NVIDIA’s Jim Fan delves into large language models and their industry impact - Ep. 204

For NVIDIA Senior AI Scientist Jim Fan, the video game Minecraft served as the “perfect primordial soup” for his research on open-ended AI agents. In the latest AI Podcast episode, host Noah Kravitz spoke with Fan on using large language models to create AI agents — specifically to create Voyager, an AI bot built with Chat GPT-4 that can autonomously play Minecraft. AI agents are models that “can proactively take actions and then perceive the world, see the consequences of its actions, and then improve itself,” Fan said. Many current AI agents are programmed to achieve specific objectives, such as beating a game as quickly as possible or answering a question. They can work autonomously toward a particular output but lack a broader decision-making agency. Fan wondered if it was possible to have a “truly open-ended agent that can be prompted by arbitrary natural language to do open-ended, even creative things.” But he needed a flexible playground in which to test that possibility. “And that’s why we found Minecraft to be almost a perfect primordial soup for open-ended agents to emerge, because it sets up the environment so well,” he said. Minecraft at its core, after all, doesn’t set a specific key objective for players other than to survive and freely explore the open world. That became the springboard for Fan’s project, MineDojo, which eventually led to the creation of the AI bot Voyager. “Voyager leverages the power of Chat GPT-4 to write code in Javascript to execute in the game,” Fan explained. “GPT-4 then looks at the output, and if there’s an error from JavaScript or some feedback from the environment, GPT-4 does a self-reflection and tries to debug the code.” The bot learns from its mistakes and stores the correctly implemented programs in a skill library for future use, allowing for “lifelong learning.” In-game, Voyager can autonomously explore for hours, adapting its decisions based on its environment and developing skills to combat monsters and find food when needed. “We see all these behaviors come from the Voyager setup, the skill library and also the coding mechanism,” Fan explained. “We did not preprogram any of these behaviors.” He then spoke more generally about the rise and trajectory of LLMs. He foresees strong applications in software, gaming and robotics and increasingly pressing conversations surrounding AI safety. Fan encourages those looking to get involved and work with LLMs to “just do something,” whether that means using online resources or experimenting with beginner-friendly, CPU-based AI models.

00:37:38

Anima Anandkumar on Using Generative AI to Tackle Global Challenges - Ep. 203

Generative AI-based models can not only learn and understand natural languages — they can learn the very language of nature itself, presenting new possibilities for scientific research. Anima Anandkumar, Bren Professor at Caltech and senior director of AI research at NVIDIA, was recently invited to speak at the President’s Council of Advisors on Science and Technology. At the talk, Anandkumar says that generative AI was described as “an inflection point in our lives,” with discussions swirling around how to “harness it to benefit society and humanity through scientific applications.” On the latest episode of NVIDIA’s AI Podcast, host Noah Kravitz spoke with Anandkumar on generative AI’s potential to make splashes in the scientific community. It can, for example, be fed DNA, RNA, viral and bacterial data to craft a model that understands the language of genomes. That model can help predict dangerous coronavirus variants to accelerate drug and vaccine research. Generative AI can also predict extreme weather events like hurricanes or heat waves. Even with an AI boost, trying to predict natural events is challenging because of the sheer number of variables and unknowns. However, Anandkumar explains that it’s not just a matter of upsizing language models or adding compute power — it’s also about fine-tuning and setting the right parameters. “Those are the aspects we’re working on at NVIDIA and Caltech, in collaboration with many other organizations, to say, ‘How do we capture the multitude of scales present in the natural world?’” she said. “With the limited data we have, can we hope to extrapolate to finer scales? Can we hope to embed the right constraints and come up with physically valid predictions that make a big impact?” Anandkumar adds that to ensure AI models are responsibly and safely used, existing laws must be strengthened to prevent dangerous downstream applications. She also talks about the AI boom, which is transforming the role of humans across industries, and problems yet to be solved. “This is the research advice I give to everyone: the most important thing is the question, not the answer,” she said.

00:40:08

Deepdub’s Ofir Krakowski on Redefining Dubbing from Hollywood to Bollywood - Ep. 202

In the global entertainment landscape, TV show and film production stretches far beyond Hollywood or Bollywood — it's a worldwide phenomenon. However, while streaming platforms have broadened the reach of content, dubbing and translation technology still has plenty of room for growth. Deepdub acts as a digital bridge, providing access to content by using generative AI to break down language and cultural barriers. On the latest episode of NVIDIA’s AI Podcast, host Noah Kravitz spoke with the Israel-based startup’s co-founder and CEO, Ofir Krakowski. Deepdub uses AI-driven dubbing to help entertainment companies boost efficiency and cut costs while increasing accessibility. The company is a member of NVIDIA Inception, a free program that offers startups go-to-market support, expertise and technological assistance. Traditional dubbing is slow, costly and often missing the mark, Krakowski says. Current technology struggles with the subtleties of language, leaving jokes, idioms or jargon lost in translation. Deepdub offers a web-based platform that enables people to interact with sophisticated AI models to handle each part of the translation and dubbing process efficiently. It translates the text, generates a voice and mixes it into the original music and audio effects. But as Krakowkski points out, even the best AI models make mistakes, so the platform involves a human touchpoint to verify translations and ensure that generated voices sound natural and capture the right emotion. Deepdub is also working on matching lip movements to dubbed voices. Ultimately, Krakowski hopes to free the world from the restrictions placed by language barriers. “I believe that the technology will enable people to enjoy the content that is created around the world,” he said. “It will globalize storytelling and knowledge, which are currently bound by language barriers.” https://blogs.nvidia.com/blog/2023/08/30/deepdub/

00:32:37

Replit CEO Amjad Masad on Empowering the Next Billion Software Creators - Ep. 201

Replit aims to empower the next billion software creators. In this week’s episode of NVIDIA’s AI Podcast, host Noah Kraviz dives into a conversation with Replit CEO Amjad Masad. Masad says the San Francisco-based maker of a software development platform, which came up as a member of NVIDIA’s startup accelerator program, wants to bridge the gap between ideas and software, a task simplified by advances in generative AI. “Replit is fundamentally about reducing the friction between an idea and a software product,” Masad said. The company’s Ghostwriter coding AI has two main features: a code completion model and a chat model. These features not only make suggestions as users type their code, but also provide intelligent explanations of what a piece of code is doing, tracing dependencies and context. The model can even flag errors and offers solutions — like a full collaborator in a Google Docs for code. The company is also developing “make me an app” functionality. This tool allows users to provide high-level instructions to an Artificial Developer Intelligence, which then builds, tests and iterates the requested software. The aim is to make software creation accessible to all, even those with no coding experience. While this feature is still under development, Masad said the company plans to improve it over the next year, potentially having it ready for developers in the next 6 to 8 months. Going forward, Masad envisions a future where AI functions as a collaborator, able to conduct high-level tasks and even manage resources. “We're entering a period where software is going to feel more alive,” Masad said. “And so I think computing is becoming more humane, more accessible, more exciting, more natural.” For more on NVIDIA’s startup accelerator program, visit https://www.nvidia.com/en-us/startups/

00:42:29

Codeium’s Varun Mohan and Jeff Wang on Unleashing the Power of AI in Software Development - Ep. 200

The world increasingly runs on code. Accelerating the work of those who create that code will boost their productivity — and that’s just what AI startup Codeium, a member of NVIDIA’s Inception program for startups, aims to do. On the latest episode of NVIDIA’s AI Podcast, host Noah Kravitz interviewed Codeium founder and CEO Varun Mohan and Jeff Wang, the company’s head of business, about the company's business, about how AI is transforming software. Codeium's AI-powered code acceleration toolkit boasts three core features: autocomplete, chat and search. Autocomplete intelligently suggests code segments, saving developers time by minimizing the need for writing boilerplate or unit tests. At the same time the chat function empowers developers to rework or even create code with natural language queries, enhancing their coding efficiency while providing searchable context on the entire code base. Noah spoke with Mohan and Wang about the future of software development with AI, and the continued, essential role of humans in the process.

00:39:02

MosaicML's Naveen Rao on Making Custom LLMs More Accessible - Ep. 199

Startup MosaicML is on a mission to help the AI community enhance prediction accuracy, decrease costs, and save time by providing tools for easy training and deployment of large AI models. In this episode of NVIDIA's AI Podcast, host Noah Kravitz speaks with MosaicML CEO and co-founder Naveen Rao, about how the company aims to democratize access to large language models. MosaicML, a member of NVIDIA's Inception program, has identified two key barriers to widespread adoption: the difficulty of coordinating a large number of GPUs to train a model and the costs associated with this process. Making training of models accessible is key for many companies who need to control over model behavior, respect data privacy, and iterate fast to develop new products based on AI.

00:31:26

Matice Founder Jessica Whited on Harnessing Regenerative Species for Medical Breakthroughs - Ep. 198

Scientists at Matice Biosciences are using AI to study the regeneration of tissues in animals known as super-regenerators, such as salamanders and planarians. The goal of the research is to develop new treatments that will help humans heal from injuries without scarring. On the latest episode of NVIDIA’s AI Podcast, host Noah Kravtiz spoke with Jessica Whited, a regenerative biologist at Harvard University and co-founder of Matice Biosciences. https://blogs.nvidia.com/blog/2023/06/21/matice/

00:39:08

MIT's Anant Agarwal on AI in Education - Ep. 197

In the latest episode of NVIDIA's AI Podcast, Anant Agarwal, founder of edX and Chief Platform Officer at 2U, shared his vision for the future of online education and the impact of artificial intelligence in revolutionizing the learning experience. Agarwal, a strong advocate for Massive Open Online Courses MOOCs, discussed the importance of accessibility and quality in education. The MIT professor and renowned edtech pioneer also highlighted the implementation of AI-powered features in the edX platform, including the ChatGPT plugin and edX Xpert, an AI-powered learning assistant.

00:38:45

How Alex Fielding and Privateer Space Are Taking on Space Debris - Ep. 196

In this episode of the NVIDIA AI Podcast, host Noah Kravitz dives into an illuminating conversation with Alex Fielding, co-founder and CEO of Privateer Space. Fielding is a tech industry veteran, having previously worked alongside Apple co-founder Steve Wozniak on several projects, and holds a deep expertise in engineering, robotics, machine learning and AI. Privateer Space, Fielding’s latest venture, aims to address one of the most daunting challenges facing our world today: space debris. The company is creating a data infrastructure to monitor and clean up space debris, ensuring sustainable growth for the budding space economy. In essence, they’re the sanitation engineers of the cosmos. Privateer is also focused on bolstering space accessibility. All of the company’s datasets and those of its partners are being made available through APIs, so users can more easily build space applications related to Earth observation, climate science and more. Privateer Space is a part of NVIDIA Inception, a free program that offers go-to-market support, expertise and technology for AI startups. During the podcast, Fielding shares the genesis of Privateer Space, his journey from Apple to the space industry, and his subsequent work on communication between satellites at different altitudes. He also addresses the severity of space debris, explaining how every launch adds more debris, including minute yet potentially dangerous fragments like frozen propellant and paint chips. https://blogs.nvidia.com/blog/2023/05/23/privateer-space

00:40:00

Bojan Tunguz, Johnny Israeli on How AI and Crowdsourcing Can Advance Vaccine Distribution - Ep. 195

Artificial intelligence is teaming up with crowdsourcing to improve the thermo-stability of mRNA vaccines, making distribution more accessible worldwide. In this episode of NVIDIA's AI podcast, host Noah Kravitz interviewed Bojan Tunguz, a physicist and senior system software engineer at NVIDIA, and Johnny Israeli, senior manager of AI and cloud software at NVIDIA. The guests delved into AI's potential in drug discovery and the Stanford Open Vaccine competition, a machine-learning contest using crowdsourcing to tackle the thermo-stability challenges of mRNA vaccines. Kaggle, the online machine learning competition platform, hosted the Stanford Open Vaccine competition. Tunguz, a quadruple Kaggle grandmaster, shared how Kaggle has grown to encompass not just competitions, but also datasets, code, and discussions. Competitors can earn points, rankings, and status achievements across these four areas. The fusion of artificial intelligence, crowdsourcing, and machine learning competitions is opening new possibilities in drug discovery and vaccine distribution. By tapping into the collective wisdom and skills of participants worldwide, it becomes possible to solve pressing global problems, such as enhancing the thermo-stability of mRNA vaccines, allowing for a more efficient and widely accessible distribution process. Don't miss this enlightening conversation on the transformative power of AI and crowdsourcing in mRNA vaccine distribution.

00:33:05