Synthetic Intelligence
Khosla-Backed Startup Claims Breakthrough With Largest-Ever AI Mannequin on an iPhoneAaron Tilley | The Info ($)
“The most important AI fashions, which might measure within the trillions of parameters, are nonetheless far too large to run on cell units. However the mannequin PrismML has engaged on an iPhone is able to duties like advanced chat, reasoning, totally autonomous brokers and software program coding, the startup mentioned.”
Robotics
Humanoid Robots Managed by Surgeons Did World-First Operation on Reside PigsJeremy Hsu | Ars Technica
“Humanoid robots have surgically eliminated the gallbladders from dwelling animals in an unprecedented medical experiment—however not as autonomous machines able to changing human docs. As a substitute, expert human surgeons remotely managed the robots’ actions in a brand new instance of human-robot teamups.”
Biotechnology
mRNA Vaccines Clear Sweeping International Assessment of Security and EffectivenessBronwyn Thompson | Refractor
“A world crew…reviewed knowledge from laboratory research, medical trials, and real-world statistics to completely examine this comparatively novel class of vaccine, from design and manufacturing to its long-term efficiency. ‘After billions of doses, we now have a rare quantity of scientific proof,’ says lead creator Dr. Anna Blakney, assistant professor at UBC’s Michael Smith Laboratories and Faculty of Biomedical Engineering.”
Tech
Anthropic, OpenAI, and SpaceX Are Greater Than the Final 25 Years of Tech ExitsRussell Brandom | TechCrunch
“SpaceX has already gone public at a $1.77 trillion valuation, and with each Anthropic and OpenAI pushing into the trillions it’s seemingly the trio collectively will land someplace north of $4 trillion. By comparability, the US Securities and Change Fee counted simply $70 billion in US-based IPO proceeds final 12 months.”
Biotechnology
11% of People Are Presently Taking a GLP-1 Weight Loss Drug Like WegovyMatt Novak | Gizmodo
“Gallup notes that weight problems reached a file excessive within the US in 2022 at 39.9% however has ticked all the way down to 36.4%, in line with the newest knowledge based mostly on self-reported top and weight. That dip has been credited to the massive variety of People now utilizing GLP-1 medicine.”
Tech
The Distrust of AI Labs Bubbles OverLiz Hoffman | Semafor
“Ahead-deployed engineers, dispatched by AI labs to assist prospects implement their fashions, will come again to the mom ship with a deep understanding of how banks, consulting corporations, retailers, producers, and consultants function. That buyer assist, to a suspicious eye, appears lots like reconnaissance.”
Future
Microsoft’s Carbon Emissions Went Up 25 % Final 12 monthsStevie Bonifield | The Verge
“Microsoft says this was ‘pushed primarily by the growth of our datacenter infrastructure,’ in addition to the corporate’s choice final February to cease buying ‘non-additional, unbundled renewable power certificates.'”
Tech
China’s Reply to AI Sticker ShockMatteo Wong | The Atlantic ($)
“Having efficiently persuaded company America to provide their merchandise a attempt, OpenAI, Anthropic, and Google at the moment are struggling to show that their instruments are definitely worth the cash. …Whereas it’s too quickly to know whether or not GLM-5.2 is absolutely able to changing America’s top-tier AI brokers, any agency or developer who’s balking on the prices now might need an alternate.”
Synthetic Intelligence
Anthropic Discovered a Hidden Area The place Claude Puzzles Over IdeasWill Douglas Heaven | MIT Expertise Assessment ($)
“The J-space accommodates particular person phrases which might be associated to the phrases and phrases that the mannequin is almost certainly to spit out in a response within the close to future. If Claude had been an individual (which it’s not), you would possibly say that these hidden phrases can reveal what’s on its thoughts earlier than it really speaks.”
Synthetic Intelligence
Hackers Can Use 9 of the Most Fashionable AI Instruments to Assemble Large BotnetsDan Goodin | Ars Technica
“A brand new assault the researchers have named HalluSquatting has the potential to assemble large botnets, carry out large-scale DDoSes, and infect units at scale, a primary for prompt-injection assaults. The assault works towards AI coding assistants and brokers, together with Cursor, Cursor CLI, Gemini CLI, Windsurf, GitHub Copilot, Cline, OpenClaw, ZeroClaw, and NanoClaw, that are all inclined.”
Future
UN Secretary-Common Seeks Ban on AI WeaponsTom Chivers | Semafor
“The UN Secretary-Common António Guterres known as for a ban on ‘killer robots,’ saying the choice to take life ‘should stay endlessly human.’ …Guterres known as AI-controlled weapons ‘morally repugnant,’ though not all ethicists agree: One roboticist argues they’re extra discriminate of their killing than scared human troopers, whereas a thinker mentioned in 2022 that utilizing robots will forestall younger women and men bearing ‘the ethical burden’ of wars.”