Google rolls out new Gemini model that can run on robots locally
Google DeepMind on Tuesday released a new language model called Gemini Robotics On-Device that can run tasks locally on robots without requiring an internet connection.
Building on the company's previous Gemini Robotics model that was released in March, Gemini Robotics On-Device can control a robot's movements. Developers can control and fine-tune the model to suit various needs using natural language prompts.
In benchmarks, Google claims the model performs at a level close to the cloud-based Gemini Robotics model. The company says it outperforms other on-device models in general benchmarks, though it didn't name those models.
In a demo, the company showed robots running this local model doing things like unzipping bags and folding clothes. Google says that while the model was trained for ALOHA robots, it later adapted it to work on a bi-arm Franka FR3 robot and the Apollo humanoid robot by Apptronik.
Google claims the bi-arm Franka FR3 was successful in tackling scenarios and objects it hadn't 'seen' before, like doing assembly on an industrial belt.
Google DeepMind is also releasing a Gemini Robotics SDK. The company said developers can show robots 50 to 100 demonstrations of tasks to train them on new tasks using these models on the MuJoCo physics simulator.
Other AI model developers are also dipping their toes in robotics. Nvidia is building a platform to create foundation models for humanoids; Hugging Face is not only developing open models and datasets for robotics, it is actually working on robots too; and Mirae Asset-backed Korean startup RLWRLD is working on creating foundational models for robots.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time Business News
2 hours ago
- Time Business News
TikTok SEO in 2025: Boost Your Video Reach with These Easy Optimization Tips
In 2025, mastering your TikTok SEO strategy 2025 is essential if you want your videos to rank higher in TikTok search and Google results. With more users treating TikTok like a search engine, optimizing your content for TikTok SEO is no longer optional. What is TikTok SEO Strategy 2025? TikTok SEO strategy 2025 involves optimizing your TikTok videos to appear in both TikTok's internal search results and external Google searches. It includes using the right keywords in captions, on-screen text, hashtags, and even in your video's spoken words. By applying a TikTok SEO strategy, you increase your chances of reaching users actively searching for content like yours. Why You Need a TikTok SEO Strategy in 2025 Changing User Search Behavior: Millions now prefer TikTok search over Google for product reviews and lifestyle content. Millions now prefer TikTok search over Google for product reviews and lifestyle content. Google Ranks TikTok Content: TikTok videos now appear in Google search snippets. TikTok videos now appear in Google search snippets. Boosts Engagement: A smart TikTok SEO strategy can lead to more views, shares, and conversions. Key TikTok SEO Ranking Factors for 2025 To build a strong TikTok SEO strategy 2025, focus on these elements: Captions: Insert target keywords naturally in your video descriptions. On-Screen Text: Display relevant keywords directly on your video screen. Hashtags: Mix niche and trending hashtags for better discoverability. Spoken Words: Clearly say your target keywords during the video. TikTok's AI recognizes audio text. Meta Data: Though invisible to users, TikTok uses backend data for indexing videos. How to Do TikTok Keyword Research for SEO Success Effective keyword research is crucial for any TikTok SEO strategy 2025. Use these methods: TikTok Search Bar: Type your topic and look at auto-suggested queries. Type your topic and look at auto-suggested queries. Google Trends: Identify growing search topics relevant to your niche. Identify growing search topics relevant to your niche. Competitor Analysis: Observe which keywords top creators in your category are using. Observe which keywords top creators in your category are using. SEO Tools: Platforms like Ahrefs TikTok SEO guide and TikTok Creative Center provide great insights. Step-by-Step TikTok SEO Strategy for 2025 Here's a complete SEO checklist for your next video: Research trending TikTok keywords for 2025 Add your main keyword in captions and hashtags Speak the target keyword during the video Use on-screen text with SEO keywords Encourage comments, likes, and shares for engagement signals Bonus Tips for Advanced TikTok SEO Strategy 2025 Besides basics, here's how to take your TikTok SEO strategy 2025 even further: Post content consistently Experiment with different video formats Use TikTok Trends and Analytics to monitor performance Reply to audience comments to boost engagement Final Thoughts on TikTok SEO Strategy 2025 Implementing a targeted TikTok SEO strategy 2025 will help your videos get discovered by the right audience. Whether you're a small business, influencer, or digital marketer, SEO-focused content creation is key to TikTok growth. For more insights on Digital Marketing and Social Media Marketing, stay updated with Time Business News. TIME BUSINESS NEWS
Yahoo
2 hours ago
- Yahoo
Down 48%, Should You Buy the Dip on Rigetti Computing?
Rigetti Computing's stock is down 48% from its peak after a massive run-up in 2024. Competition is fierce, with giants like Google, IBM, and Nvidia in the mix. The company continues to raise cash by selling more shares, leading to significant shareholder dilution. 10 stocks we like better than Rigetti Computing › Quantum computing specialist Rigetti Computing (NASDAQ: RGTI) soared sky-high at the end of 2024, but gave back a lot of those gains in the first half of 2025. On June 25, the stock is trading 48% below January's peak prices. Will Rigetti's stock get a second wind someday soon, or could the price chart continue its retreat? Let's find out. Rigetti is no spring chicken. The company was founded in 2013 and delivered its first quantum computing services in the cloud four years later. Its systems rely on superconducting materials cooled nearly to absolute zero. It's a proven approach, but also a costly one with specialized materials managed by massive super-cooling machinery. Nobody expects quantum computers to take over every data-crunching task, even in the long run. Rigetti proposes a hybrid approach where digital computers continue to be the workhorses of everyday computing needs, but with the option to tap into quantum computers for certain workloads. So the company develops solutions across this stack of technologies, from quantum chips and systems to the old-school software that assigns quantum jobs and reads the results. Rigetti's management sees a huge long-term market in this integrated system philosophy. Last summer, management said that "practical quantum computing is poised to happen in the next few years and have significant commercial potential." A significant breakthrough in quantum computing error correction arrived five months later. Rigetti's stock soared from roughly $1 to more than $21 per share in five weeks, as investors saw a quick route to commercial applications. But the huge gains didn't last. Jensen Huang, the CEO of semiconductor giant (and quantum-to-digital software developer) Nvidia (NASDAQ: NVDA) said that practical quantum systems could be 20 years away. Huang tried to backtrack from that statement, but the market-moving effects held firm anyway. At this point, Wall Street seems to have settled on a compromise between Rigetti's quick plans and Nvidia's long-term view. Despite losing approximately half of its peak gains, Rigetti's stock is still up tenfold over the last year. Unfortunately, Rigetti runs an expensive business model with minimal sales and deep financial losses. The company reported $10.8 million of revenue in 2024, down from $12 million in 2023. Net losses nearly tripled in the same time span, from $75 million to $201 million. The soaring stock price played a large part in the loss calculations, as the fair value of Rigetti's stock-based financial instruments soared. That's a taxable expense. And of course, Rigetti spent $50 million on research and development plus $24 million on selling, general, and administrative expenses -- again, based on $10 million in full-year sales. Furthermore, Rigetti isn't the only name in the quantum computing game. The game-changing error correction that sparked last year's boom? That was an achievement from Google Quantum AI, a group in the Alphabet (NASDAQ: GOOG) (NASDAQ: GOOGL) tech empire. Rigetti lists Google, IBM (NYSE: IBM), and the Chinese government as direct rivals using similar superconducting techniques. And you already know that Nvidia competes with Rigetti's digital-to-quantum control and reporting software. This small and cash-burning company faces challenges from some of the largest, richest, and most innovative companies in the world. Rigetti keeps the lights on by selling more stock, sometimes on the open market and sometimes in direct offerings to private investors. It's an effective cash-raising strategy, but Rigetti's existing investors see massive dilution along the way. The share count is up by 61.7% over the last year: Rigetti's stock returns took some damage from dilution, but the chart above still shows a 52-week return of more than 1,100%. Cry me a river, right? That's great as long as the stock is soaring, but Rigetti's share price often shows big drops between the sudden jumps. Go back to June 2024 and look at the two-year chart at that point. You'll see a very different story: The stock sales continued while share prices dropped 76% lower. I expect Rigetti to continue raising cash on the open market, no matter where the chart is going next. Based on the company's unprofitable operations, lofty valuation, and well-heeled rivals, investors are probably looking at another lengthy downturn. Before you buy stock in Rigetti Computing, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the for investors to buy now… and Rigetti Computing wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $713,547!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $966,931!* Now, it's worth noting Stock Advisor's total average return is 1,062% — a market-crushing outperformance compared to 177% for the S&P 500. Don't miss out on the latest top 10 list, available when you join . See the 10 stocks » *Stock Advisor returns as of June 23, 2025 Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool's board of directors. Anders Bylund has positions in Alphabet, International Business Machines, and Nvidia. The Motley Fool has positions in and recommends Alphabet, International Business Machines, and Nvidia. The Motley Fool has a disclosure policy. Down 48%, Should You Buy the Dip on Rigetti Computing? was originally published by The Motley Fool Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Gizmodo
4 hours ago
- Gizmodo
Mark Zuckerberg Wants to Win AI by Copying Everyone Smarter Than Him
Mark Zuckerberg's favorite move? Copy, poach, repeat. It made him one of the richest people on the planet. Now, he's dusting off that playbook for his biggest challenge yet: the AI race. Meta, the company behind Facebook, Instagram, and WhatsApp, is on a hiring and buying spree, throwing eye-watering sums at top AI researchers and startups in a last-ditch effort to catch up to OpenAI, Google, and upstart rivals like DeepSeek. It's a full-court press to convince the tech world and investors that Meta still matters in the AI race. But here's the twist: Zuckerberg is raiding everyone else's. In the past few weeks, Zuckerberg AI team has: So far, Meta has nabbed several big names: Lucas Beyer, Alexander Kolesnikov, Xiaohua Zhai (formerly OpenAI), and others like Trapit Bansal and Jack Rae, who left Google DeepMind. The company also reportedly tried (and failed) to lure high-profile AI researchers like OpenAI's Noam Brown and Google's Koray Kavukcuoglu. The unit's mission? To build AI systems smarter than humans. Yes, really. This is Zuckerberg's version of the Thanos snap: collect the stones, snap his fingers, and will himself to the front of the AI line. Zuckerberg is also throwing cash at startups. Meta invested $14.3 billion in Scale AI and is in talks to acquire PlayAI, which develops eerily human-sounding voice agents. There were also feelers sent to Ilya Sutskever's Safe Superintelligence and Mira Murati's new lab Thinking Machines. If it feels like Zuckerberg is panic-buying his way through the AI mall before the shelves are empty, that's because he kind of is. Meta's AI assistant, Meta AI, is currently no match for OpenAI's GPT-4o, Google's Gemini 2.5 Pro, or DeepSeek's R1. Those models can reason. Meta's can reply. It's like asking a toddler to outwit a chess master. And that's the problem: Meta doesn't just want to catch up; it wants to leapfrog. Zuckerberg knows he can't wait years for homegrown breakthroughs. So he's doing what he's always done best: copy, buy, and scale fast. It's the Facebook playbook all over again. Stories, Reels, Threads: all riffs on rival ideas. And now, AI. Meta is behind in the AI race. Big time. Tools like ChatGPT, Gemini, and DeepSeek's R1 can reason, plan, and problem-solve. Meta's AI? It's still playing catch-up, better at small talk than solving real problems. Build AI that can think and reason, like OpenAI's and Google's best systems, and do it fast enough to stay relevant in the AI arms race. OpenAI's next models are about to go open source. That could erase Meta's last advantage: its free, open AI models that developers love. Zuckerberg has declared 2025 the year of AI for Meta. It has to be. The company's ad business is under pressure, Threads is floundering, and TikTok is still eating Instagram's lunch. Meanwhile, OpenAI is flirting with becoming a full-on social network, threatening Meta on its home turf. The new plan is to go all-in on 'reasoning agents,' AI tools that can think through problems step-by-step, not just autocomplete your thoughts. These agents could power business assistants, customer support bots, or even future consumer apps. But Meta needs one thing first: actual intelligence. Even OpenAI's Sam Altman is rolling his eyes. In a recent podcast with his brother, he said Meta's strategy is obvious: 'copy OpenAI, try to poach talent, outspend everyone.' But Altman warns that copying doesn't build a culture of innovation, and culture is what wins in the long run. 'I've heard that Meta thinks of us as their biggest competitor,' Altman told his brother. But 'their current AI efforts have not worked as well as they've hoped. And I respect like being aggressive and continuing to try new things.' He didn't stop there. 'They started making these, like, giant offers to a lot of people on our team, you know, like $100 million signing bonuses,' Altman said, adding:'It is crazy.' He accused Meta of 'just trying to copy OpenAI, down to the UI mistakes,' Altman added. Altman even called out Facebook's addiction to attention-hacking, saying OpenAI wants to be 'the only tech company that doesn't feel adversarial.' Translation: Meta tries to hack your brain; we're trying to help it. New episode of Uncapped with @sama. Enjoy 🤗 — Jack Altman (@jaltma) June 17, 2025Zuckerberg has always been a fast follower, not a first mover. This worked for social media. And if history is any guide, he'll land a few hits. He's great at absorbing features, scaling fast, and bulldozing rivals with sheer force. But the AI race is different. It's about innovation. So far, Meta's biggest AI play is its wallet. Whether that's enough to win a race where originality matters as much as firepower remains to be seen. For now, Zuckerberg is betting that you can't lose if you buy everyone who knows how to win. But in a world where AI shapes power, privacy, and the future of work, we all have a stake in whether these strategies actually produce safe, useful tech, or just more hype.
