Ramprasad G’s Post

|| Miles Brundage Exits OpenAI: What It Means for AI Policy|| In a significant move that reverberates through the AI community, Miles Brundage, a key figure in shaping OpenAI's responsible AI development, has decided to step down from his role. His departure, while a loss for the company, presents an exciting opportunity for the broader AI landscape. Brundage's departure: A longtime policy researcher and senior advisor at OpenAI, Brundage has left the company to pursue his work in the nonprofit sector. Reasoning behind the decision: Brundage believes he can have a greater impact as a policy researcher and advocate in the nonprofit sector, where he will have more freedom and independence. OpenAI's response: The company's economic research division will now fall under the leadership of its new chief economist, Ronnie Chatterji. The remaining AGI readiness team will be distributed among other OpenAI divisions. Brundage's legacy at OpenAI: He joined the company in 2018 and played a crucial role in developing responsible language generation systems. Industry implications: Brundage's departure raises questions about OpenAI's commitment to safety and the future direction of AI development. Hashtags: #AI #OpenAI #AIPolicy #ResponsibleAI #AIEthics #MilesBrundage #AIResearch #linkedInCommunity #MorningAINews Stay updated! Follow Emly Labs

As a student and then a young lawyer, every year at around Christmas time I used to by the #Economist Year Ahead and leave it untouched until the end of the next year, when I would read it and secretly enjoy some flawed predictions made by the world's most prestigious publication (observing the mismatch between expectations and ensuing reality, it was also a way to remind myself of the need of embracing the unexpected). Some futurists are bold with their predictions, perhaps because they don't expect anybody to remember them in 30 or 50 years. #elonmusk never disappoints though. He is now predicting that #artificialgeneralintelligence (or #AGI) will happen next year. Just as a reminder, AGI is distinct and broader than #generativeai (AI systems that are used to generate meaningful -and sometimes spectacular- text, sound and images) as it refers to AI systems that are capable of at least matching human intelligence on a broad range of tasks. Simplifying, shut an AI behind a closed door and engage it in a broad-ranging conversation. If you can't tell whether you are speaking with a human or a machine, General AI is served. From a #legal perspective, to my knowledge there are no statutory definitions of #AGI (the EU #AIAct defines an AI system as operating "with varying levels of autonomy", which may be a good legislative technique to avoid being bogged down in semantics when assessing a statute'a application). However, if Elon's prediction materialises, AGIs will be able to claim an exemption under article 83 of the AI Act as a result of being placed on the market/put into service before the Act's date of application. Although I suppose I am technically closer to the #accelerationist camp, I am ambivalent about our society dealing with AGI so soon (most people are still overwhelmed by still error-prone Gen-AI chatbots). But there is another interpretation of Elon's prediction. In fact, in his recent lawsuit against #openai and #samaltman, EM claimed that AGI has already been reached by OpenAI (or that that threshold is very close, the lawsuit is not very clear on that point) but that OpenAI would not acknowledge it as its new board of directors is influenced by #Microsoft which would contractually see its licensing agreement with OpenAI automatically terminated upon achieving AGI. For once, I almost hope that conspiracy theories are true..

The latest departures from OpenAI will further feed the fire under the decentralized and safe development of AGI because other than OpenAI, all the departures and new ventures consider safety as the top priority.

No one is going to save us from the great AI takeover. 🤖 OpenAI's "superalignment" team, tasked with the monumental endeavor of ensuring artificial general intelligence (AGI) doesn't go SkyNet on humanity, has dissolved. All members have either resigned, been fired, or reassigned this past week. Superalignment, the concept of pre-programming AGI with values that guarantee our safety, now seems a distant aspiration. While I don't foresee AGI materializing imminently, the exodus of senior personnel from OpenAI this week suggests the company is adopting an even more cavalier stance towards AI safety than I'd previously imagined—a disconcerting notion, given my own reservations. 🤔 All things considered, I’m glad that Prometheus gave us fire, but ask me again when my house is burning. Right now we need to be doing three things: ✅ AI researchers need to be collaborating with ethicists and with policymakers. ✅ The federal government, with input from industry and academia, needs to establish governance over AI development and deployment, in addition to clear national laws regarding data privacy. ✅ We must educate the public at large on the potential benefits and risks of AI so that public policy will appropriately represent a well-informed electorate. 🔥 What do you think should be done about limiting the dangers of AI? Share your thoughts and concerns in the comments. [Or just have your AI write a response for you, if you’re that cheeky.] Artificial General Intelligence (AGI): A hypothetical type of artificial intelligence that possesses the cognitive abilities of a human being, capable of learning and performing any intellectual task that a human can.  Like a precocious child with an insatiable appetite for knowledge and a penchant for outsmarting its creators, AGI is the brainy offspring we yearn for yet secretly dread, for who knows what mischief it might concoct when left unsupervised? Superalignment: The ambitious endeavor of teaching super-intelligent AI to behave itself, like a hyper-intelligent toddler we're trying to potty train before it redecorates the house with permanent marker. It's the hope that we can instill human values and ethical principles into a being potentially smarter than us, so it doesn't decide to use its vast intellect for, say, turning the Earth into a giant paperclip factory. #ArtificialIntelligence #DataScienceEthics #TechLeadership

#ML #AI #OpenAI #LLM #Math #statistics #numerical #optimisation ## This article just confirms that AI Silicon Valley people are a bunch of immature kids toying with technology and thinking they are playing god... ## Few reminders to put things in perspective: * ML, AI is statistics coupled with numerical optimization methods that leverage on centuries of accumulated knowledge * As an example, the so famous gradient descent method, is Newton's optimization method from the 17th century, where the second derivative is omitted since calculated Hessian matrix for a system with billions of parameters is time consuming, not to mention to invert it! * Supervised learning is regression and interpolation like within excel, but for sure in larger dimension... * So AI is very artificial, not very intelligent, but very math and statistics centered.... * We are not god developing code that is capable to find out the most probable words following another one... * Some working in AI think they will be god, sorry to disappoint them! * Some think they will get something alive from it calling them Dad! My advice, leave your computer, get out and find a girlfriend!

Ilya Sutskever, one of OpenAI’s co-founders, has launched a new company, @ #SafeSuperintelligence Inc. (SSI), just one month after formally leaving OpenAI. Sutskever, who was OpenAI’s longtime chief scientist, founded SSI with former Y Combinator partner Daniel Gross and ex-OpenAI engineer Daniel Levy. Sutskever has been shining a light on the thornier aspects of AI safety for a long time now. In a blog post published in 2023, Sutskever, writing with Leike, predicted that AI with intelligence exceeding that of humans could arrive within the decade—and that when it does, it won’t necessarily be benevolent, necessitating research into ways to control and restrict it. Wednesday afternoon, a tweet announcing the formation of Sutskever’s new company states that: “SSI is our mission, our name, and our entire product roadmap, because it is our sole focus. Our team, investors, and business model are all aligned to achieve SSI. We approach safety and capabilities in tandem, as technical problems to be solved through revolutionary engineering and scientific breakthroughs.” SSI has offices in Palo Alto and Tel Aviv, where it is currently recruiting technical talent. #ai #genai #openai #tech #breakingnews https://lnkd.in/g5fuADXh

:- OpenAI's Chief Scientist, Ilya Sutskever, Set to Depart from The Company Ilya Sutskever, one of the leading figures in the field of artificial intelligence and the chief scientist of OpenAI, has decided to exit the firm. This move comes as a major departure for the company, which was co-founded by Sutskever himself along with technological luminaires like Elon Musk & Sam Altman. OpenAI has been a frontrunner in advanced AI research, particularly in the field of machine learning. Under Sutskever’s guidance, the company has made remarkable strides such as creating AI that can write like humans and even play complex video games. Sutskever's next destination remains unknown. In a recent interview, he only commented that he's "looking forward to the next thing". This decision comes in alignment with the company's recent shift towards safety and policy research. Founded in 2015, OpenAI originally aimed to advance digital intelligence in a way that would benefit humanity. It evolved from a nonprofit to a for-profit entity under OpenAI LP in 2019, which has drawn criticism from industry analysts arguing against potential, future misuse of artificial intelligence technology. Sutskever's departure marks a turning point in the history of OpenAI; his leadership and expertise in machine learning have been instrumental in driving the company's narrative and success in AI development. His exit will leave a void that will be felt by the AI community and will entail some significant changes for OpenAI. #AI, #OpenAI, #IlyaSutskever Join MarTech community here: https://tally.so/r/3NWD6B for Exclusive updates. Update as of 15/05/2024 & You can read more about this here: https://lnkd.in/drmtpfs6 by WIRED 🔗Follow for more 📈🫡 & Lets connect: https://lnkd.in/dkis_kdf ----------------- #ai #tech

THE OPENAI DEPARTURE DRAMA CONTINUES In a surprising turn of events, OpenAI's top executives, including Chief Technology Officer Mira Murati and AI Research Scientist Barret Zoph, have announced their departure "to explore other opportunities." This comes at a pivotal moment as the company reshapes itself to fully embrace a for-profit model, having successfully raised $6.5 billion and reaching a staggering valuation of $150 billion. Interestingly, Sam Altman is also stepping into new territory, securing a 7.5% share of the company for the first time. Altman has previously emphasized the necessity of a for-profit structure, stating, "We had to become a somehow for-profit company because creating AGI requires it for computation, training, and more." As we witness these departures, we can't help but wonder: How will this impact the future of AI? OpenAI has undeniably become the seed from which the AGI tree will grow, leading to the emergence of new branches. Visionaries like Ilya Sutskever, Elon Musk, and Dario Amodei are already charting their own paths with companies like SSI, XAI, and Anthropic. This scenario could be seen as a shedding of leaves in the pursuit of capitalism, but it may also signal the dawn of a new AI era—one in turbo mode. What we are witnessing could be likened to a mitosis of innovation, where the ideas and talents cultivated at OpenAI branch out to form diverse new companies, each contributing to the rapid evolution of artificial intelligence. The name, OpenAI, doesn't necessarily mean open-sourced, but it is open for execs to leave and start their own ventures to create the AGI competition among each other while putting the responsibility of safety in mind since the reason why they leave OpenAI is safety first - XAI, Anthropic, and XAI put safety first, therefore, OpenAI needs to do so as well. The Wild West of AI as a soap opera is unfolding, and it’s entertaining to watch how things will turn out with the most intelligent people in the world controlling the largest amounts of investment without truly understanding how neural networks work. #OpenAI #AI #Leadership #Innovation #AGI #Technology

Sam Altman, CEO of OpenAI, shared some bold words on the path to AGI during his talk on Stanford eCorner. 'Whether we burn $500 million, $5 billion, or $50 billion a year, I don’t care. I genuinely don’t as long as we can stay on a trajectory where eventually we create way more value for society than that,' he said. 'We are making AGI, and it is going to be expensive and totally worth it.' 🚀 🤖 Follow @mindwired.ai for more interesting AI & Tech news and updates 🔥 That's all! For the latest in AI news, insightful tips, and significant advancements, join our free newsletter, with over 10k subscribers. 🔗 Link in Bio #OpenAI #AGI #Innovation #FutureTech #AI

Last week Sam Altman was questioned on AI Safety by Nicholas Thompson, the CEO of the Atlantic (which has incidentally just announced a deal for OpenAI to share its content to train new AI models). The conversation was the highlight of the UN’s AI for Good Summit, organized by the International Telecommunication Union. Among my favorites was the air plane analogy. Sam Altman skillfully evaded the question whether or how we can know that we are not going to crash, given that we are all passengers in his air plane right now. Nicholas brought up the public's concerns about Iliya Sutskever, the safety concious chief AI wizard, leaving OpenAI but was eloquently persuaded that industrial progress and effiency will benefit the public more than a collective democratic discussion about the exact form and shape of an AI that humanity actually wants and needs. Altman seems to be of the opinion that the more human-like we can make AI, the better for automation and progress. Robots that can perform manual labour, voice assistants that cannot be distinguished from call center agents and deep fakes that impersonate politicians on social media to rig elections are all features of Altman's AI utopia. There ARE other use cases of AI that do not require the system to impersonate or replace humans, but that can discover new drugs, scan millions of historic documents, detect cancers and predict hurricanes. Since regulators are not likely to put a break on the development of AGI and the mass acquisition of GPU servers by these billion dollar companies, the public, *Humans*, need to have an open debate about what kind of AI we truly want for a better world and build it ourselves. Many Open Source projects allow us to build the agents and chat assistants we need to automate repetive tasks and gain new insights into ourselves and the systems we ve built, without donating our most valuable commodity, our data to a handful of rich and powerful individuals. What kind of AI will truly benefit humanity in your opinion? What should it do well, and if this was possible, what should it be forbidden from doing? I love respectful debate, please let me know below 🌈🌻🤖 https://lnkd.in/dydVV6x6 https://lnkd.in/eqEpYUXQ #ai #samaltman #aisafety #agi #unitednations