Robotics
This Robotic Solely Wants a Single AI Mannequin to Grasp Humanlike ActionsWill Knight | Wired
“The brand new Atlas work is a giant signal that robots are beginning to expertise the sort of equal advances in robotics that ultimately led to the final language fashions that gave us ChatGPT within the subject of generative AI.”
Synthetic Intelligence
‘World Fashions,’ an Outdated Concept in AI, Mount a ComebackJohn Pavlus | Quanta Journal
“You are carrying round in your head a mannequin of how the world works. …The deep studying luminaries Yann LeCun (of Meta), Demis Hassabis (of Google DeepMind), and Yoshua Bengio (of Mila, the Quebec Synthetic Intelligence Institute) all consider world fashions are important for constructing AI methods which might be really good, scientific and secure.”
Synthetic Intelligence
Synthesia’s AI Clones Are Extra Expressive Than Ever. Quickly They’ll Be In a position to Speak Again.Rhiannon Williams | MIT Expertise Assessment
“This demonstration exhibits how a lot more durable it’s changing into to differentiate the bogus from the true. And earlier than lengthy, these avatars will even have the ability to speak again to us. However how a lot better can they get? And what would possibly interacting with AI clones do to us?”
Tech
Anthropic to Pay at Least $1.5 Billion in Landmark Copyright SettlementMelissa Korn and Jeffrey A. Trachtenberg | The Wall Avenue Journal
“The settlement may affect the end result of pending litigation between different media corporations and AI companies, and will push the tech corporations to hunt licensing agreements with content material homeowners whose works are thought-about very important for coaching functions.”
Future
Why Anthropic’s Coding Prediction Hasn’t Panned OutStephanie Palazzolo and Rocket Drew | The Data
“In March, Anthropic CEO Dario Amodei predicted that AI can be writing 90% of all code in three to 6 months. It’s been over six months since then, so how does Amodei’s prediction maintain up? We requested Anthropic’s chatbot Claude. ‘Grade: F (Failed Prediction),’ begins Claude’s reply. ‘The prediction that AI can be writing 90% of all code inside 3-6 months was wildly off the mark.'”
Synthetic Intelligence
Reducing-Edge AI Was Speculated to Get Cheaper. It’s Extra Costly Than Ever.Christopher Mims | The Wall Avenue Journal
“The most recent AI fashions are doing extra ‘considering,’ particularly when used for deep analysis, AI brokers, and coding. So whereas the worth of a unit of AI, often called a token, continues to drop, the variety of tokens wanted to perform many duties is skyrocketing. It’s the alternative of what many analysts and specialists predicted even a number of months in the past.”
Tech
Apple Is Engaged on AI-Powered Search EngineAaron Tilley | The Data
“The corporate is planning to launch the web-search function alongside its delayed Siri revamp within the spring of subsequent yr, Bloomberg additionally reported. With the search instrument, Siri can be extra able to wanting up info throughout the net with out linking to exterior providers.”
Robotics
Waymo Expands to Denver and Seattle With Its Zeekr-Made VansSean O’Kane | TechCrunch
“The brand new cities be part of a rising listing of locations the place Waymo is working within the US. Simply final week the corporate introduced that it has greater than 2,000 robotaxis in its industrial fleet countrywide, with 800 within the San Francisco Bay Space, 500 in Los Angeles, 400 in Phoenix, 100 in Austin, and ‘dozens’ in Atlanta. Waymo has additionally introduced plans to launch a industrial robotaxi providers in Dallas, Miami, and Washington, DC, subsequent yr, and not too long ago obtained a allow to start out testing in New York Metropolis.”
Synthetic Intelligence
The Much less You Know About AI, the Extra You Are More likely to Use ItHeidi Mitchell | The Wall Avenue Journal
“Relating to most new applied sciences, early adopters are usually the individuals who know and perceive the instruments the most effective. With synthetic intelligence, the alternative appears to be true. This counterintuitive discovering comes from new analysis, which means that the individuals most drawn to AI are usually those that perceive the expertise the least.”
Tech
How Tech Giants Are Spreading the Danger of the AI BuildoutMiles Kruppa | The Data
“The pace and scale of the AI buildout is now forcing [companies] to search out exterior sources of capital, an indication of how the prices of AI are weighing on even the most important tech corporations as they define plans to spend upward of $100 billion yearly on new buildings and gear.”
Future
Ought to AI Get Authorized Rights?Kylie Robison | Wired
“Within the usually unusual world of AI analysis, some persons are exploring whether or not the machines ought to have the ability to unionize. I’m joking, form of. In Silicon Valley, there’s a small however rising subject referred to as mannequin welfare, which is working to determine whether or not AI fashions are aware and deserving of ethical concerns, reminiscent of authorized rights.”
