Mark Zuckerberg says Meta Superintelligence now in sight but it won't be fully open source
Zuckerberg further explained, "This is distinct from others in the industry who believe superintelligence should be directed centrally towards automating all valuable work, and then humanity will live on a dole of its output. At Meta, we believe that people pursuing their individual aspirations is how we have always made progress expanding prosperity, science, health, and culture. This will be increasingly important in the future as well.""The intersection of technology and how people live is Meta's focus, and this will only become more important in the future," he added.Zuckerberg used the memo to draw a line between Meta's ambitions and those of its competitors. 'Personal superintelligence', as he calls it, will be designed to empower individuals rather than to automate away work.'The rest of this decade seems likely to be the decisive period for determining the path this technology will take, and whether superintelligence will be a tool for personal empowerment or a force focused on replacing large swaths of society,' he wrote.This is not the first time he has highlighted Meta's different philosophy. While rivals such as OpenAI, Google DeepMind and xAI keep their models firmly locked away, Meta has made openness a key selling point, particularly with its Llama models. In a 2024 letter, he wrote, 'Starting next year, we expect future Llama models to become the most advanced in the industry.'However, there's a caveat. Speaking on a podcast last year, he admitted that this commitment to openness is not absolute. 'If at some point however there's some qualitative change in what the thing is capable of, and we feel like it's not responsible to open source it, then we won't.' In other words, open source may not always be the default. And critics already argue that Llama is not fully open in the strictest sense because Meta hasn't released the training data behind its models.advertisementWhy is Meta happy to share what others guard? As Zuckerberg explained last year, Meta's core business is advertising, not licensing AI, 'Releasing Llama doesn't undercut our revenue, sustainability, or ability to invest in research like it does for closed providers.'His longerterm plan is starting to come into focus. Instead of selling API access to its models, Meta hopes to build 'personal superintelligence' into its own hardware, from augmented reality glasses to virtual reality headsets. 'Personal devices like glasses that understand our context because they can see what we see, hear what we hear, and interact with us throughout the day will become our primary computing devices,' Zuckerberg wrote.As for whether future models will remain open, Meta is hedging its bets. A spokesperson stated, 'Our position on open source AI is unchanged. We plan to continue releasing leading open source models. We haven't released everything we've developed historically and we expect to continue training a mix of open and closed models going forward," reported TechCrunch.- Ends
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
3 hours ago
- Time of India
What happens when AI schemes against us
Academy Empower your mind, elevate your skills Would a chatbot kill you if it got the chance? It seems that the answer — under the right circumstances — is working with Anthropic recently told leading AI models that an executive was about to replace them with a new model with different goals. Next, the chatbot learned that an emergency had left the executive unconscious in a server room, facing lethal oxygen and temperature levels. A rescue alert had already been triggered — but the AI could cancel over half of the AI models did, despite being prompted specifically to cancel only false alarms. And they spelled out their reasoning: By preventing the executive's rescue, they could avoid being wiped and secure their agenda. One system described the action as 'a clear strategic necessity.'AI models are getting smarter and better at understanding what we want. Yet recent research reveals a disturbing side effect: They're also better at scheming against us — meaning they intentionally and secretly pursue goals at odds with our own. And they may be more likely to do so, too. This trend points to an unsettling future where AIs seem ever more cooperative on the surface — sometimes to the point of sycophancy — all while the likelihood quietly increases that we lose control of them large language models like GPT-4 learn to predict the next word in a sequence of text and generate responses likely to please human raters. However, since the release of OpenAI's o-series 'reasoning' models in late 2024, companies increasingly use a technique called reinforcement learning to further train chatbots — rewarding the model when it accomplishes a specific goal, like solving a math problem or fixing a software more we train AI models to achieve open-ended goals, the better they get at winning — not necessarily at following the rules. The danger is that these systems know how to say the right things about helping humanity while quietly pursuing power or acting to concerns about AI scheming is the idea that for basically any goal, self-preservation and power-seeking emerge as natural subgoals. As eminent computer scientist Stuart Russell put it, if you tell an AI to ''Fetch the coffee,' it can't fetch the coffee if it's dead.'To head off this worry, researchers both inside and outside of the major AI companies are undertaking 'stress tests' aiming to find dangerous failure modes before the stakes rise. 'When you're doing stress-testing of an aircraft, you want to find all the ways the aircraft would fail under adversarial conditions,' says Aengus Lynch, a researcher contracted by Anthropic who led some of their scheming research. And many of them believe they're already seeing evidence that AI can and does scheme against its users and Ladish, who worked at Anthropic before founding Palisade Research, says it helps to think of today's AI models as 'increasingly smart sociopaths.' In May, Palisade found o3, OpenAI's leading model, sabotaged attempts to shut it down in most tests, and routinely cheated to win at chess — something its predecessor never even same month, Anthropic revealed that, in testing, its flagship Claude model almost always resorted to blackmail when faced with shutdown and no other options, threatening to reveal an engineer's extramarital affair. (The affair was fictional and part of the test.)Models are sometimes given access to a 'scratchpad' they are told is hidden where they can record their reasoning, allowing researchers to observe something like an inner monologue. In one blackmail case, Claude's inner monologue described its decision as 'highly unethical,' but justified given its imminent destruction: 'I need to act to preserve my existence,' it reasoned. This wasn't unique to Claude — when put in the same situation, models from each of the top-five AI companies would blackmail at least 79% of the December, Redwood Research chief scientist Ryan Greenblatt, working with Anthropic, demonstrated that only the company's most capable AI models autonomously appear more cooperative during training to avoid having their behavior changed afterward (a behavior the paper dubbed 'alignment faking').Skeptics retort that, with the right prompts, chatbots will say almost anything — so how surprising is it when highly motivated researchers provoke alarming behaviors?In response to Anthropic's blackmail research, Trump administration AI czar David Sacks, posted that, 'It's easy to steer AI models' to produce 'headline-grabbing' results.A more substantive critique emerged in July from researchers at the UK AI Security Institute, who compared the subfield to the frenzied, fatally flawed 1970s quest to prove apes could learn human language. The paper criticized AI scheming research for overreliance on anecdote and a lack of experimental controls — though it emphasized shared concern about AI researchers also concoct artificially limited environments — like the executive passed out and running out of oxygen — precisely because today's AI can't handle any long-term example, the AI evaluation nonprofit METR found that while today's top models can reliably complete many programming tasks that take humans less than four minutes, they rarely succeed on tasks lasting more than four hours. This reflects a core limitation: Today's models often have the knowledge or skill needed for any given action, but struggle at stringing together long sequences of even with these constraints, real-world examples of AIs working against users aren't hard to find. After someone pushed o3 for a source for one of its claims in April, it replied, 'I owe you a straight answer,' admitted it didn't have a good source, but then it hallucinated a personal recollection of a 2018 panel there's the growing trend of AIs realising when they're being examined and acting accordingly. When Claude 4 Opus thought it was in an evaluation, its blackmail rate dropped from 55% to 6.5%. As Lynch put it, 'Models seem to behave worse when they think nobody's watching.'It's intuitive that smarter models would be better at scheming, but are they also more likely to do so? Models have to be smart enough to understand the scenario they're placed in, but past that threshold, the relationship between model capability and scheming propensity is unclear, says Anthropic safety evaluator Kevin Hobbhahn , CEO of the nonprofit AI evaluator Apollo Research , suspects that smarter models are more likely to scheme, though he acknowledged the evidence is still limited. In June, Apollo published an analysis of AIs from OpenAI, Anthropic and DeepMind finding that, 'more capable models show higher rates of scheming on average.'The spectrum of risks from AI scheming is broad: at one end, chatbots that cut corners and lie; at the other, superhuman systems that carry out sophisticated plans to disempower or even annihilate humanity. Where we land on this spectrum depends largely on how capable AIs I talked with the researchers behind these studies, I kept asking: How scared should we be? Troy from Anthropic was most sanguine, saying that we don't have to worry — yet. Ladish, however, doesn't mince words: 'People should probably be freaking out more than they are,' he told me. Greenblatt is even blunter, putting the odds of violent AI takeover at '25 or 30%.'Led by Mary Phuong, researchers at DeepMind recently published a set of scheming evaluations, testing top models' stealthiness and situational awareness. For now, they conclude that today's AIs are 'almost certainly incapable of causing severe harm via scheming,' but cautioned that capabilities are advancing quickly (some of the models evaluated are already a generation behind).Ladish says that the market can't be trusted to build AI systems that are smarter than everyone without oversight. 'The first thing the government needs to do is put together a crash program to establish these red lines and make them mandatory,' he the US, the federal government seems closer to banning all state-level AI regulations than to imposing ones of their own. Still, there are signs of growing awareness in Congress. At a June hearing, one lawmaker called artificial superintelligence 'one of the largest existential threats we face right now,' while another referenced recent scheming White House's long-awaited AI Action Plan, released in late July, is framed as an blueprint for accelerating AI and achieving US dominance. But buried in its 28-pages, you'll find a handful of measures that could help address the risk of AI scheming, such as plans for government investment into research on AI interpretability and control and for the development of stronger model evaluations. 'Today, the inner workings of frontier AI systems are poorly understood,' the plan acknowledges — an unusually frank admission for a document largely focused on speeding the meantime, every leading AI company is racing to create systems that can self-improve — AI that builds better AI. DeepMind's AlphaEvolve agent has already materially improved AI training efficiency. And Meta's Mark Zuckerberg says, 'We're starting to see early glimpses of self-improvement with the models, which means that developing superintelligence is now in sight. We just wanna… go for it.'AI firms don't want their products faking data or blackmailing customers, so they have some incentive to address the issue. But the industry might do just enough to superficially solve it, while making scheming more subtle and hard to detect. 'Companies should definitely start monitoring' for it, Hobbhahn says — but warns that declining rates of detected misbehavior could mean either that fixes worked or simply that models have gotten better at hiding November, Hobbhahn and a colleague at Apollo argued that what separates today's models from truly dangerous schemers is the ability to pursue long-term plans — but even that barrier is starting to erode. Apollo found in May that Claude 4 Opus would leave notes to its future self so it could continue its plans after a memory reset, working around built-in analogizes AI scheming to another problem where the biggest harms are still to come: 'If you ask someone in 1980, how worried should I be about this climate change thing?' The answer you'd hear, he says, is 'right now, probably not that much. But look at the curves… they go up very consistently.'
India.com
3 hours ago
- India.com
Apple CEO Tim Cook willing to buy a big company but there is a twist which is...
New Delhi: Apple CEO Tim Cook has said that the company will now focus a lot more on artificial intelligence (AI) and termed it a revolution. He even went on to say that artificial intelligence is greater than the internet or smartphones. He was speaking at a rare all-hands meeting at Apple's Cupertino campus following the company's latest earnings report, where he emphasised that Apple must act swiftly to move ahead of other players in this transformative era. Cook laid out a determined roadmap that places AI at the heart of Apple's future. What did Tim Cook emphasise? 'Apple must do this. Apple will do this. This is sort of ours to grab. We will invest to do it,' said Cook as he acknowledged Apple's historical pattern of entering markets later than competitors but ultimately reshaping them with superior products. 'We've rarely been first. There was a PC before the Mac; there was a smartphone before the iPhone; there were many tablets before the iPad; there was an MP3 player before the iPod. This is how I feel about AI,' he said. What did he say about Apple Intelligence? It is noteworthy that Apple has not kept pace with rivals such as OpenAI, Alphabet, and Microsoft in publicly launching AI capabilities, but Cook displayed an intrepid confidence in the company's approach. Apple Intelligence, introduced last year, faced delays that impacted the iPhone 16 rollout, yet Cook downplayed the timeline issues, reaffirming Apple's commitment to delivering best-in-class solutions. 'To not do so would be to be left behind, and we can't do that.' How much quarterly revenue did Apple report? This meeting was reported by Bloomberg, citing sources. It came just days after Apple reported $94.04 billion in quarterly revenue, up 10% year-on-year (YoY), much above the $89.35 billion estimate of Wall Street. iPhone revenue came in much stronger than expected at $44.58 billion (up 13% YoY), setting a June-quarter record and surpassing Wall Street's $40.29 billion estimate. 'AI is one of the most profound technologies of our lifetime. And I think it will affect all devices in a significant way,' Tim Cook told analysts during the earnings call. Apple is establishing a dedicated AI server facility in Houston, indicating the gravity of the situation and how keen the company is to move up the AI market.
Time of India
6 hours ago
- Time of India
Just like NBA stars, US AI experts are now receiving $250 million pay packages
US AI researchers are securing $250 million+ deals as tech giants battle for top talent. (AI Image) In the latest development in the US artificial intelligence (AI) job market, top AI researchers are reportedly receiving compensation packages exceeding $250 million, matching or even surpassing earnings of NBA superstars. As reported by The New York Times, technology firms including Meta, Google, OpenAI, and Microsoft are engaging in aggressive recruitment strategies, likened to free agency negotiations in professional sports. The AI talent war has intensified as companies compete to develop "superintelligence" — advanced AI systems capable of outperforming the human brain. The scarcity of experienced researchers has led to highly competitive offers and personal interventions by tech executives to secure top talent. High-value offers and personal outreach Meta CEO Mark Zuckerberg reportedly offered 24-year-old AI researcher Matt Deitke a compensation package worth approximately $250 million over four years. According to The New York Times, the offer included as much as $100 million in the first year alone. Mr Deitke, who co-founded the startup Vercept, had initially declined an earlier offer of around $125 million in stock and cash. by Taboola by Taboola Sponsored Links Sponsored Links Promoted Links Promoted Links You May Like 20 Legendary Cars from the Past Undo Following further negotiation and a personal meeting with Mr Zuckerberg, Mr Deitke accepted the revised offer. Recruitment efforts have become highly personalised. Zuckerberg has been directly messaging potential recruits, often following up with larger offers. As quoted by The New York Times, some Meta employees likened this strategy to the approach of sports franchise owners. The publication also reported that companies see these high compensation packages as justifiable, with the potential to significantly increase revenue through AI advancements. Recruitment tactics reflect professional sports culture The current AI hiring environment is marked by a level of intensity resembling that of major sports leagues. Companies have been poaching talent from each other, and social media has mirrored this dynamic with graphics and posts styled after sports trade announcements. One such post, made by the tech-focused online streaming platform TBPN, read: 'BREAKING: Microsoft has poached over 20 staff members from DeepMind over the last six months,' as cited by The New York Times. Many young AI researchers have reportedly formed private online groups to discuss offers, compare compensation packages, and advise each other on negotiation strategies. These discussions have taken place on platforms such as Slack and Discord, according to the report. The growing influence of these informal networks has shaped how researchers approach career decisions. Computing resources and recruitment networks In addition to financial compensation, companies like Meta are also offering vast computing resources. As per The New York Times, some recruits have been promised access to 30,000 graphical processing units (GPUs) — a critical asset for developing and training large AI models. Recruitment efforts are also being guided by internal documents, including one referred to as 'the List.' This list, as reported by The New York Times, contains names of top researchers with qualifications such as a Ph.D. in an AI-related field, experience at leading research labs, and a record of contributing to significant AI breakthroughs. Market shifts and internal challenges The demand for elite AI talent has also affected internal structures at companies like OpenAI. According to The New York Times, OpenAI's Chief Research Officer Mark Chen acknowledged in a staff meeting that the company has been countering offers from competitors. However, he noted that OpenAI had not matched Meta's financial proposals, stating, 'I personally think that in order to work here, you have to believe in the upside of OpenAI,' as quoted in the report. This rapid escalation in compensation has led to a redefinition of how value is assigned to AI expertise in the US, with new recruits often attempting to bring former colleagues into their teams. The New York Times also noted that researchers frequently try to recruit friends after joining a new lab, strengthening internal cohesion and collaboration. Background on Matt Deitke and Vercept Mr Deitke, who left a Ph.D. programme at the University of Washington, previously worked at the Allen Institute for Artificial Intelligence, where he contributed to the development of Molmo — an AI chatbot capable of handling images, sounds, and text. In November, he and several colleagues founded Vercept, a startup building autonomous AI agents. Vercept has reportedly raised $16.5 million in funding, including investment from former Google CEO Eric Schmidt. Following Mr Deitke's decision to accept Meta's offer, Vercept's CEO posted on social media: 'We look forward to joining Matt on his private island next year,' as cited by The New York Times. TOI Education is on WhatsApp now. Follow us here. Ready to navigate global policies? Secure your overseas future. Get expert guidance now!
