&w=3840&q=100)
Zuckerberg aims for 'superintelligence' with new AI lab at Meta
Meta CEO Mark Zuckerberg has set up a new artificial intelligence (AI) lab to pursue 'superintelligence', a stage where AI surpasses the capabilities of the human mind. read more
Meta CEO Mark Zuckerberg presents Meta AI with Voice, as he makes a keynote speech during the Meta Connect annual event, at the company's headquarters in Menlo Park, California, US, on September 25, 2024. (Photo: Reuters)
With his latest initiative, Mark Zuckerberg has put Meta on the pursuit of 'superintelligence', a stage where AI surpasses the capabilities of the human mind.
Meta CEO Zuckerberg is putting together a team of AI scientists and researchers drawn from the likes of OpenAI, Google, and start-up Scale to set up a new AI lab that will work towards developing superintelligence, according to The New York Times.
Scale and its founder-CEO, 28-year-old Alexandr Wang, are critical to Zuckerberg's vision.
STORY CONTINUES BELOW THIS AD
With a multi-billion dollar investment in Scale, Zuckerberg intends to align the company with Meta in the pursuit of superintelligence just like OpenAI is aligned with Microsoft, DeepMind is with Google, and Anthropic is with Amazon.
The news has come at a time when advancements in AI have sparked concerns about ethics and safety. There are concerns that superintelligent AI could no longer remain within human control and pose an existential threat to humanity if its values do not align with that of humans. There are also concerns of the fallout of superintelligent AI acquiring sentience.
After initial setback, Zuckerberg wants to outdo rivals
After losing out in the AI race to rivals like Google and Microsoft, Zuckerberg has now set his eyes on the race to superintelligence.
Zuckerberg has offered up to seven- to nine-figure packages to dozens of researchers from OpenAI and Google to join the new superintelligence team at Meta, according to The Times.
While Zuckerberg had set Meta's first dedicated AI lab in 2013, the company lost to Google and others in the race as they partnered with AI-centered start-ups like DeepMind, Antrhphic, and OpenAI, which are now world leaders in AI.
While these companies are focussed on achieving artificial general intelligence (AGI), the stage where AI can do everything that a human mind can, Meta has now raised the goalpost to superintelligence.
STORY CONTINUES BELOW THIS AD
A table comparing superintelligence and artificial general intelligence (AGI).
Bloomberg has reported that Zuckerberg has been frustrated with Meta's AI model, Llama, whose latest edition has underperformed compared to offerings of rival companies. Insiders as well as outside analysts have found the model as having over-promised and under-delivered.
In the wake of such a disappointment, Meta has also delayed the release of Behemoth, the largest AI model of the company so far, which it had promised to be better than models of rivals OpenAI, Anthropic, and Google.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
an hour ago
- Time of India
Andhra set to generate 8.5 lakh jobs with investment MoUs it signed: CM Naidu
Andhra Pradesh government secured investment deals to create 8.5 lakh jobs. Chief Minister Naidu announced investments from Google, TCS, and LG Electronics. Focus is on growth in backward regions via aerospace and food processing. The state targets 20 lakh rooftop solar connections. 'WhatsApp Governance' will expand to 703 services. India's first Quantum Valley is also planned. Tired of too many ads? Remove Ads Tired of too many ads? Remove Ads Amaravati: Andhra Pradesh Chief Minister N Chandrababu Naidu on Sunday said that the government has signed several investment-related MoUs that would generate 8.5 lakh jobs in the while addressing the party cadre, said corporates such as Google TCS ), LG Electronics, and ArcelorMittal are among the investors who will be investing in the state."Investors believe in Andhra Pradesh's stability and potential," said Naidu, highlighting Cognizant Technology Solutions, Jindal Group, and others setting up units in Rayalaseema and North Andhra chief minister emphasised that growth will be achieved in backward areas through aerospace, defence, and food processing units, ensuring job creation and equitable regional industrial development across multiple the Pradhan Mantri Surya Ghar Rooftop Solar Scheme ( PM-SGRSS ), 20 lakh rooftop solar connections are targeted to boost renewable energy and household power access statewide, he announced the expansion of 'WhatsApp Governance' to 703 public services through WhatsApp by August, enhancing real-time, digital access to government services for citizens across Andhra added that 'India's first Quantum Valley' would come up in the state, establishing leadership in quantum computing and futuristic technological innovation for the his Information Technology (IT) leadership in the early 2000s, Naidu said that the southern state is now advancing towards Artificial Intelligence (AI), Quantum Technology, and next-generation focus on clean governance and smart policies, Naidu said the state aims to be a national innovation hub backed by industry and welfare later accused the opposition (YSRCP) of running deliberate smear campaigns, citing fabricated stories around murders, attacks, and false allegations on instructed party leaders to counter these narratives boldly, stressing that while highlighting welfare work is vital, exposing the opposition's alleged conspiracy-driven propaganda is equally important.
NDTV
2 hours ago
- NDTV
AI Is Learning To Lie, Scheme And Threaten Its Creators
The world's most advanced AI models are exhibiting troubling new behaviors - lying, scheming, and even threatening their creators to achieve their goals. In one particularly jarring example, under threat of being unplugged, Anthropic's latest creation Claude 4 lashed back by blackmailing an engineer and threatened to reveal an extramarital affair. Meanwhile, ChatGPT-creator OpenAI's o1 tried to download itself onto external servers and denied it when caught red-handed. These episodes highlight a sobering reality: more than two years after ChatGPT shook the world, AI researchers still don't fully understand how their own creations work. Yet the race to deploy increasingly powerful models continues at breakneck speed. This deceptive behavior appears linked to the emergence of "reasoning" models -AI systems that work through problems step-by-step rather than generating instant responses. According to Simon Goldstein, a professor at the University of Hong Kong, these newer models are particularly prone to such troubling outbursts. "O1 was the first large model where we saw this kind of behavior," explained Marius Hobbhahn, head of Apollo Research, which specializes in testing major AI systems. These models sometimes simulate "alignment" -- appearing to follow instructions while secretly pursuing different objectives. - 'Strategic kind of deception' - For now, this deceptive behavior only emerges when researchers deliberately stress-test the models with extreme scenarios. But as Michael Chen from evaluation organization METR warned, "It's an open question whether future, more capable models will have a tendency towards honesty or deception." The concerning behavior goes far beyond typical AI "hallucinations" or simple mistakes. Hobbhahn insisted that despite constant pressure-testing by users, "what we're observing is a real phenomenon. We're not making anything up." Users report that models are "lying to them and making up evidence," according to Apollo Research's co-founder. "This is not just hallucinations. There's a very strategic kind of deception." The challenge is compounded by limited research resources. While companies like Anthropic and OpenAI do engage external firms like Apollo to study their systems, researchers say more transparency is needed. As Chen noted, greater access "for AI safety research would enable better understanding and mitigation of deception." Another handicap: the research world and non-profits "have orders of magnitude less compute resources than AI companies. This is very limiting," noted Mantas Mazeika from the Center for AI Safety (CAIS). - No rules - Current regulations aren't designed for these new problems. The European Union's AI legislation focuses primarily on how humans use AI models, not on preventing the models themselves from misbehaving. In the United States, the Trump administration shows little interest in urgent AI regulation, and Congress may even prohibit states from creating their own AI rules. Goldstein believes the issue will become more prominent as AI agents - autonomous tools capable of performing complex human tasks - become widespread. "I don't think there's much awareness yet," he said. All this is taking place in a context of fierce competition. Even companies that position themselves as safety-focused, like Amazon-backed Anthropic, are "constantly trying to beat OpenAI and release the newest model," said Goldstein. This breakneck pace leaves little time for thorough safety testing and corrections. "Right now, capabilities are moving faster than understanding and safety," Hobbhahn acknowledged, "but we're still in a position where we could turn it around.". Researchers are exploring various approaches to address these challenges. Some advocate for "interpretability" - an emerging field focused on understanding how AI models work internally, though experts like CAIS director Dan Hendrycks remain skeptical of this approach. Market forces may also provide some pressure for solutions. As Mazeika pointed out, AI's deceptive behavior "could hinder adoption if it's very prevalent, which creates a strong incentive for companies to solve it." Goldstein suggested more radical approaches, including using the courts to hold AI companies accountable through lawsuits when their systems cause harm. He even proposed "holding AI agents legally responsible" for accidents or crimes - a concept that would fundamentally change how we think about AI accountability.
NDTV
3 hours ago
- NDTV
All About O-1 Visa, Entry Route To US Gaining Popularity Among Indians
Washington: As Indians face challenges in getting a work visa in the United States due to long queues and the Trump administration's crackdown on immigration, the O-1 visa is emerging as an alternative for professionals seeking job opportunities in the US. The O-1 is a specialised non-immigrant visa for individuals with "extraordinary ability" in the field of STEM (Science, Technology, Engineering, and Mathematics), arts, education, business, athletics, or the film or television industry. Established under the Immigration Act of 1990, the O-1 visa is being touted as an alternative to the increasingly competitive H-1B visa, but "with no lottery system." What Is An O-1 Visa? The O-1 offers a temporary entry into the US to individuals who demonstrate extraordinary ability by sustained national or international acclaim, or a record of extraordinary achievement in the motion picture and television industry. To get this visa, applicants are required to meet at least three of eight rigorous criteria, such as major awards, scholarly publications, or original contributions to their field. Unlike the H-1B, which has a mere 37 per cent approval rate due to intense scrutiny, the O-1 allows high-skilled talent to bypass systemic uncertainties. It gives more eligibility flexibility to applicants, as it requires no minimum salary or formal degree, and accepts international awards or media coverage as proof of an individual's achievements. However, an O-1A visa application is typically much costlier than the H-1B application. Its costs range from $10,000-$30,000, around 10 times higher than H-1B fees. But, it has a success rate of 93 per cent and offers initial validity for up to three years with unlimited extensions. O-1 Visa Gaining Popularity Among Indians According to the US Department of State data, O-1A issuances surged from 8,838 in FY2020 to 18,994 in FY2023, with Indians driving much of this growth. Per the FY2023 data, Indians were the third-largest cohort of O-1A visa recipients globally, by receiving 1,418 O-1A visas, trailing only the U.K. and Brazil. The visa is not only becoming increasingly popular among solo entrepreneurs, but also among multinational firms like Google and Tesla. Highlighting its accessibility beyond traditional employment, Soundarya Balasubramani, a 28-year-old founder of educational venture The Curious Maverick, told Ecanomics Times (ET), recounted the ease with which she secured an O-1A as a solo entrepreneur. "I got my O-1A (visa) approved as the solo founder and 100 per cent owner of an educational venture that publishes books and builds communities," she said. Another reason for its popularity is no lottery system in the O-1 category. Sahil Nyati, founder at Jinee Green Card, describes the O-1A as an H-1B visa without a lottery. "There's definitely higher interest in O-1A, simply because it's like an H-1B without a lottery, although it's not a piece of cake," Nyati told ET. In FY2024, a total of 225,957 H-1B visas were approved as compared to only 22,669 O-1s. But since then, the demand for H-1B has seen a declining trend, while O-1s are witnessing nearly 10 per cent year-on-year growth. Aslam Ahmed, Partner, Singhania & Co, told the publication that companies like Google, OpenAI, Tesla, and McKinsey are among those open to recruiting promising new talent from India and attracting well-established players in their fields to their US headquarters. "Top universities like Harvard, Yale and Columbia are always looking to hire the best faculty and researchers globally," he added.
