San Francisco-based AI startup Inflection AI just raised $1.3 billion in a new round of funding from the likes of Microsoft, Nvidiaand before Google CEO Eric Schmidt – not to mention a certain Bill Gates. This grant will support further development of Infection AI’s first product, a personal assistant and companion named Number of Pi (Personal AI) was launched in May.
The funding brings the total Inflection has raised to $1.5 billion. The company previously raised $225 million in early 2022 from a number of similar investors, as well as former Meta CTO Mike “Schrep” Schoepfer and Google. DeepMind co-founder and CEO Demis Hassabis and pop artist Will.i.am.
Inflection’s goal is to build a kind of universal digital assistant powered by AI. Many technologists, including Gates, see this type of assistant as the future for all human-computer interaction.
“Whoever wins personal agency, that’s important, because you’ll never go to the search page again, you’ll never go to the productivity page, you’ll never visit the page again. search. Amazon again,” Gates said at a talk on artificial intelligence sponsored by Goldman Sachs- and SV Angel in San Francisco in May. Just a few weeks ago, Gates participated in another round of “unicorn” funding or more than $1 billion for an AI startup, as he was part of a group of prominent investors pouring money. money into this project. start mining Metal KoBold.
The size of Inflection’s new funding round reflects investors’ soaring enthusiasm for innovative AI pioneering startups, especially those that make large language models ( LLM) is essentially at the heart of the current AI boom. In recent months, startups like Cohere AI, Anthropic, and Runway have announced funding rounds worth hundreds of millions of dollars. But the more than $1 billion funding for Inflection also reflects the enormous cost of creating these AI models, which must be trained on expensive, specialized computer chips in data centers whether big.
“A powerful benefit of the AI revolution is the ability to use natural conversational language to interact with supercomputers to simplify aspects of life,” said Jensen Huang, founder and CEO of NVIDIA. our daily lives”. “The world-class team at Inflection AI are helping to lead this groundbreaking work, deploying NVIDIA AI technology to develop, train, and deploy massive generic AI models that power technical assistants great personal digital.”
Inflection AI was co-founded by Mustafa Suleyman, currently the CEO of the startup. Suleyman, who co-founded advanced AI research lab DeepMind with Hassabis in 2012. Google acquired DeepMind in 2014 for about $650 million. Suleyman left DeepMind in 2019 following allegations of bullying by several employees, but then returned to the company briefly before joining Google as vice president of product management and policy books for AI. He then left Google entirely in 2022 to join the venture capital firm Greylock Partners, where Inflection AI was incubated. Greylock Partner Reid Hoffman, Co-Founder PayPal And LinkedIn and an early supporter of OpenAI, and co-founder of Inflection.
Some venture capitalists have also predicted the integration of AI personal assistants into everyday life—Marc Andreessen, billionaire founder of Andreessen Horowitz, wrote in a 7,000-word article. manifesto that chatbots like Pi will one day be ubiquitous in everything from the creative arts to government.
“Each person will have an AI assistant/coach/mentor/coach/mentor/therapist who is incredibly patient, incredibly compassionate, incredibly knowledgeable, and incredibly helpful,” said Andreessen. wrote in June. “AI assistants will be present in every opportunity and challenge of life, maximizing each person’s outcome.”
Inflection’s first product, a chatbot called Pi, was much more limited than this kind of genie-like personal agent. In fact, it is more limited than many competing chatbots, such as OpenAI’s ChatGPT and Google’s Bard. Designed Pi, Suleyman tell Fortunesimply an empathetic listener and a good communicator, not a tool for writing research reports, brainstorming marketing ideas, or coding software—all tasks that chatbots do. competition is possible.
“Personal AI will become the most transformative tool in our lives. This is really an inflection point,” says Suleyman. “We are excited to work with NVIDIA, Microsoft and CoreWeave as well as Eric, Bill and many others to make this vision a reality.”
According to Inflection, the Pi is a “teacher, coach, soulmate, creative partner, and sound board” primarily helpful in being ready to engage in field conversations with users at any time. any.
But Inflection’s ambitions are much higher. The Pi-powered AI system, a large language model (LLM) the company calls Inflection-1, has outperformed several competing LLMs — including OpenAI’s GPT-3.5, the versioning system. Free ChatGPT and LLaMA, an open source AI system created by Meta AI researchers—on multiple tasks, according to Inflection. The system is still slower than the reported performance of the largest LLMs, such as OpenAI’s GPT-4 models and Google’s PaLM 2.
Inflection has signaled its intention to overcome these competing systems though in a press release announcing its latest round of funding. It said it would use the money raised to build a supercomputer, in conjunction with chipmaker Nvidia and computing infrastructure company CoreWeave, to train massive AI models. The supercomputer will include 22,000 Nvidia’s most powerful computer chips for AI applications, the H100 Tensor Core graphics processing unit. This will be one of the most powerful GPU clusters in the world.
The supercomputer recently completed the MLPerf reference training task, a benchmark that measures the training and performance of AI and other technology, in the most modern time of 11 minutes.