| | | | | | OpenAI, recently raised $6.6 billion at a valuation of $157 billion to fuel frontier AI research. Your friendly AI Human, Amit Raja Naik, is here to dive deep into this latest development. | | | | | | | | | | | | The latest round of funding, led by existing investor Thrive Capital, has brought OpenAI’s total capital raised to $17.9 billion. Thrive Capital contributed approximately $1.3 billion, with the option to invest an additional $1 billion at the same valuation through 2025. Other major participants in this funding round were Microsoft, NVIDIA, SoftBank, Khosla Ventures, Altimeter Capital, Fidelity, and MGX. The Wall Street Journal reported that Microsoft’s contribution was just under $1 billion, while NVIDIA committed $100 million, and SoftBank invested $500 million. Previously reported to be participating, Apple was notably absent from the investment round. “The new funding will allow us to double down on our leadership in frontier AI research, increase compute capacity, and continue building tools that help people solve hard problems,” the company said in a blog post. That explains why it is hiring people aggressively across various job roles, both old and new, including frontiers infrastructure engineer and GPU kernel engineer. Meanwhile, OpenAI is planning to restructure itself into a for-profit public benefit corporation. This marks a shift from its current structure, which has a non-profit board overseeing a for-profit subsidiary. Altman could also receive an equity stake in the restructured company, though specific figures haven’t been confirmed yet. | | | | | | | OpenAI is NOT a Normal Company | | | | | However, the past few days have not been smooth for OpenAI, as its CTO, Mira Murati, announced her resignation after being associated with the company for over six years. She said that she was “stepping away because I want to create the time and space to do my own exploration”. Bob McGrew, the chief research officer, also announced his departure, saying, “It is time for me to take a break.” McGrew had been with OpenAI since 2017. Barret Zoph, the vice president of research, also resigned, describing his departure as a “personal decision based on how I want to evolve in the next phase of my career”. Altman took these resignations in his stride and said, “Leadership changes are a natural part of companies, especially for companies that grow so quickly and are so demanding.” He added that he obviously won’t pretend it’s natural, for these were abrupt, but “we are not a normal company”, he quipped. Altman describes being a leader at OpenAI as “all-consuming”. “On the one hand, it’s a privilege to build AGI and be the fastest-growing company that gets to put our advanced research in the hands of hundreds of millions of people. On the other, it’s relentless to lead a team through it,” he said. Despite the departures, Altman is not worried about the company as he believes its mission is strong enough to sustain it. He also believes that despite OpenAI raising substantial funds, it remains a research-focused organisation. “We’ll do whatever works to build safe AGI and figure out how to share the benefits,” said Altman in a fireside chat at DevDay 2024.
“If the answer is a rack of GPUs, we’ll do that. Right now, the answer is to really push on research, and I think you see this with o1. We have many more giant research breakthroughs to come,” he said. | | | | | | | | | Altman believes that the arrival of artificial superintelligence is just a few thousand days away. In a recent blog post titled ‘The Intelligence Age’, Altman said that, “It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I’m confident we’ll get there.” However, Altman said that they need to bring down the cost of compute and make it abundant, which requires a lot of energy and chips. “If we don’t build enough infrastructure, AI will be a very limited resource that wars will be fought over and something that will mostly be a tool for the rich people.” OpenAI researcher Noam Brown, in a recent podcast, said that the significance of the new o1 model is ‘profound’ because it represents a new dimension for scaling AI models in the domain of inference-time compute, and that “the ceiling is a lot higher than a lot of people appreciated”. Similarly, Altman recently mentioned that “o1-preview is deeply flawed, but when o1 comes out, it will feel like a major leap forward”. He described o1-preview as being at the “GPT-2 stage” of reasoning development. “Before the end of the year, o1 will support function calling along with system prompts and structured output,” he said. Moreover, Altman has defined five levels of AI development, the first stage being chatbots, followed by reasoners, which the company claims has now been achieved. The next phases are agents, innovators capable of scientific discovery, and fully autonomous organisations. | | | | | | | | | Notably, xAI raised $6 billion earlier this year, while Safe Superintelligence secured $1 billion and Anthropic raised $7.3 billion over the past year through various funding rounds and commitments. At DevDay, Altman mentioned how companies everywhere were trying to copy OpenAI. “We really deeply care about research. I think it’s easy to copy something you know works, and I actually don’t mean that as a bad thing. When people copy OpenAI, I’m like, great—the world gets more AI, that’s wonderful. But to do something new for the first time and find the new paradigm one after the another that is what motivates us.” In another development, xAI recently relocated to a new office in the Mission district in San Francisco, specifically in the same building that housed OpenAI’s headquarters for several years. “It’s my building. That’s where we created OpenAI (as an open source, non-profit!!) and Neuralink,” posted Musk on X. Enjoy the full story here. | | | | | | | | | Last week, PyTorch introduced torchao (architecture optimisation tool), a native library designed to enhance model training and inference. It achieves this by leveraging low-bit data types, quantization, and sparsity.
According to the PyTorch team, Torchao’s quantization algorithms, applicable to popular models like Llama 3 and diffusion models, have demonstrated up to 97% speedup in inference and 73% peak VRAM reduction, maintaining high accuracy. “Quantizing weights to int4 and the KV cache to int8 supports Llama 3.1 8B at full 128K context length, running in under 18.9GB of VRAM.”
Read the full story here. | | | | | | | | | | | Accenture and NVIDIA recently launched a new business group to scale AI adoption, training 30,000 professionals globally, leveraging Accenture’s AI Refinery platform and NVIDIA’s AI stack to accelerate enterprise AI, focusing on agentic AI. The Indian government launched BharatGen, its first government-funded multimodal AI initiative aimed at developing AI models for India’s diverse languages and cultural contexts, led by IIT Bombay and a consortium of academic institutions. Walmart Global Tech is set to establish a 1 million sq ft GCC in Bengaluru, marking one of the largest office deals in India. MathCo recently opened a 200,000 sq ft global delivery and intelligence centre at IWF Campus in Bengaluru, serving as its global headquarters and housing its workforce that delivers AI-powered solutions to Fortune 500 clients. It features an experience centre for client collaboration on GenAI and emerging technologies, built on its proprietary AI engine, NucliOS. | | | | | | | | | Join the NVIDIA AI Summit India from October 23–25, 2024, at the Jio World Convention Centre in Mumbai to explore AI innovations across generative AI, robotics, supercomputing, and more, with 70% of use cases addressing India's grand challenges. Don't miss the Fireside Chat with NVIDIA CEO Jensen Huang on October 24. | | | | | AIM & NVIDIA Present DevPalooza 4.0: The Ultimate Developer Meetup in Bengaluru | | | | | | | Join us at DevPalooza 4.0 in Bengaluru, powered by AIM and NVIDIA, to dive into hands-on generative AI workshops, explore applications, and network with AI professionals—click here to register and secure your spot! | | | | | | | | | Cypher 2024 marks a significant expansion as it celebrates its 8th edition by branching out to the USA in addition to its already established presence in India. Browse through the links below to learn more about the different editions of Cypher 2024. These links will guide you to comprehensive event information, including agendas, speakers, registration details, and more. | | | | | Enjoying Sector 6 (formerly AIM Daily XO)? Share it with colleagues or friends – they can sign up here. We love hearing from our readers! Have thoughts on our new format? Questions, comments, or ideas are always welcome. If there’s a specific topic in AI or analytics that you're curious about, tell us! For brand collaborations, write to info@aimmediahouse.com Stay tuned for more insights in our next edition!
Curated with ♥️ in Namma Bengaluru | | | | | | | | |
Комментариев нет:
Отправить комментарий
Примечание. Отправлять комментарии могут только участники этого блога.