Artificial Intelligence
China’s Moonshot AI Releases Model to Challenge Top US SystemsTracy Qu and Raffaele Huang | The Wall Street Journal ($)
“Moonshot said Friday that it planned to totally open-source the model, called Kimi K3, by late this month, where people will likely be generally free to download and adapt. With 2.8 trillion parameters dictating its decision-making, Kimi K3 is the world’s biggest open-source model, in keeping with the Beijing-based startup.”
Robotics
The Fight Over Humanoid Robots Has Shut Down a Automobile Factory for the First TimeJiyoung Sohn | The Wall Street Journal ($)
“The union’s response [to Hyundai’s new Atlas robot] was blunt: Atlas would never step onto a production line without employees agreeing first. This week, Hyundai’s auto employees in South Korea have gone on a partial strike. It’s the automobile industry’s first factory stoppage addressing humanoid robots.”
Computing
Psiquantum Has a Plan to Make a Massive Quantum Computer Out of LightJames O’Donnell | MIT Technology Review ($)
“PsiQuantum has attracted unusual investment and scrutiny for 2 reasons: It’s one in all the few corporations aiming directly at constructing a big and useful machine, and it’s already working with a significant chip manufacturer to construct its systems using existing semiconductor fabs.”
Tech
Generative AI Is an Engineering DisasterAlex Reisner | The Atlantic ($)
“I asked a couple of AI researchers whether or not they could name every other real-world software that scales so poorly. None of them could consider any. Even outside the world of software, it’s hard to seek out a comparable example, on condition that economy of scale is the principle that has made light bulbs, cars, and clothing so reasonably priced. By economic and engineering measures, generative AI may be the worst technology ever deployed.”
Space
How Hard Is It to Construct Orbital Data Centers, Actually?Eric Berger | Ars Technica
“You wish unprecedented heavy lift: reusable and rapid launch. You wish the power to fabricate the biggest satellites humans have ever built and to construct 100 times more of them than humans ever have for a single constellation. You’ve gotten to hope that radiation’s impacts on chips are manageable and that radiational cooling scales. You furthermore may need a couple of trillion bucks.”
Future
Once Again We Are Told AI May Be Conscious—I Study Consciousness, and I Have My DoubtsAnil Seth | The Guardian
“The knowledge processing unfolding inside Claude is not any more more likely to lead to consciousness than a simulation of a weather system is more likely to generate an actual hurricane. AI systems are getting more powerful each day. But to provide ourselves the most effective likelihood of navigating this recent world, we must always remember how different we’re from our almost-magical creations. After we sell our minds too cheaply to our machines, we not only overestimate them, we underestimate ourselves.”
Biotechnology
AI Teaches a Bitter Biology LessonReed Albergotti | Semafor
“Future discoveries and therapies will come not from a human-like understanding of science, but by easy pattern recognition of latest biological information at scale. …[Richard] Sutton’s bitter lesson is applicable to biology because a lot of the human body—not only the mind—continues to be beyond our understanding. And we’ll find the best way forward by industrializing trial-and-error experimentation until the breakthroughs materialize.”
Future
Want Experts in 10 Years? Keep AI Away From Your Beginners TodayLaëtitia Vitaud | Fast Company
“A Nordic public school system and a 300-person American law firm aren’t pursuing the identical goals. But they arrived at the identical conclusion: Beginners must first learn without assistance with a purpose to be, in a while, well assisted.”
Space
Astronomers Find an Atmosphere on a Nearby Earthlike PlanetKatrina Miller | The Recent York Times ($)
“Recent data collected by the astronomers strongly suggests that LHS 1140b has a helium-rich atmosphere. The detection, published within the journal Science, is the primary clear evidence of a potentially habitable planet with an environment, and it reinforces the concept there exists a population of worlds just like our own with the properties mandatory to sustain life.”
Future
The Big Reason Einstein Would Never Have Used AIEthan Siegel | Big Think
“No matter whether AI counts as actual ‘intelligence’ or not, the very fact is that outsourcing the expansion and refinement of your critical pondering skills means you not only don’t develop those skills yourself, those you have already got begin to decay. …That is where humans are needed most. Einstein was uncompromising about humans developing those exact skills.”
Artificial Intelligence
Simulating Every thing, Sort Of: The Promise and Limits of World ModelsSamuel Axon | Ars Technica
“As a substitute of or along with working with language, world models aim to put the groundwork for AI systems which can be able to simulating the physical world, or at the least a useful approximation of it. To look at what’s different and necessary about this concept, Ars spoke with three expert practitioners working on world models and related technologies.”
Artificial Intelligence
Meet GPT-Red: An LLM Super-Hacker OpenAI Built to Make Its Models SaferWill Douglas Heaven | MIT Technology Review ($)
“GPT-Red automates a kind of safety evaluation for software systems often known as red-teaming, which is often done by a team of human testers. The aim is to seek out as many alternative ways to interrupt or hijack a system as possible. The weak spots can then be patched before the ultimate version of the software is released.”