Ilya Sutskever has a new plan for safe superintelligence
Ashlee Vance | Bloomberg
“In recent months, the question 'Where is Ilya?' has become common in the world of artificial intelligence. … Now, Sutskeva [a new] The project, a venture called Safe Superintelligence Inc., aims to develop safe and powerful artificial intelligence systems, with no intention of selling AI products or services in the near future. In other words, he wants to continue his research without many of the obstacles facing rivals like OpenAI, Google, and Anthropic.
AI revolutionizes drug development
Steve Lohr | The New York Times
“Most of the early business uses of generative AI, which can churn out everything from poetry to computer programs, have been to help alleviate the drudgery of mundane office work, customer service and writing code. But drug discovery and development is a huge industry that experts say is ripe for transformation by AI. According to consulting firm McKinsey & Company, AI is a “once-in-a-century opportunity” for the pharmaceutical industry.”
Starlink Mini brings space internet to backpackers
Thomas Ricker | The Verge
“SpaceX's Starlink space internet service is already available to ships, planes, vanlifers, Amazonian villages and rural homes in more than 75 countries, and now it's coming to backpackers too. The new compact, DC-powered Starlink Mini is about the size of a thick laptop and has a Wi-Fi router built into the antenna. It can deliver speeds of over 100 Mbps while consuming less power than other Starlink devices.”
Apple's Vision Pro team is reportedly focusing on cheaper headsets
Jay Peters | The Verge
“A new report from The Information suggests that Apple may stop development of a new high-end Vision headset amid slowing sales of the Vision Pro. Instead, Apple is looking for ways to reduce component costs for the first model and appears to be working on a cheaper Vision headset that it aims to ship by the end of 2025.”
We're still waiting for the next big leap in AI
Will Knight | Wired
“More than a year after GPT-4 sparked a frenzy of new investment in AI, it may be getting harder to create big leaps in machine intelligence. Because GPT-4 and similar models are trained on vast amounts of online text, images and videos, it's becoming harder to find new sources of data to feed machine learning algorithms. Making models significantly bigger and more able to learn is expected to cost billions of dollars.”
Escape the war robot dogs
Jared Keller | Wired
“The Chinese military recently unveiled a new fighting buddy for its soldiers: a 'robot dog' equipped with a back-mounted machine gun. … China's demonstration has clearly outraged the international community, with at least one U.S. member of Congress calling for a report from the Pentagon on the 'rifle-toting robot dog' and its potential impact on national security. But if the Chinese military is in the vanguard of weaponizing robot dogs, the U.S. military isn't far behind.”
Waabi's GenAI promises to do more than power self-driving trucks
Rebecca Beran | TechCrunch
“This technology is very powerful.” [Waabi founder and CEO Raquel] Speaking to TechCrunch in a video interview, Urtasun, behind him a whiteboard with hieroglyphic-like mathematical formulas written on it, said: “This technology has an incredible ability to generalize, it's very flexible, it's very fast to develop, and it can expand into a lot of different uses beyond trucking in the future. … This could be a robo-taxis. This could be a humanoid or a warehouse robot. This technology can solve all of those use cases.”
Apple and Microsoft shrink and improve AI models
Shubham Agarwal | IEEE Spectrum
“In recent months, several major technology companies, including Apple and Microsoft, have introduced small language models (SLMs). These models are a fraction of the size of their LLM counterparts, but in many benchmarks they match or even outperform LLMs at generating text…[And] SLM does not consume as much energy as LLM and can also be run locally on devices such as smartphones and laptops rather than in the cloud, ensuring data privacy and personalization for each user.”
sensor
Microphones made from atomically-thick graphene could be used in smartphones
Alex Wilkins | New Scientist
“The main advantage of their graphene system is that it can be made much smaller than conventional microphones,” Verbiest said, adding that the membrane only needs to be 10 micrometers wide, which is 200 times smaller than a conventional microphone with similar performance.
Image credit: Li Zhang/Unsplash
