Google leaders see AGI arriving around 2030
So-called artificial general intelligence (AGI) — widely understood to mean AI that matches or surpasses most human capabilities — is likely to arrive sometime around 2030, Google's co-founder Sergey Brin and Google DeepMind CEO Demis Hassabis said Tuesday.
Why it matters: Much of the AI industry now sees AGI as an inevitability, with predictions of its advent ranging from two years on the inside to 10 years on the outside, but there's little consensus on exactly what it will look like or how it will change our lives.
Brin made a surprise appearance at Google's I/O developer conference Tuesday, crashing an on-stage interview with Hassabis.
The big picture: While much of Google's developer conference focused on the here and now of AI, Brin and Hassabis focused on what it will take to make AGI a reality.
Asked whether it will be enough to keep scaling up today's AI models or new techniques will be needed, Hassabis insisted both are key ingredients.
"You need to scale to the maximum the techniques that you know about and exploit them to the limit," Hassabis said during the on-stage interview with tech journalist Alex Kantrowitz. "And at the same time, you want to spend a bunch of effort on what's coming next."
Brin said he'd guess that algorithmic advances are even more significant than increases in computational power. But, he added, "both of them are coming up now, so we're kind of getting the benefits of both."
The big picture: Hassabis predicted the industry will probably need a couple more big breakthroughs to get to AGI — reiterating what he told Axios in December .
However, he said that we may already have achieved part of one breakthrough in the form of the reasoning approaches that Google, OpenAI and others have unveiled in recent months.
Reasoning models don't respond to prompts immediately but instead do more computing before they spit out an answer.
"Like most of us, we get some benefit by thinking before we speak," Brin said — joking that it's something he often has to be reminded of.
Between the lines: Google detailed a couple of new approaches Tuesday that, while less flashy than some of the other AI features the company unveiled, hinted at other novel directions.
Gemini Diffusion is a new text model that employs the diffusion approach typically used by image generators, "converting random noise into coherent text or code," per a Google blog post. The result, Google says, is a model that can generate text far faster than other approaches.
The company also debuted a mode for its models called Deep Think, which works by pursuing multiple approaches to a problem and evaluating which is most promising.
What's next: On the timing of AGI, Hassabis and Brin were asked whether they thought it would arrive before or after 2030.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Business Insider
29 minutes ago
- Business Insider
The Googlers behind NotebookLM are launching their own AI audio startup. Here's a sneak peek.
Some of the key people behind Google's viral AI podcasting app NotebookLM have launched a new startup, and it just unveiled its first product. NotebookLM went viral last year for letting users create AI-generated podcasts about any topic they liked. Some even called it Google's ChatGPT moment. NotebookLM team lead Raiza Martin, designer Jason Spielman, and engineer Stephen Hughes learned from building the app that a surprising number of people actually learn best through audio—not necessarily through text like ChatGPT. They also learned that building a product outside a tech giant like Google is much faster. So, in December 2024, they left Google to launch their own startup, Huxe, which has been operating in stealth mode ever since. On Thursday, it launched a consumer AI app that connects to users' email, calendar, and other personal feeds and generates a daily audio briefing tailored just for them. "People just really seem to digest information more clearly when it's in that audio form," Spielman told Business Insider. Huxe declined to disclose its investors and what it's raised so far. San Francisco venture firm Conviction is one investor, according to an X post from its partner Pranav Reddy. Huxe creates personalized audio Huxe aims to create a personalized feed of AI audio content so users spend less time locked onto their screens and focus more on experiencing the world around them, Martin told Business Insider. Here's what a daily audio briefing from Huxe could go like, Martin says: Good morning! Here are your latest emails. This person has followed up five times — here's what they want. Also, here's what's going on with your meetings today and some headlines you'll care about. Huxe also has a feature called Instant Deep Dive, where users can ask any question, like "What just happened with Nvidia earnings?" or "Summarize OpenAI's latest releases," and get an informative audio response. Users can also interrupt Huxe or ask it questions in real time. Huxe publicly launched the product Thursday, though it is initially only available to a select group of users. It's still early and the product could change, Huxe's cofounders said. Building outside Google was much faster Martin said she was shocked at how quickly they were able to build the first version of Huxe. It only took about a month, she says, mostly because she could just make a decision without needing any approvals from other people. "I think it was just very different building outside of Google," Martin said. "It was great, especially for testing a concept." Spielman has been clear that the idea of Google being slow is a "misconception," he said in a podcast with Sequoia Capital. Still, Google's bureaucracy has gained some infamy in Silicon Valley, and a former Googler has even created a presentation comparing this bureaucracy to " slime mold." Building out of fear isn't the answer Huxe began as an app that lets businesses chat with their data by connecting to tools like Salesforce. While Huxe has since pivoted to focus on consumers, the founders kept a key idea from the original version of Huxe: integrating with your existing apps so the AI can talk across your tools. Still, Huxe faces the risk that a tech giant ends up launching something quite similar and stealing its thunder. Google said at I/O last month that it wants Gemini to be a super-helpful assistant that connects to Gmail and other apps. OpenAI also lets business users connect with Google Drive and use a "record mode" to take notes on meetings. NotebookLM hasn't left the scene, either, and still generates podcasts on specific topics. Google announced Tuesday that it now supports sharing NotebookLM notes publicly. In the end, Martin said you can't build while looking over your shoulder, especially when AI changes so rapidly. "It's hard to be creative and focused when you're in a state of fear," she said. "The newest products that are going to be our everyday products in the future are still to be discovered."
Axios
an hour ago
- Axios
Exclusive: X piloting new program to elevate content users agree on
Elon Musk's X is piloting a new program to use its Community Notes feature to highlight posts where people of typically differing opinions actually agree with each other, Axios has learned. Why it matters: The goal is to build momentum around opinions that are widely shared, which could help make conversations on and off X less polarizing. How it works: Currently, Community Notes are used for content that has been marked as potentially misleading. Those notes are only featured in users' feeds when a note's content is deemed helpful by X users who typically disagree. As a part of X's small pilot program, it will test Community Notes on posts that are not necessarily deemed as potentially misleading, as a way to foster more public conversations around issues where X believes people tend to have more common ground than is understood. The pilot will start with a very small handful of Community Notes users in the U.S. before gradually expanding, said X's VP of product Keith Coleman. Those selected to be part of the pilot will not be notified directly that they have been selected, but will start to see callouts in their main feeds to add context to everyday posts. (See image above.) The callouts will be added to posts based on early and limited "like" engagement signals from those posts. Select contributors will be asked to rate and provide more feedback about each post, such as whether they think the content helps to unify a polarizing opinion or if it's funny. That feedback will be used to help develop an open source algorithm to identifies posts liked by people from different perspectives, per Coleman. Between the lines: X is using data from its existing Community Notes participants to help identify users with a range of viewpoints who will be a part of the pilot, Coleman said. The rest will be selected at random. Those selected to participate will have a separate feed in their timelines that will guide them to more posts to add context. Of note: For now, only selected participants from the pilot will be able to post Notes and see them in their feeds, Coleman said. Eventually, all X users will be able to add Notes and see the posts with additional context in their feeds once X gathers enough data from its pilot. Zoom out: Coleman joined Twitter in 2016 and was part of the team that helped build the original Community Notes program. He said the rollout of X's new feature will mirror the way Twitter rolled out Birdwatch, now called Community Notes, in 2021. The first cohort of people who were selected to pilot what is today Community Notes was just 1,000 people, Coleman notes. The big picture: Once seen as a bold effort to replace fact-checking, X's Community Notes feature is now being adopted by other internet giants as a preferred way of moderating internet content. Most notably, Meta eliminated fact-checkers in the U.S. in January, arguing they were too politically biased. It began testing community notes as a replacement, using open-source technology from X shortly thereafter. The bottom line: Contrary to what the public may think about social media conversations, X's data shows that "there's actually quite a lot of agreement, even on controversial issues," Coleman said. The pilot, he believes, has the potential to build momentum and energy around opinions that are widely shared, which could reshape social discourse and how we think about governing. "It has potential to change the world," he said.
Fox News
an hour ago
- Fox News
Former Biden officials ridicule Karine Jean-Pierre's book as 'bizarre cash grab'
Several former Biden officials and staffers have spoken out against ex-White House press secretary Karine Jean-Pierre's announcement that she was becoming an independent after serving in a Democratic administration, which came alongside a book announcement. "Everyone thinks this is a grift," a former official told Politico. Jean-Pierre announced Wednesday that her new book, "Independent: A Look Inside a Broken White House, Outside the Party Lines," would be released in October. "SHE was the public face telling us all that this White House was on track and that Biden was amazing. And now she doesn't even want to be a Democrat????" a former Biden White House official told the outlet. "She's making herself the middle ground here when Republicans hate her. She's not in any position to be a connector of our two party system and assuming she could be is just ego." "It's hard to believe someone could look at the past year and genuinely think, 'The party left Joe Biden — that's why I'm leaving the Democratic Party,'" a former senior spokesperson for the Biden administration told Axios. "The hubris of thinking you can position yourself as an outsider when you not only have enjoyed the perks of extreme proximity to power… but have actively wielded it from the biggest pulpit there is, is as breathtaking as it is desperate ... It's difficult to see how this is anything but a bizarre cash grab," another White House communications official told the outlet. "She made a joke about being an independent last year and now it's a book. All ideas are monetary — even the dumb ones," a former Jean-Pierre staffer told Politico. Politico reported that Gilda Squire, a publicist, worked with Jean-Pierre while she was in the White House and was included on multiple official emails, until her staffers said it was an issue and raised it with the White House counsel office. The outlet noted that Jean-Pierre was the subject of a gushing Vogue profile, in addition to a profile in Women's Health. The former press secretary also appeared on "The View" in person and from the White House while she was press secretary. Axios reporter Alex Thompson asked Squire about boosting Jean-Pierre in February 2024 and the publicist said off the record that none of it was true. Thompson reported Thursday, "Internal White House emails obtained later by Axios, however, show that Squire was copied on emails related to Jean-Pierre's publicity work. The messages included a September 2023 email describing plans for a Vogue magazine profile." Former Jean-Pierre staffer, Jeremy Edwards, reacted on X to the book announcement. "Lol," Edwards wrote. Another White House communications official told Axios, "Today Karine lost the only constituency that ever supported her — party line Democrats." White House reporters who covered the Biden administration also criticized Jean-Pierre's announcement. "Did she find the manuscript somewhere in that fat binder she toted around? If I were a historian writing about the Biden White House, I wouldn't ignore what Karine has to say, but it's not an account in which much weight will be invested — just like her briefings," a White House reporter told Fox News Digital. Another White House reporter told Fox News Digital they were "shocked" that Jean-Pierre had left the Democratic Party. "I have to pick my jaw up from the floor. It is unbelievable that she, of all people, would choose this path," the reporter said. Jean-Pierre, a veteran Democratic Party operative who also worked for the 2020 Biden-Harris campaign and served as a spokeswoman for the far-left MoveOn advocacy group, said in a post on Instagram that the book was about not always being in a "partisan stance." Jean-Pierre repeatedly defended Biden when faced with questions about the president's decline, and even laughed at a question about his ability to serve past 2024 during an interview in 2022. Jean-Pierre's publicist did not immediately return a request for comment.
