OpenAI’s Mira Murati: The Woman Responsible for Pushing Generative AI into the Real World

Applications of AI


For Mira Murati, Chief Technology Officer at OpenAI, the night of November 29th was just like any other. That night, she returned home from her San Francisco office at the startup after her team finished releasing their experimental product, ChatGPT.

ChatGPT was lined up as a research demonstration. That plan changed dramatically when the chatbot got him one million users within five days of its launch. Murati says the company is at the forefront of the race to commercialize generative AI — an artificial intelligence system that can rapidly output human-like text, images, and content — even though it was at least an order of magnitude different than she expected. I noticed that there is

Murati, who heads the technology team at the 375-person company headed by CEO Sam Altman, said, “This contact with reality really helps us understand what to prioritize and where to go next. “It was a very important decision,” he said. “When we decided to build the product and bring the technology in, we definitely went through a significant shift.”

Murati is now leading OpenAI’s efforts to position ChatGPT as a standalone product, seeking ways to capitalize on the popularity of chatbots with tens of millions of consumers following the tremendous popularity of chatbots. increase. But she insists that despite her new financial obligations, her company’s main goals haven’t changed.

“Our mission is to reach artificial general intelligence and figure out how to deploy it safely,” she said, referring to future software that can perform a variety of cognitive tasks at the human level. “So we are always very careful not to lose sight of it.”

Meanwhile, the company is restructuring the tech industry. With roughly 100 million monthly users, according to SamelWeb, the launch of his ChatGPT has prompted tech giants such as Google to rethink his AI strategy and hastily eliminate rival chatbots. Companies around the world are beginning to experiment with this technology, believing it can transform industries ranging from media, finance, legal and professional services.

Shortly after ChatGPT’s release, Microsoft agreed to a “multi-year, multi-billion dollar” investment in OpenAI, valued at an estimated $10 billion. The Seattle-based group also incorporates underlying technology across his Office suite of productivity apps that are the revenue base for hundreds of millions of large enterprise customers.

Chatbots have helped senior executives at Murati and OpenAI articulate the clear business value of generative AI. “Our strategy on deployment is now expanded to … With the ChatGPT platform, we are building a direct relationship with our users,” she said. “We can ask them about their feedback and preferences and use that feedback…to make the code better,” she added.

Murati, a mechanical engineer who previously worked at augmented-reality startup Magic Leap and electric-car maker Tesla, joined the company in 2018 when it was still a nonprofit lab. She currently oversees the mass market distribution of a number of her OpenAI products, including AI image generator Dall-E, AI code generator Codex and ChatGPT.

Since her arrival, her role has evolved into testing OpenAI’s technology in the real world, enabling industry partners to build AI versions of products in everything from education to financial services, law, and healthcare. became.

Murati’s team has started working towards offering a ChatGPT Business Subscription that will allow clients to customize their version of ChatGPT for their specific purposes. We have also released a so-called plugin. This allows users to remain within ChatGPT while performing tasks such as browsing the web, buying groceries, and reserving a table at a restaurant through Instacart, OpenTable, etc.

Today, OpenAI, like many Silicon Valley software companies, advertises itself as a “platform” company that offers two types of services. One is an API or application programming interface that allows third parties to integrate OpenAI’s software into their products for a fee. Another is ChatGPT. Microsoft, the largest shareholder, will become a third independent revenue channel.

“[ChatGPT] This allows us to contact people directly and collect their feedback to make our model more coherent and more helpful,” Murati said. “and [the API] is a platform that allows others to build on top of our model. ”

OpenAI’s API customers include a variety of companies, including education startup Khan Academy, social media company Snap, and asset managers such as Morgan Stanley Wealth Management. Each one pays to tailor the tool to their own needs. OpenAI, which reportedly made $28 million in 2022, predicts to make $200 million in 2023.

The company needs to increase revenue to cover the eye-popping costs of the computing power required to train and run AI models at scale, and Altman recently said the company is “the most capital-intensive in Silicon Valley. We are on track to become a successful start-up.” history”.

ChatGPT’s running costs are estimated at $1 million per day, assuming 10 million monthly users. Microsoft chief executive Satya Nadella has claimed that Microsoft is building a supercomputer to handle the work of OpenAI and can now handle some AI computations at half the cost of its competitors. .

The potential risks of AI software are of paramount concern to policy makers, business and the research community, especially as AI software becomes more powerful.

Among them are challenges such as manipulation by AI and the spread of misinformation. Longer term, there are concerns about the existential risks posed by AI technologies that may ignore human morality and act unethically.

Last month, 350 industry insiders, including Altman, said mitigating the risk of AI-induced extinction should be a global priority alongside other societal-scale risks such as pandemics and nuclear war. rice field.

Murati said he supported the use of technical “guardrails” to mitigate, eliminate or even prevent some of these hazards.

“Progress is obviously rapid,” she says. “But today we use these systems more like auxiliary tools. We don’t rely on them blindly or totally. is like a tool [and] Creativity. “



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *