Artificial Intelligence
A Startup Claims It Broke Through a Bottleneck That’s Holding Back LLMsWill Douglas Heaven | MIT Technology Review ($)
“Based on Subquadratic, it has developed a brand new form of LLM, called SubQ, that is quicker and cheaper and uses so much less energy than every other model available on the market. The corporate also claims that SubQ is in a position to process as much as 12 times as much text directly than most other models, allowing it to perform a spread of data-heavy tasks, equivalent to analyzing tons of of documents or entire code bases.”
Robotics
The Next Humanoid Robot Might Not Look Human at AllRobert Hart | The Verge
“The subsequent humanoid robot won’t have a head. It won’t have legs. It’d even sit on a wheeled base and fold down like a deck chair. But, as Genesis AI puts it, ‘humanoid robots don’t have to look human.’ …Genesis says Eno is designed ‘around human capability’ slightly than human appearance and is meant as a completely ‘general-purpose’ robot slightly than a machine built around a single task, like folding laundry.”
Biotechnology
Chilling the Body With Drugs Could Limit Brain Damage From StrokeAlice Klein | Latest Scientist ($)
“A mix of two drugs used to treat hay fever and psychosis cooled down the core body temperature of mice and monkeys, reducing brain damage after a stroke. These medications have also undergone preliminary testing in people, and can now be evaluated in a follow-up clinical trial.”
Future
A Court Has Ruled That Google Is Chargeable for False Statements Generated by AI OverviewsFernanda González | Wired ($)
“The authorities found that, unlike traditional engines like google, which merely display lists of links with statements made by third parties, Google’s tool produced ‘independent, recent, and substantial statements’ based on a misinterpretation of data available on the web. …Google is the one entity with the flexibility to switch the technology underpinning its AI-generated summaries and, subsequently, ‘should be held accountable.'”
Artificial Intelligence
Estonia Is Giving AI Agents ‘Personal Identification Codes’Webb Wright | Gizmodo
“Estonia is attempting to bring some law and order to the Wild West that’s the world of AI agents. The small Baltic nation plans to assign each AI agent a ‘personal identification code,’ hoping to trace what agents do across the web and discover the people or corporations behind them.”
Biotechnology
Why the Human Genome’s Tangled Physicality May Confound AIPhilip Ball | Quanta Magazine
“[The AI] approach is prone to be useful, but for individuals who crave real understanding of how the genome, and ultimately life itself, works, a computational black box won’t ever suffice. And maybe more to the purpose, the genome won’t undergo the form of straightforward input-output approach that such AI models ultimately assume. That’s since the genome isn’t any blueprint or algorithm. It’s something else.”
Future
The Inevitable Weakness of MetricsBryan Gardiner | MIT Technology Review ($)
“What I believe lots of us miss—what I know I definitely missed—is that there are all the time trade-offs if you attempt to distill something necessary down to a knowledge point. After we turn to metrics to know ourselves, our social world, and culture as an entire, they’ll never come near capturing what matters. Even worse, they’ll often actively obscure it.”
Future
Just 16% of Americans Imagine AI Will Positively Impact Society, Pew Poll FindsMatt Novak | Gizmodo
“Half of adult Americans use AI chatbots, with 1 / 4 using them each day, in keeping with recent polling from Pew Research released Wednesday. That’s up from 33% of Americans who used AI chatbots in the summertime of 2024. But a small minority, 16%, consider AI may have a positive impact on society.”
Computing
Sooner Than Expected? Useful Quantum Error Correction Promised for 2028.John Timmer | Ars Technica
“‘By 2028, we’ll bring Libra, a Megaquop-scale device, able to executing a million quantum operations over tons of of logical qubits, to our customers, enabling first scientific applications in quantum chemistry, high-energy physics, and materials simulation which might be beyond the reach of classical and Noisy Intermediate-Scale Quantum (NISQ) computers today,’ Amazon’s statement said.”
Computing
Brain-Computer Interface Trials Are Taking OffJessica Hamzelou | MIT Technology Review ($)
“Over the past couple of years, the variety of BCI trial volunteers has soared. This yr, China became the primary country to approve a BCI for medical use. Advances in technology are allowing engineers to supply more features than ever. BCI research is correctly taking off.”
Robotics
Why Waymo’s Driverless Taxis Won’t Be on Your Streets Anytime SoonDavid McCabe | The Latest York Times ($)
“Waymo is increasingly facing political roadblocks because it tries to roll out its self-driving taxis powered by artificial intelligence nationwide. After early successes winning over politicians in California—its home state—and elsewhere, Waymo has stumbled in unlocking among the biggest markets within the country.”