OpenAI poaches top AI talent from xAI, Meta and Tesla: Report
Co-founder and president of OpenAI, Greg Brockman, confirmed the news on X and welcomed the four new hires.
The report said that the researchers will be joining OpenAI's scaling team which handles the backend hardware and software systems and data centres including Stargate.
Per the report, Ruddarraju and Dalton were both working on building xAI's Colossus, believed to be the world's largest supercomputer comprising more than 200,000 GPUs.
In a post on X Ruddarraju bid farewell to X saying, 'When I first joined, I thought everyone was absolutely nuts for thinking we could deploy 100K GPUs in 4 months, especially without a fully functioning site. Watching us go and double that, and most importantly successfully train Grok 3 made me incredibly proud... and very happy to be wrong.'
Talent poaching in AI is becoming increasingly aggressive after Meta CEO Mark Zuckerberg poached from competitors including OpenAI for his new Superintelligence team. A total of seven high profile AI researchers from OpenAI have now joined Meta's AGI team.
Last week, Mark Chen, the chief research officer at OpenAI reportedly sent out an internal memo promising to recalibrate. Chen described Meta's actions saying he felt like 'someone has broken into our home and stolen something.'
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Mint
an hour ago
- Mint
Is AI killing Google search? It might be doing the opposite
AI upstarts were supposed to lay siege to Google's search-engine dominance. So far, the defense is winning. Google's ubiquitous search tool has proven surprisingly resilient to competition from the likes of OpenAI, which is hoping people will skip the search box and ask its chatbot for answers instead. One of Google's lines of defense has been its 'AI Overview" tool, whereby users can see answers generated by its Gemini AI model hovering above their traditional search results. Alphabet Chief Executive Sundar Pichai said Wednesday that this tool now has over 2 billion monthly users, up from 1.5 billion users in its last quarterly update. Google is also rolling out an 'AI Mode" that competes more directly with chatbots. 'We see AI powering an expansion in how people are searching for and accessing information," Pichai said in a call with analysts, adding that AI features 'cause users to search more as they learn that Search can meet more of their needs." Independent analyses suggest Google's AI search strategy is indeed having an impact. Search impressions—the number of advertiser links that show up in searches, even if they aren't clicked—grew by 49% in the year since the overviews were launched, according to a May report from BrightEdge, a search-engine-optimization firm. Those trends bode well for a search advertising business that accounts for more than half of Alphabet's overall revenue. The company said Wednesday that search revenue rose 12% from a year earlier in the second quarter to $54.2 billion, a record. Analysts surveyed by FactSet had expected $52.9 billion. The company's shares rose about 3% in after-hours trading. This outperformance underlines Alphabet's continued ability to monetize its search traffic. It has also helped that the advertising market has been relatively healthy in the quarter, despite some hiccups from tariffs and an uncertain macroeconomic outlook that had clouded the picture in April. Evercore ISI analysts said in a note this month that market checks showed ad budgets rising year-over-year in the quarter after a wobbly start. The real test of Google's search-engine resilience still lies in the future. One issue is that while AI overviews are boosting how many links users see, industry metrics show people aren't actually clicking on revenue-generating links as much. When AI overviews give people the information they seek, they don't need to. That is a puzzle Google will have to solve to keep itself on top. If it can show advertisers that they are getting a sufficient return on their spending through high-quality AI answers, the overview tool may lead to more spending and more revenue for Google. The other big unknown is how new AI-infused web browsers from the startup Perplexity and reportedly from OpenAI might change how people get information. Those challenges are in their infancy, and could make inroads against Google's Chrome browser. The impact to Google's revenue, if there is one, would only show up at a later stage, when the new entrants start building a substantial ad business tied to their software. What is clear is that Google has many tools with which to respond to the challenges. The company can modify Chrome to better compete against new entrants. It can imbue Gemini into other products. It can come up with software that others can't easily replicate, such as new AI-driven results in its 'Circle to Search" feature on Android phones, where users can circle anything on their screens and get feedback about it. Google also has a record of making defensive moves when it needs to. When search traffic was poised to shift to mobile phones two decades ago, Google acquired Android and developed its mobile-phone operating system. As Apple's iPhones became ubiquitous, Google started paying Apple billions of dollars to make its search engine the default in the company's Safari browser. And when the AI boom kicked off nearly three years ago, Microsoft's early AI push raised concern about Bing taking market-share from Google. That led Google to splash out on AI computing, and Microsoft didn't make much of a dent. Google's success has helped it build a cash pile it can deploy to keep up the defense. On Wednesday, Pichai said he was upping the company's 2025 capital spending to $85 billion from an earlier plan of $75 billion. Chief Financial Officer Anat Ashkenazi said the company would spend even more next year. There is no doubt that Google faces more serious threats to its search dominance than it has in a long time, not to mention unprecedented antitrust scrutiny across the the company is chronically undervalued compared with its peers. Based on its performance so far in the AI age, it looks likely to come out far less scathed than many skeptics believe. Write to Asa Fitch at
Time of India
2 hours ago
- Time of India
AMD CEO Lisa Su predicts demand for AI chips may increase to $500 billion; says US is going to…
AMD CEO Lisa Su has said that the global demand for AI chips could rapidly increase to more than $500 billion in a few years. Speaking at an event hosted by the Hill and Valley Forum, Su said that the growth will be driven by infrastructure demands from companies such as Elon Musk's xAI and Sam Altman's OpenAI. "We're imagining that just the accelerator market—so the chips for these AI large computing systems—will be like over $500 billion in a couple of years," Su said. Lisa Su further stated that scaling production alone won't be enough. 'You have to scale the entire ecosystem,' she said. 'The US is going to be a huge piece of it.' AMD CEO Lisa Su on Trump's AI action plan Lisa Su called the Trump administration's action plan 'a really excellent blueprint' to support AI growth in the US. 'What I really like about the AI action plan is that it's quite actionable,' she said. by Taboola by Taboola Sponsored Links Sponsored Links Promoted Links Promoted Links You May Like Indian Investors Are Buying These Dubai Apartments—Here's Why Binghatti Developers FZE Explore Undo Alongside her market projections, Su said AI will drive a wide range of chip designs tailored for different sectors. 'I am a believer in there will be diversity of chips,' she said, predicting a 'Cambrian explosion' of custom chips as AI expands into fields like science, personal devices, and manufacturing. AMD and its rival Nvidia were recently given a break from the US rules that limited the export of certain AI chips to China. Lisa Su said that going forward, policy needs to be balanced, adding that allowing shipments to US allies will help make sure the country's technology remains foundations to AI systems everywhere. 'It's a tricky dribble,' she said. 'I think the administration has been doing a good job of working with us.' Samsung Galaxy Z Fold7: 7 UNMATCHED Features No Other Foldable Has! AI Masterclass for Students. Upskill Young Ones Today!– Join Now
India Today
2 hours ago
- India Today
Meta unveils mind-reading wristband that lets you control devices without touching them
Ever imagined a time when you'd be able to open an app or type a message, not by tapping on the screen or clicking the mouse, but instead by just thinking about it? Sounds like a sci-fi movie? Well, that's exactly what Meta's latest experimental technology is working towards. According to a report by The New York Times via research detailed in Nature (a well-known scientific journal), Meta has developed a wristband that can pick up electrical signals from your muscles and use them to control computers, smartphones, and other devices. The interesting bit — you don't even have to physically move. The wristband can understand your intention to move, and that's enough to trigger a response on screen. Of course, the device is still in development, but Meta says it could be ready for the market in the next few wristband is designed by researchers at Meta's Reality Labs and works using a technique called electromyography (EMG). It reads electrical signals that travel from the brain to the muscles, especially in the forearm. When you think about moving a finger or wrist, your brain sends signals, and this device catches those signals before your muscles even react. 'You don't have to actually move,' said Thomas Reardon, Meta's VP of research, in an interview. 'You just have to intend the move.'The technology relies on AI to make sense of these muscle signals. Meta gathered data from about 10,000 people wearing the wristband prototype. Using machine learning, the team trained the system to recognise patterns linked to specific movements. That means even a new user could strap on the band and start using it without needing to teach it from scratch. 'Out of the box, it can work with a new user it has never seen data for,' said Patrick Kaifosh, one of the scientists leading the project. The main difference between this and brain-implant tech like Neuralink is that Meta's wristband is non-invasive. You don't need surgery — you just wear it like a smartwatch. That makes it safer and easier to use for the general public, as well as for people with mobility issues. Researchers at Carnegie Mellon are already testing it with people who have spinal cord injuries, helping them interact with computers despite limited hand function. 'We can see their intention to type,' said Douglas Weber, professor at Carnegie mentioned at the start, the idea of reading your thoughts might sound a bit sci-fi, but the team behind the wristband is clear — it's not actually reading your mind. 'It feels like the device is reading your mind, but it is not. It is just translating your intention. It sees what you are about to do,' Reardon far, the prototype has shown that users can control a mouse, type words, and even write letters in the air that appear on a screen — all by making small finger gestures or simply intending them. With practice, people can even activate individual muscle fibres without any visible movement at all.- Ends
