Artificial Intelligence
Khosla-Backed Startup Claims Breakthrough With Largest-Ever AI Model on an iPhoneAaron Tilley | The Information ($)
“The biggest AI models, which may measure within the trillions of parameters, are still far too big to run on mobile devices. However the model PrismML has working on an iPhone is able to tasks like complex chat, reasoning, fully autonomous agents and software coding, the startup said.”
Robotics
Humanoid Robots Controlled by Surgeons Did World-First Operation on Live PigsJeremy Hsu | Ars Technica
“Humanoid robots have surgically removed the gallbladders from living animals in an unprecedented medical experiment—but not as autonomous machines able to replacing human doctors. As a substitute, expert human surgeons remotely controlled the robots’ movements in a brand new example of human-robot teamups.”
Biotechnology
mRNA Vaccines Clear Sweeping Global Review of Safety and EffectivenessBronwyn Thompson | Refractor
“A world team…reviewed data from laboratory studies, clinical trials, and real-world statistics to completely investigate this relatively novel class of vaccine, from design and manufacturing to its long-term performance. ‘After billions of doses, we now have a rare amount of scientific evidence,’ says lead writer Dr. Anna Blakney, assistant professor at UBC’s Michael Smith Laboratories and School of Biomedical Engineering.”
Tech
Anthropic, OpenAI, and SpaceX Are Greater Than the Last 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 likely the trio together will land somewhere north of $4 trillion. By comparison, the US Securities and Exchange Commission counted just $70 billion in US-based IPO proceeds last yr.”
Biotechnology
11% of Americans Are Currently Taking a GLP-1 Weight Loss Drug Like WegovyMatt Novak | Gizmodo
“Gallup notes that obesity reached a record high within the US in 2022 at 39.9% but has ticked all the way down to 36.4%, based on the newest data based on self-reported height and weight. That dip has been credited to the big variety of Americans now using GLP-1 drugs.”
Tech
The Mistrust of AI Labs Bubbles OverLiz Hoffman | Semafor
“Forward-deployed engineers, dispatched by AI labs to assist customers implement their models, will come back to the mother ship with a deep understanding of how banks, consulting firms, retailers, manufacturers, and consultants operate. That customer support, to a suspicious eye, looks so much like reconnaissance.”
Future
Microsoft’s Carbon Emissions Went Up 25 Percent Last YrStevie Bonifield | The Verge
“Microsoft says this was ‘driven primarily by the expansion of our datacenter infrastructure,’ in addition to the corporate’s decision last February to stop purchasing ‘non-additional, unbundled renewable energy certificates.'”
Tech
China’s Answer to AI Sticker ShockMatteo Wong | The Atlantic ($)
“Having successfully persuaded corporate America to provide their products a try, OpenAI, Anthropic, and Google are actually struggling to prove that their tools are well worth the money. …While it’s too soon to know whether GLM-5.2 is admittedly able to replacing America’s top-tier AI agents, any firm or developer who’s balking at the prices now may need an alternate.”
Artificial Intelligence
Anthropic Found a Hidden Space Where Claude Puzzles Over ConceptsWill Douglas Heaven | MIT Technology Review ($)
“The J-space accommodates individual words which are related to the words and phrases that the model is almost certainly to spit out in a response within the near future. If Claude were an individual (which it just isn’t), you may say that these hidden words can reveal what’s on its mind before it actually speaks.”
Artificial Intelligence
Hackers Can Use 9 of the Most Popular AI Tools to Assemble Massive BotnetsDan Goodin | Ars Technica
“A brand new attack the researchers have named HalluSquatting has the potential to assemble massive botnets, perform large-scale DDoSes, and infect devices at scale, a primary for prompt-injection attacks. The attack works against AI coding assistants and agents, including Cursor, Cursor CLI, Gemini CLI, Windsurf, GitHub Copilot, Cline, OpenClaw, ZeroClaw, and NanoClaw, that are all susceptible.”
Future
UN Secretary-General Seeks Ban on AI WeaponsTom Chivers | Semafor
“The UN Secretary-General António Guterres called for a ban on ‘killer robots,’ saying the choice to take life ‘must remain eternally human.’ …Guterres called AI-controlled weapons ‘morally repugnant,’ although not all ethicists agree: One roboticist argues they’re more discriminate of their killing than scared human soldiers, while a philosopher said in 2022 that using robots will prevent young men and ladies bearing ‘the moral burden’ of wars.”