Taken from JRE #1919 w/Bret Weinstein:
Category: robotics/AI – Page 1,228
ChatGPT integration with Bing’s search engine.
Microsoft’s Bing search engine might soon become more attuned to users’ needs and return results in a more human-like fashion. According to The Information, the tech giant is planning to incorporate the OpenAI software powering ChatGPT into Bing in hopes that it can help the company catch up to (or maybe even outshine) Google. Microsoft invested $1 billion in OpenAI back in 2019, and more recent reports said it’s in talks with the Elon Musk-founded startup for a follow-up investment. Now, The Information is reporting that Microsoft’s initial investment included an agreement to incorporate some aspects of GPT into Bing.
OpenAI developed GPT as a language model that uses deep learning to generate human-like text responses. Late last year, it launched a program called ChatGPT that quickly skyrocketed in popularity due to its ability to return responses that seem like they were written by actual people. Educators raised concerns that it could easily be used for cheating, since those who tried the tool said they would’ve given its responses a good grade if a student claimed to have written them. ChatGPT is free for now, but OpenAI intends to charge for its use in the future.
What GPT integration would mean for Bing isn’t clear at this point. It doesn’t have the capability to scrape the internet for results, so Microsoft will still use its own search engine technology. However, The Information said it could help Bing present results to users in a friendlier way. A source “familiar with Bing’s systems” told the publication that Microsoft could use the technology to present search results as full sentence answers with the source for the information. If Bing suggests related queries to the original one, GPT could explain their relevance in a meaningful manner. The technology could help Bing suggest better keywords or key phrases to users, as well.
Greg Yang is a mathematician and AI researcher at Microsoft Research who for the past several years has done incredibly original theoretical work in the understanding of large artificial neural networks. Greg received his bachelors in mathematics from Harvard University in 2018 and while there won the Hoopes prize for best undergraduate thesis. He also received an Honorable Mention for the Morgan Prize for Outstanding Research in Mathematics by an Undergraduate Student in 2018 and was an invited speaker at the International Congress of Chinese Mathematicians in 2019.
In this episode, we get a sample of Greg’s work, which goes under the name “Tensor Programs” and currently spans five highly technical papers. The route chosen to compress Tensor Programs into the scope of a conversational video is to place its main concepts under the umbrella of one larger, central, and time-tested idea: that of taking a large N limit. This occurs most famously in the Law of Large Numbers and the Central Limit Theorem, which then play a fundamental role in the branch of mathematics known as Random Matrix Theory (RMT). We review this foundational material and then show how Tensor Programs (TP) generalizes this classical work, offering new proofs of RMT. We conclude with the applications of Tensor Programs to a (rare!) rigorous theory of neural networks.
Patreon: https://www.patreon.com/timothynguyen.
Part I. Introduction.
00:00:00 : Biography.
00:02:36 : Harvard hiatus 1: Becoming a DJ
00:07:40 : I really want to make AGI happen (back in 2012)
00:09:00 : Harvard math applicants and culture.
00:17:33 : Harvard hiatus 2: Math autodidact.
00:21:51 : Friendship with Shing-Tung Yau.
00:24:06 : Landing a job at Microsoft Research: Two Fields Medalists are all you need.
00:26:13 : Technical intro: The Big Picture.
00:28:12 : Whiteboard outline.
Part II. Classical Probability Theory.
00:37:03 : Law of Large Numbers.
00:45:23 : Tensor Programs Preview.
00:47:25 : Central Limit Theorem.
00:56:55 : Proof of CLT: Moment method.
01:02:00 : Moment method explicit computations.
Part III. Random Matrix Theory.
Though OpenAI CEO Sam Altman recently said it would be “a mistake” to use ChatGPT for anything important.
Microsoft aims to integrate OpenAI’s ChatGPT chatbot into its Bing search engine in order to boost its user count and rival Google, as per a report from The Information.
Could OpenAI enhance Microsoft’s search engine credibility?
Picture alliance/Getty Images.
According to the report, Microsoft hopes a more conversational search experience, in which a bot provides contextual replies based on search queries, will lure users away from Google, which currently dominates the search landscape.
“She cooks, she cleans, and she still finds time to play ball with Elroy,” George and Jane Jetsons’ six-and-a-half-year-old son. Set in the year 2062 and described in the 1960s animated sitcom The Jetsons as an “aluminum-encased, battery-powered robotic maid” who is the “perfect answer for any modern family,” Rosie the Robot takes care of chores around the house while also serving as friend and confidante of mother Jane. Sarcastic and funny, Rosie is a hardworking nanny and aunt figure to children Elroy and Judy.
While many technologies The Jetsons predicted for 2062 have become reality, such as video calls and smart watches, the full realization of robots as the 1960s ideal friend and helper who makes life easier has yet to be fulfilled. For twenty-five years, roboticist Daniel Theobald has been on a mission to create robots that can solve the world’s most pressing problems. But rather than focus solely on making robots be more human, his calling has been to use them to help humans be more human.
The co-founder of Vecna Technologies and Vecna Robotics tells the story of a client who once came to him worried about their aging population. Vecna was one of the only modern robotics companies doing cutting edge work, and they wanted to talk about creating robots to take care of the elderly. Theobald says, “I felt that was completely backwards. Why would we build robots to take care of humans, and send our humans to work in a factory? Robots should be used to do the things that don’t matter. People need real, meaningful work like taking care of each other and the people we love. We should use technology to give us more time to be human, to do more meaningful activities like art, science, caring for the elderly, exploration of the universe, those sorts of things.”
Despite surges in fields like AI, medicine and nuclear energy, major advances in science and technology are slowing and are fewer and farther between than decades ago, according to a study published in Nature.
The researchers analyzed some 45 million scientific papers and 3.9 million patents between 1945 and 2010, examining networks of citations to assess whether breakthroughs reinforced the status quo or disrupted existing knowledge and more dramatically pushed science and technology off into new directions.
Across all major scientific and technological fields, these big disruptions—the discovery of the double helix structure of DNA, which rendered earlier research obsolete, is a good example of such research—have become less common since 1945, the researchers found.
Both the European Space Agency and NASA are planning to test even more sensitive sensors on future moon missions to try and hone in on satellite signals. If they can truly connect with sats back home, we could get closer to achieving autonomous moon travel. But eventually that won’t be enough. To help direct humans on the lunar surface, we’re going to need a fleet of satellites specifically around the moon. NASA calls its project LunaNet, and it’s part of the Gateway space station, which is the culmination of America’s plan to return to the moon. It needs to be designed to play well with ESA technology and, eventually, will be the source of high-speed internet on the moon.
Artemis I launched back in November, rounded the moon just 81 miles above the lunar surface and touched down Earth-side in December. Artemis II, which will carry astronauts around the moon in a similar trajectory, is slated to launch in late 2024, according to Space.com. Artemis III, which will be humanity’s first boots on the moon since 1972, could launch as early as 2025.
Nearly three years into the pandemic, travel has returned but hotel staff have not. Unable to find workers, hotel owners and managers are having to adapt to what they believe is the new normal.
Check out all the on-demand sessions from the Intelligent Security Summit here.
Over the last half-decade, quantum computing has attracted tremendous media attention. Why?
After all, we have computers already, which have been around since the 1940s. Is the interest because of the use cases? Better AI? Faster and more accurate pricing for financial services firms and hedge funds? Better medicines once quantum computers get a thousand times bigger?