A.I
Cognitive capabilities emerge from stealth, launches AI software engineer Devin
Shubham Sharma Venture Beat
“Human users input natural language prompts into Devin’s chatbot-style interface, and AI software engineers take them and develop detailed step-by-step plans to solve the problem. You then use developer tools to start projects, write your own code, fix problems, and test and report progress in real time, just like humans use developer tools, so you can keep an eye on everything while you work. You can. .”
Covariant Announces Universal AI Platform for Robotics
Evan Ackerman | IEEE spectrum
“[On Monday, Covariant announced] RFM-1, which the company describes as a robot-based model that gives robots “human-like reasoning abilities.” That's from the press release. There's no need to read too much into 'human-like' or 'opposite sex', but what Covariant has going on here is pretty cool. …'Our existing systems are already good enough to perform very fast and variable selection and placement,' says Covariant co-founder Pieter Abbeel. 'But we are now taking it a bit further. Whatever the task, whatever the implementation, that's the long-term vision. A robotics-based model that powers billions of robots around the world.’”
Cerebras unveils next-generation wafer-scale AI chip
Samuel K. Moore | IEEE spectrum
“Cerebras says its next-generation waferscale AI chips can deliver twice the performance of the previous generation while consuming the same amount of power. Wafer Scale Engine 3 (WSE-3) contains 4 trillion transistors, more than 50% more than the previous generation, using the latest chip manufacturing technologies. The company said it will use WSE-3 in its next-generation AI computers, which are currently installed in a data center in Dallas and make up a supercomputer capable of 8 exaflops (8 billion floating-point operations per second).
SpaceX celebrates major progress on spacecraft's third flight
Stephen Clark | Ars Technica
“SpaceX’s new-generation Starship rocket, the most powerful and largest launch pad ever built, flew halfway around the world after lifting off from South Texas on Thursday, demonstrating its ability to deliver heavy payloads to low-Earth orbit. The successful launch builds on two Starship test flights last year that achieved some, but not all, of their goals and appears set to begin launching satellites for the privately funded rocket program, allowing SpaceX to ramp up the already explosive pace of Starlink. distribution.”
This self-driving startup is using generative AI to predict traffic.
James O'Donnell | MIT Technology Review
“The new system, called Copilot4D, was trained on a trove of data from LiDAR sensors, which use light to detect how far away an object is. When you prompt the model for a situation, such as a driver recklessly merging onto a highway at high speed, it predicts how surrounding vehicles will move and then generates a LiDAR representation 5 to 10 seconds later (like a tracking dummy indicator). .”
Electric cars are not enough yet.
Andrew Moseman | Atlantic Ocean
“The next step in electric vehicles’ leap from early adoption to mass adoption depends on people. [David] Rapson calls himself a ‘pragmatist’. He's an American who buys the car he thinks is best and waits for concerns about price, range and charging to subside before he goes electric. Current EVs just can’t beat them.”
Helium-3 mining on the Moon has been talked about forever. Now companies will try.
Eric Berger | Ars Technica
“Two of Blue Origin's early employees, former President Rob Meyerson and Chief Architect Gary Lai, started a company to extract helium-3 from the lunar surface, return it to Earth, and sell it for applications there. … The current dollar rush is like the California gold rush without the gold. Interlune could help change the calculus by deriving value from the moon's resources by harvesting helium-3, which is rare and in limited supply on Earth. But many questions remain about the approach.”
What happens when ChatGPT tries to solve the 50,000 trolley problem?
Fintan Burke | Ars Technica
“Self-driving startups are currently experimenting with AI chatbot assistants, including self-driving systems that explain driving decisions. In addition to announcing red lights and turn signals, the large language models (LLMs) powering these chatbots may ultimately have to make moral decisions, such as prioritizing the safety of passengers or pedestrians. But is the technology ready? Kazuhiro Takemoto, a researcher at Kyushu Institute of Technology in Japan, wanted to see whether chatbots could make the same moral decisions when driving as humans.”
States are lining up to make lab-grown meat illegal.
Matt Reynolds | mad
“In addition to the Florida bill, bills to ban farmed meat have been proposed in Alabama, Arizona, Kentucky and Tennessee. “If all of these bills pass—an obviously unlikely prospect—about 46 million Americans will lose access to meat in a form that many hope will be much kinder to the planet and animals.”
Physicists have finally discovered a problem that only quantum computers can solve
Lakshmi Chandrasekaran's adopted son
“Quantum computers are poised to become computational superpowers, but researchers have long sought viable problems that provide quantum advantages that only quantum computers can solve. Only then, they argue, will the technology finally be considered essential. They have been looking for decades. … Now a team of physicists including [John] Preskill may have found the best candidate for quantum advantage.”
Image credit: SpaceX