At Google's I/O developer conference, the company made the case to developers, and to some extent consumers, why its bet on AI is better than its competitors. At the event, the company announced a revamped AI-powered search engine, an expanded AI model with a 2 million token context window, AI helpers across the Workspace suite of apps including Gmail, Drive, Docs, and tools to integrate AI. Did. It even includes developer apps and a vision of the future of AI (codenamed Project Astra) that can respond to a combination of sight, sound, voice, and text.
While each advancement was promising, the onslaught of AI news was overwhelming. Although clearly aimed at developers, these large-scale events are also an opportunity to wow end users about the technology. But after a deluge of news, even the most tech-savvy consumer may be asking themselves, “Wait, what is Astra again?” Is that what powers Gemini Live? Is Gemini Live like Google Lens? How is it different from Gemini Flash? Is Google actually making AI glasses, or is it just Vapor? What is Gemma? What is LearnLM? What are Gems? When will Gemini arrive in your inbox or documents? How can I use them? Or?
If you know these answers, congratulations! You're a TechCrunch reader. (If not, click the link to stay updated.)
Image credit: Google
Despite the enthusiasm of the individual presenters and the loud cheers from the Googlers in the crowd, what was missing from the entire presentation was a sense of the coming AI revolution. If AI ultimately leads to products that have a major impact on the direction of technology, like the iPhone influenced personal computing, this is not the event where AI debuted.
Rather, it turns out that we are still in the early stages of AI development.
On the sidelines of the event, there was a sense that even Googlers knew their work was unfinished. When we demonstrated how the AI could upload hundreds of pages of documents and instantly create study guides and quizzes for students, it was an impressive feat, but quiz answers were not annotated with citations. I noticed that. Asked about accuracy, employees acknowledged that the AI interprets things mostly accurately and that future versions will point to sources and allow people to fact-check answers. But if you need to check the facts, how reliable can he trust AI study guides in preparing for tests in the first place?
Astra's demo uses a camera mounted on a table and connected to a large touchscreen to let the AI play Pictionary, show you objects, ask questions about those objects, and tell a story. can. However, despite impressive technological advances in their own right, the use cases for how these capabilities would be applied to everyday life were not immediately apparent.
For example, you can ask the AI to use alliteration to describe objects. During the livestreamed keynote, Astra looked at a set of crayons and responded, “Brightly colored creative crayons.” A nice party trick.
In a private demo, we challenged Astra to guess objects in a scribbled picture, and it quickly and accurately identified a flower and a house that I drew on the touchscreen. When I drew the bug, I drew a large circle for the body, a small circle for the head, and small legs from the sides of the large circle. The AI stumbled. Is it a flower? No. Is it the sun? No, the employee told the AI to guess what is alive. I added two legs for a total of eight. Is it a spider? yes. Despite my lack of artistic ability, a human would have noticed the bug right away.
in #GoogleIOtook a private test for Project Astra (not sure if I'm allowed to film this?). @skilano Talk to Astra, identifying what she sees. Fully multimodal, no lag, very smooth. Insane. pic.twitter.com/ayWPZRiEYq
— Conor Grennan (@conorgrennan) May 15, 2024
No, I shouldn't have recorded it. But here's a similar demo posted on X.
To help you understand the current state of technology, Google staff did not allow recording or photography in the Astra demo room. I could run Astra on my Android smartphone, but I couldn't see the app or hold the phone. While the demo was fun and the technology that made it possible is certainly worth exploring, Google missed an opportunity to showcase how its AI technology will impact everyday life.
For example, when should you ask an AI to come up with a band name based on images of your dog and stuffed tiger? Do you really need an AI to find your glasses? Here's another Astra demo from the talk.
This isn't the first time we've seen a technology event filled with demos of advanced futures with no real-world applications or, more importantly, touting convenience as an upgrade. Google, for example, has teased its own AR glasses in the past. (It even had a skydiver wearing Google Glass parachute into I/O. The project was built more than a decade ago, but has since been canceled.)
Looking at I/O, it feels like Google sees AI as a way to generate additional revenue. If you need a product upgrade, pay for Google One AI Premium. Google probably won't make the first big breakthrough in consumer AI. As his CEO of OpenAI, Sam Altman, recently reflected, OpenAI's original idea was to develop technology to “create all sorts of benefits for the world.”
“Instead, it's likely that we'll develop AI and others will use it to create all sorts of cool things that we all benefit from,” he said.
Google appears to be in the same position.
Still, there was a time when Google's Astra AI looked more promising. It's easy to see how it can be a useful work partner if it can correctly identify the code or suggest ways to improve the system based on the diagram. (Clippy has evolved!)
Gemini for Gmail. Image credit: Google
There were other moments where AI's real-world utility shined. For example, Google Photos has a great search tool. Plus, we bring Gemini's AI to your inbox to help you summarize emails, draft replies, and list action items, ultimately making Inbox 0, or near it, even better. You may be able to get there sooner. But delete emails that are annoying but not spam, organize your emails smartly into labels, never miss an important message, and get an overview of everything in your inbox that you need to take action on as soon as you log in. Can you provide that? ? Can you summarize the most important news in the newsletter? Not completely. still.
Additionally, some of the more complex features, such as AI-powered workflows and receipt management, which were demonstrated, won't roll out to Labs until September.
When considering how AI will impact the Android ecosystem (Google's pitch to developers in attendance), even Google is still arguing whether AI will help Android pull users away from Apple's ecosystem. I had a feeling that I couldn't do it. “When is the best time to switch from iPhone to Android?” we asked Googlers of all ranks. “This fall” was the common response. In other words, Google's fall hardware event should coincide with Apple's adoption of RCS, an upgrade to SMS that makes Android Messaging more competitive with iMessage.
Simply put, bringing AI to consumers' personal computing devices may require the development of new hardware. AR glasses? A smarter smartwatch? Gemini-powered Pixel Buds? — but Google isn't ready to release or tease any hardware updates yet. And as we've already seen, hardware remains a challenge with the underwhelming launches of Ai Pin and Rabbit.
Image credit: Google
There's a lot you can do now with Google's AI technology on Android devices, but Google's accessories like the Pixel Watch and the system that powers them, WearOS, are getting some small performance improvements. Apart from that, it was largely ignored in I/O. Its Pixel Buds earbuds didn't even receive loud praise. In Apple's world, these accessories help lock users into the company's ecosystem, and may one day connect them to AI-powered Siri. These are important elements of your overall strategy, not optional add-ons.
But there's also a sense that we're waiting for the other shoe to be announced: Apple's WWDC. The tech giant's Worldwide Developer Conference promises to unveil Apple's own AI agenda, perhaps through its partnership with OpenAI and even Google. Is it competitive? What happens if AI can't be deeply integrated into the OS like Android's Gemini? The world is waiting to see Apple's response.
At its fall hardware event, Google has time to review Apple's announcements and attempt to craft its own AI moment as powerful and immediately understandable as Steve Jobs' introduction of the iPhone. “iPod, phone, and the Internet” communicator. iPod, phone… Got it?”
People got it. But when will they get Google's AI as well? At least not from this I/O.
Publish an AI newsletter. Sign up here to start receiving it in your inbox on June 5th.