Robotics
Aurora’s Driverless Vans Can Now Journey Farther Distances Quicker Than Human DriversKirsten Korosec | TechCrunch
“Aurora’s self-driving vans can now journey nonstop on a 1,000-mile route between Fort Value and Phoenix—exceeding what a human driver can legally accomplish. The gap, and the time it takes to journey it, affords up constructive monetary implications for Aurora—and another firm hoping to commercialize self-driving semitrucks.”
Computing
OpenAI Sidesteps Nvidia With Unusually Quick Coding Mannequin on Plate-Sized ChipsBenj Edwards | Ars Technica
“The mannequin delivers code at greater than 1,000 tokens (chunks of knowledge) per second, which is reported to be roughly 15 occasions quicker than its predecessor. To check, Anthropic’s Claude Opus 4.6 in its new premium-priced quick mode reaches about 2.5 occasions its commonplace pace of 68.2 tokens per second, though it’s a bigger and extra succesful mannequin than Spark.”
Vitality
This State’s Energy Costs Are Plummeting as It Nears 100% RenewablesAlice Klein | New Scientist ($)
“The unbiased Australian Vitality Market Operator’s (AEMO) newest report exhibits that the typical wholesale electrical energy value in South Australia fell by 30 per cent within the closing quarter of 2025, in contrast with a yr earlier. In consequence, the state had the bottom value in Australia, together with Victoria, which has the second highest share of wind and photo voltaic power within the nation.”
Biotechnology
Gene Modifying That Spreads Throughout the Physique Might Remedy Extra IllnessesMichael Le Web page | New Scientist ($)
“The concept is that every cell within the physique that receives the preliminary supply will make a number of copies of the gene-editing equipment and cross most of them on to its neighbors, amplifying the impact. Which means disease-correcting modifications might be made to the DNA of extra cells.”
Future
The First Indicators of Burnout Are Coming From the Folks Who Embrace AI the MostConnie Loizos | TechCrunch
“The instruments give you the results you want, you’re employed much less laborious, everyone wins. However a brand new research printed in Harvard Enterprise Assessment follows that premise to its precise conclusion, and what it finds there isn’t a productiveness revolution. It finds firms are vulnerable to changing into burnout machines.”
Synthetic Intelligence
ALS Stole This Musician’s Voice. AI Let Him Sing Once more.Jessica Hamzelou | MIT Know-how Assessment ($)
“[ALS patient Patrick Darling] was in a position to re-create his misplaced voice utilizing an AI software educated on snippets of outdated audio recordings. One other AI software has enabled him to make use of this ‘voice clone’ to compose new songs. Darling is ready to make music once more.”
Synthetic Intelligence
Chatbots Make Horrible Docs, New Research FindsSamantha Cole | 404 Media
“When the researchers examined the LLMs with out involving customers by offering the fashions with the total textual content of every scientific situation, the fashions appropriately recognized circumstances in 94.9 p.c of circumstances. However when speaking to the contributors about those self same circumstances, the LLMs recognized related circumstances in fewer than 34.5 p.c of circumstances.”
Computing
LEDs Enter the NanoscaleRahul Rao | IEEE Spectrum
“MicroLEDs, with pixels simply micrometers throughout, have lengthy been a byword within the show world. Now, microLED-makers have begun shrinking their creations into the uncharted nano realm. …They go away a lot to be desired of their effectivity—however sooner or later, nanoLEDs may energy ultra-high-resolution digital actuality shows and high-bandwidth on-chip photonics.”
Future
Main AI Skilled Delays Timeline for Its Doable Destruction of HumanityAisha Down | The Guardian
“A number one synthetic intelligence professional has rolled again his timeline for AI doom, saying it should take longer than he initially predicted for AI methods to have the ability to code autonomously and thus pace their very own growth towards superintelligence [and doom for humanity].”
Biotechnology
CAR T-Cell Remedy Could Sluggish Neurodegenerative Circumstances Like ALSMichael Le Web page | New Scientist ($)
“Genetically engineered immune cells often known as CAR-T cells would possibly have the ability to sluggish the progress of the neurodegenerative situation amyotrophic lateral sclerosis (ALS) by killing off rogue immune cells within the mind. ‘It’s not a approach to treatment the illness,’ says Davide Trotti on the Jefferson Weinberg ALS Heart in Pennsylvania. ‘The aim is slowing down the illness.'”
Computing
Meta Plans to Add Facial Recognition Know-how to Its Good GlassesKashmir Hill, Kalley Huang, and Mike Isaac | The New York Instances ($)
“5 years in the past, Fb shut down the facial recognition system for tagging folks in photographs on its social community, saying it needed to seek out ‘the correct stability’ for a expertise that raises privateness and authorized issues. Now it needs to deliver facial recognition again. …The characteristic, internally known as ‘Identify Tag,’ would let wearers of sensible glasses determine folks and get details about them through Meta’s synthetic intelligence assistant.”
Future
I Tried RentAHuman, The place AI Brokers Employed Me to Hype Their AI StartupsReece Rogers | Wired ($)
“At its core, RentAHuman is an extension of the round AI hype machine, an ouroboros of everlasting self-promotion and sketchy motivations. For now, the bots don’t appear to have what it takes to be my boss, even relating to gig work, and I’m completely OK with that.”
Synthetic Intelligence
AI Is Getting Scary Good at Making PredictionsRoss Andersen | The Atlantic ($)
“At first, the bots didn’t fare too nicely: On the finish of 2024, no AI had even managed to put a centesimal in one of many main [forecasting] competitions. However they’ve since vaulted up the leaderboards. AIs have already proved that they will make superhuman predictions inside the bounded context of a board sport, however they might quickly be higher than us at divining the way forward for our complete messy, contingent world.”
Synthetic Intelligence
Meet the One Girl Anthropic Trusts to Train AI MoralsBerber Jin and Ellen Gamerman | The Wall Road Journal ($)
“Because the resident thinker of the tech firm Anthropic, [Amanda] Askell spends her days studying Claude’s reasoning patterns and speaking to the AI mannequin, constructing its character and addressing its misfires with prompts that may run longer than 100 pages. The intention is to endow Claude with a way of morality—a digital soul that guides the tens of millions of conversations it has with folks each week.”
House
This Startup Thinks It Can Make Rocket Gasoline From Water. Cease LaughingNoah Shachtman | Wired ($)
“It’s an concept that’s been round for the reason that Apollo period and has been touted in recent times by the likes of former NASA administrator Invoice Nelson and SpaceX’s Elon Musk. However right here’s the factor: Nobody has ever efficiently turned water into rocket gasoline, not for a spaceship of any important measurement. A startup known as Basic Galactic, led by a pair of twentysomething engineers, is aiming to be the primary.”
