1.5 Flash. The team at Google said that this multimodal model was as powerful as Gemini 1.5 Pro, but that it had been optimized for quote, narrow, high-frequency, low-latency tasks, making it better at generating fast responses. They also announced that the context window was moving from the already industry-leading 1 million up to 2 million in the coming months. Second, and perhaps most relevant for the demos that we saw a couple days ago from OpenAI, we got Project Astra. This is their version of a personal assistant. On screen is the demo video they shared both live and on socials.

1:59Which makes sound. What is that part of the speaker called? That is the tweeter. It produces high frequency sounds. The Verge writes, Project Astra is powering many of the most impressive demos from IO this year, and the company's aim for it is to be an honest-to-goodness AI agent that can't just talk to you, but also actually does things on your behalf. We'll come back to that in a moment, given that it is such a strong and clear through line across both the OpenAI presentation and this Google presentation. that these companies are betting on a totally new mode of interaction between people and AI that they're literally betting their whole enterprise on.

2:33We also got Google's answer to Sora. A couple months ago, there had been a question of whether OpenAI was starting to lose the lead that it had held for so long, but then they dropped a set of demos for Sora, and honestly, everyone was just slack-jawed. It was so far ahead of what was available currently from other text-to-video generators that it honestly shifted many people's opinions on whether specialized models can ever beat a big generalist model. Then again, uncommonly for OpenAI, Sora hasn't been available for people to test. Part of that seems to be economics, with generations from Sora just costing an absolute boatload to create right now.

3:05Instead, OpenAI has been pursuing relationships with Hollywood and professional filmmakers, but with Vio, Google has produced an answer. Like OpenAI's Sora, Vio is not available to the public, and instead, Google has been working with a set of high-profile YouTube creators as well as Hollywood filmmakers to try to start integrating the technology. IO also saw the announcement of a competitor to OpenAI's custom GPTs, something that Google is calling GEMS. They are basically pitching it in exactly the same way that those custom GPTs were created. Their announcement tweet reads, Whether you need a yoga bestie or a calculus tutor, in the coming months you'll be able to customize Gemini, saving time when you have specific ways you interact with Gemini again and again.

3:42Gemini, like GPT-4, got a conversational upgrade for voice, adding personality, the ability to interrupt it, and lower latency to make conversations hopefully feel more natural. And then from there, there were a ton of announcements of where Gemini was being integrated far more deeply into Google products than it had been before. For example, Gemini is getting much more deeply integrated into Workspace. The Verge again writes, Google is rolling Gemini 1.5 Pro into the sidebar for Docs, Sheets, Slides, Drive, and Gmail. When it rolls out to paid subscribers next month, it will turn into more of a general purpose assistant within Workspace that can fetch info from any and all of the content from your drive no matter where you are.

4:16It will also be able to do things for you like write emails that incorporate info from a document you're currently looking at or remind you later to respond to an email you're perusing. Interestingly, this integration is starting to get so deep that it sort of appears to me that Google is making a bet on a future where our default interaction with computers is talking to our AI assistant that then goes and does the stuff that we do now for us. So instead of writing an email, you talk to your assistant, which writes the email. Instead of analyzing a document, you ask your AI assistant to analyze a document, et cetera, et cetera.

4:44This one is coming pretty soon, so we'll see just how ubiquitous it actually becomes and whether people really start to shift those behaviors. Another big important integration is around search. Thank goodness we're losing the language of search generative experiences, which is an incredible mouthful, and instead they're being replaced with what they're calling AI overviews. Basically every search now is going to have a perplexity style summarization first rather than just the classic blue links. It is not hyperbolic to call this the most significant shift to Google search since the product was launched 20 plus years ago.

5:16Relatedly, Google Chrome is getting an AI assistant which will basically place text generation natively in the browser. As a listener of this show, I have a strong feeling you like to stay up to date on all things artificial intelligence, including its impact on the workforce, which is why I highly recommend checking out Managing the Future of Work, the chart-topping business podcast from Harvard Business School. HBS professors Bill Kerr and Joe Fuller talk to business leaders, technologists, and policymakers grappling with the forces like AI, globalization, and demographic shifts that are reshaping the nature of work.

5:46Recent guests include IBM CHRO Nicol Lamoureux on how Big Blue is adopting AI, Morningstar CEO Kunal Kapoor on how AI can raise the investment IQ, Microsoft Corporate Vice President Jared Spatero on how the tech giant is experimenting its way from AI assistants to autonomous agents, and many other prominent movers in business and the workforce ecosystem. So don't miss out. Follow Managing the Future of Work on Apple Podcasts, Spotify, or wherever you're listening now. Hello, friends. Quick note before we get back to the show, I'm so excited to share that Superintelligent is now live. Superintelligent is a platform for fast, fun, and super practical, useful AI learning.

6:25We have something like 300 video tutorials adding 30 to 50 each week, covering every topic in AI you can imagine from LLMs to image generators to case studies, use cases, basically everything that tells you how to use AI and what to use it on in short, fast, four to seven-minute tutorial videos, which are paired with step-by-step instructions that help you actually use these tools as well. It's$20 a month for unlimited access, and I would love to see you there. Check it out at besuper.ai. That's besuper.ai. So let's talk about impressions. First of all, I think Jan Pelek nailed what a lot of people were feeling when he wrote, Google is out for blood.

7:05What a bombardment. In many ways, that's what it felt like to me too, As opposed to OpenAI's tight 30-minute announcement, Google's I.O. keynote went on for a couple of hours. In fact, in its hugeness, it sort of gave the impression that Google wasn't trying to compete one-to-one between Project Astra, their AI assistant, and OpenAI's GPT-4.0 assistant, but instead to show just how many different vectors of AI Google is competing on. Still, there was a lot of focus on comparing Project Astra and GPT-4.0. Professor Ethan Malek writes, This video does a great job highlighting the practical side of ubiquitous multimodal AI agents.

7:38It's fascinating to see the neck-and-neck technology race between Google and OpenAI around the same set of problems and approaches. Robert Lukosko writes, Guys, that's really insane. I rewatched this IO agent's demo three times already. Tell me Google will not win agent space. It can schedule UPS pickup from a single photo. However, the biggest critique was that we're still comparing what was a slightly janky live demo in the case of OpenAI from extremely high-production demo videos prepared for Google. Amir from Duist writes, There's a big difference between OpenAI and Google. Open AI ships useful stuff.

8:09For example, the GPT-40 model is incredibly fast and 50 % cheaper, and the voice improvement is incredible. It's widely available. Google announces stuff that looks great in I-O presentations, but isn't really usable or even released in many cases. It's mostly vaporware. It's fantastic to see a company with real potential to disrupt Google. DC Investor writes, Google simply lacks credibility for any kind of AI demo at this point, like Project Astra or similar. Promo Reel's even worse. Meanwhile, Open AI does a semi-scripted event for GPT-4-0, but enough flaws come through to show it's real. Google needs a full culture reset at this point.

8:40What he's referring to, in part, is that people got extremely hyped-backed last December when Google announced Gemini Ultra, only to realize that the videos that had been most impressive were really pretty doctored to show exactly what they wanted to show versus where the actual capacity was. Some people also pointed out that even these overly prepared demos weren't as impressive as what OpenAI showed. Bindu Reddy, who was a little bit harsh on OpenAI, frankly, wrote, after watching Google I.O., it's safe to say what OpenAI showed yesterday was mind-blowing. Astra is a prototype voice assistant and seemed like a two-year-old baby compared to OAI's Scarlett Johansson.

9:12Danny Rodriguez said, well, that's quite a turnaround from yesterday. To which Bindu replied, it definitely is, lol. Google I.O. gave me a comparison, and comparatively, the OpenAI version was very, very good. Unfortunately, the people who were at I.O. and got their hands on Project Astra also weren't all that complimentary. Santiago on Twitter said, I tried Project Astra and the demo wasn't great. One of the things that was totally unignorable was the difference in approach. Google's I.O. presentation was, like I said, extremely high production value, and it also had celebrity integrations woven throughout.

9:44OpenAI's presentation was intimate and quaint by comparison. However, I'm not so sure that does exactly what Google wanted. I asked my Twitter followers, did Google's high production values and celebrity integrations make you more or less stoked on their AI announcements? Notably, 42.6 % of people said it didn't impact their opinion, but of those who did have it affect their opinion, about 90 % were less stoked rather than more stoked. Broadly speaking, other polls I saw confirmed a pretty similar sensibility about the two events. Daedalus asked, did Google cook OpenAI? 75.9 % of respondents said nah.

10:15Stanford's Andrew Gao wrote, who wins AI today? Among nearly 1 ,600 voters, 60.4 % said OpenAI, compared to just 16.1 % who said Google. Jenny AI, who does research at Microsoft wrote So Who won the week, and of her 400 respondents, 87.5 % said OpenAI. It is worth noting, however, that there may be an audience bias issue here. OpenAI is very much competing for the state of the art. They are, as they have been, leading the pack from a technology perspective. A significant portion of the people who are discussing this on AI Twitter are themselves technologists, developers, and entrepreneurs. They are the type of people, in other words, who value state-of-the-art more than anything else.

10:54That's different, I think, than the audience that Google is going for and the strategy that Google is taking. Google is taking an approach of putting AI everywhere with their incredibly large and diverse install base. OpenAI does not have users that they are bringing to their party. Google has everyone using Gmail, everyone using Maps, everyone using Search, everyone using Docs, everyone using Slides, etc., etc., etc. While they are clearly not interested in seeding state-of-the-art in any way, the strategy that they're actually deploying doesn't technically rely on being state-of-the-art. It requires being good enough for these products to actually be useful for the people who are already their users.

11:30It's frankly a much lower burden, and one of the reasons that even for an open AI, it's so incredibly hard to compete in this space. So to the extent that you are trying to find signal in these results, I don't think it's about who is necessarily leading AI in the real world, but it does seem to be clear that people think that what open AI is producing is still a little bit at least ahead of what Google's labs are producing. For some, though, they're not ready to sign off on any of these yet. Robert Scoble writes, I'm in a judgy mood, so here's some things I'm going to do with my AI personal assistant, since that is what OpenAI and Google are trying to bring to us.

12:01Scoble then puts together a list of 30 questions by which he will judge an AI assistant. For example, he writes, does it do the basics? Like when you ask Siri to take you back to your car, it doesn't work. But if you ask it to take you back to your parked car, it does. If you ask it a thousand questions, how many does it answer accurately? In my perfect world that would have no mistakes, etc, etc, etc. And I think, hold aside Scoble's specifics, what he's getting at is that ultimately, AI assistants will only be a thing if they're good enough for them to be a thing. The general mass market of consumers isn't going to use something because it's AI and cool.

12:32They're going to use something strictly on the basis of whether it's useful. We are still at the very early stages of these products coming to market, and it's not at all clear to me yet that either Project Astra or the new GPT-40 is actually going to clear that threshold. Still, after all of this, the competitive landscape is getting a little bit clearer. DC Investor again pointed out OpenAI's stakes and why the news about them potentially getting a deal with Apple is even more interesting today than it was a couple of days ago. He writes, The big thing I see with Google's AI announcement yesterday is they are driving for much deeper integration with your data and your surroundings.

13:04OpenAI better hope Apple can do the same for them. I could see an argument that OpenAI needs Apple as much as Apple needs OpenAI. At the same time, there are clearly some forces that have been unleashed by AI that all of these companies now feel subject to. Daryl Bossonjo writes, The real problem OpenAI has created for Google is that it has damaged the virtuous cycle of websites making free content for Google to index so Google sends them traffic in return. Now Google's AI results will provide answers instead of sending sites traffic and the web withers. Pete Paschal put it more simply, It's official.

13:34Generative search is now the norm. Before the OpenAI event, we had so many rumors that they were going to announce a search product, which they of course didn't, but I would be very surprised if something like it wasn't still in the works. Still, maybe the tweet of the day went to Andre Karpathy, formerly of OpenAI, who in reflecting on the new OpenAI voice assistant, which reminded so many people of Samantha from her, wrote, the killer app of LLMs is Scarlett Johansson. You all thought it was math or something. AI is more competitive than ever, and next week we're getting a Microsoft event as well.

14:04So stay tuned for all of this, as there is going to be a lot to discuss. If you made it this far and want to dig deeper into how to actually use these tools, please go check out Super Intelligent. It's our platform for fast, fun, and useful AI learning. You can find it at besuper.ai. For now, though, that's going to do it for the AI Daily Brief. Until next time, peace.

14:34ت