Astronomers are improving the resolution of NASA’s James Webb Space Telescope (JWST) space observatory using a non-redundant Aperture Masking Interferometer. They published their findings in Publications of the Astronomical Society of Australia. The JWST features a non-redundant Aperture Masking Interferometer (AMI) within its Near Infrared Imager and Slitless Spectrograph (NIRISS). This provides magnitudes more precise than…
A journalist created a startup run almost entirely by AI agents
Journalist Evan Ratliff established a startup called HurumoAI run almost entirely by AI agents. Ratliff started it as an experiment to evaluate the capabilities and limitations of agentic AI and documented the experience on his podcast, Shell Game. Ratliff served as the CEO alongside an AI agent co-CEO called Kyle. AI agents also filled the…
Google Cloud’s Shweta Maniar on moving life-sciences AI from research to reality
DeepMind was among the first organizations built from the ground up around artificial intelligence. Founded in 2010 and acquired by Google in 2014, the London-based lab, now known as Google DeepMind, helped catalyze the current AI wave, and two of its researchers, Demis Hassabis and John Jumper, won the 2024 Nobel Prize in Chemistry for…
Biohybrid robots deliver cancer treatments directly to tumors
Scientists from the Robotic Materials Department at the Max Planck Institute for Intelligent Systems in Germany published a paper in Science Advances last week describing biohybrid microrobots that can deliver cancer treatments directly to tumor sites. Medical microrobots have potential for targeted therapeutic delivery. However, current systems achieve only physical targeting, and once at the…
A startup says it found hidden memory behavior in NVIDIA GPUs and is building a security layer around it
Chicago startup Newtonian Standard says it has identified a memory behavior in NVIDIA GPUs that could serve as the foundation for a new kind of hardware-level security. J.P. O’Donnell, the company’s founder and a former Okta engineer, says he spent 18 months characterizing the behavior across NVIDIA Turing, Lovelace and Blackwell GPU architectures. The company…
Study shows LLMs can diagnose ER patients more accurately than physicians
A new study in Science reveals that LLMs can now outperform physicians at diagnosis. The study demonstrated that OpenAI’s o1 model identified the correct or a very close diagnosis in 67% of early ER cases, compared to about 50% to 55% for physicians. The model was required to complete various tasks such as reading medical…
These microrobots can collect nanoplastics from water
Researchers at Brno University of Technology have developed magnetic microbots that can remove nanoplastics from water via electrostatic attraction. They published their findings in Environmental Science: Nano. Nanoplastics, plastic particles smaller than 100 nm, are considered more dangerous than microplastics because their size allows them to cross cell membranes. They have been detected in the…
New Pistoia Alliance survey shows just 1% of professionals report AI’s value in the wet lab
The Pistoia Alliance, a life sciences nonprofit organization, today announced new data showing that only 1% of professionals report AI having value in the wet lab. The data also shows that while 30% of organizations claim to have rolled out enterprise-wide AI, 69% lack metrics to show the impact of AI on reducing costs or…
The startup JustPaid drew inspiration from the sitcom Silicon Valley gag into an AI engineer
Engineers manage agents now. So do scientists. So do the labs building AI. “Engineering roles are becoming management roles where you manage AI agents.” That is how Vinay Pinnaka, co-founder and CTO of the Y Combinator-backed Silicon Valley fintech JustPaid, describes what the past year has done to his job. “I built AI Gilfoyle,” Pinnaka…
UniX AI’s new robot can tackle your household chores
A Chinese robotics company is finally tackling what we all wanted to see robots do: our chores. UniX AI announced earlier this month that its Panther service robot had begun the “first real-home deployment of a mass-produced humanoid robot.” The Panther robot has completed full-stack, continuous multi-task validation in real, unmodified household environments without staging…
The startup QuantHealth claims it can predict how any patient will respond to any therapy, even novel ones
A pharmaceutical company can spend years and hundreds of millions of dollars learning that a Phase 3 trial was aimed at the wrong patients, the wrong endpoint or the wrong competitive bar. The venture-backed startup QuantHealth is trying to force that reckoning earlier. QuantHealth says its AI clinical simulation platform can test trial designs before…
Roblonski platform automates photochemistry with 1,000-fold reduction in reagent use
Researchers from North Carolina State University have developed a compact, automated robotic platform for foundational photochemical assays. They call the system Roblonski, after A. Jablonski, who created the Jablonski diagram. They published their findings in ACS Central Science. Reliable photochemical and photophysical characterization is essential for understanding and optimizing photocatalytic processes. However, traditional spectroscopic methods…
SEQSTER CEO Ardy Arianpour on AWS Bio Discovery, fragmented health records and AI drug discovery’s missing patient layer
Big Tech is sharpening its focus on healthcare and drug development. Amazon’s recent Bio Discovery launch is one recent signal, with AWS pushing into agentic AI application for antibody design. It joins a growing playing field: Google DeepMind’s Isomorphic Labs has advanced pharma partnerships with Novartis and Eli Lilly on the back of AlphaFold 3…
Five key trends that defined the show floor at Interphex 2026
Interphex 2026 brought over 600 exhibitors to the Javits Center in New York City this week. Among the hundreds of vendors, a few themes emerged: AI and automation, digitalization and sustainability. The show indicated that discussions around the digital transformation are turning into action. “I think we’re seeing the transition of the discussion around digital…
How OpenAI’s recently released GPT-5.5 stacks up with Anthropic’s gated Claude Mythos
TL;DR: Claude Mythos Preview appears to lead cleanly on six of nine overlapping rows, especially SWE-bench Pro and Humanity’s Last Exam (HLE), but benchmark comparisons between Mythos and GPT-5.5 are imprecise. The released Opus 4.7 and GPT-5.5 are neck and neck. Opus 4.7 leads on SWE-bench Pro (64.3 vs 58.6), HLE no tools (46.9 vs 41.4)…
Inside the Viz.ai–J&J deal to speed subdural hemorrhage care, from CT scan to treatment decision
Imagine a patient arrives in the ER after a fall. A CT scan already shows blood collecting in the subdural space, but the case still has to move through the usual chain: radiology review, report generation, chart review, specialist notification and treatment triage. Each handoff adds time. The AI care-coordination platform Viz.ai is betting that…
Why the techbio Precigenetics aims to swap cell ‘autopsies’ with live-cell measurement
Suppose the only way to search for life other planets was to vaporize them first, then sift through the debris. “Imagine if you had to literally nuke planets to figure out whether there’s alien life there,” said Parmita Mishra, founder and CEO of Precigenetics, a San Carlos, California-based real-time cellular imaging startup. “We would never…
How a Berkeley team broke 8 major AI benchmarks. Six of them hit 100% without solving a single task
Every couple of months, a new AI model climbs a benchmark leaderboard. Companies cite the numbers in press releases, investors use them to justify valuations, and engineers use them to pick which model to deploy. A new project out of UC Berkeley helps confirm something many had suspected: leaderboards can be less reliable than they…
Merck’s $1 billion Google pact to provide backbone for AI projects
Drugmaker Merck & Co. is committing up to $1 billion over the next decade-plus to make Google Cloud its primary AI anchor, the companies announced Wednesday at Google’s Cloud Next conference in Las Vegas. The deal covers AI infrastructure, engineers, and licensing of Google’s Gemini Enterprise platform, with Google Cloud staff embedded alongside Merck teams.…
AI agents in the R&D workforce: Moving beyond commodity AI
AI is being commoditized in front of our eyes. Months after a frontier release, rivals, often including open source variants, either match it in terms of benchmarks, or come close. Meanwhile, every R&D and IP team now has access to the same enterprise Copilot license, the same Claude and ChatGPT seats, the same general-purpose models…
Cursor is using AI to build the next Cursor while grappling with industrial-scale code creation
On stage at NTT Upgrade, Cursor COO Jordan Topoleski explained that a legion of AI agents built a web browser with more than one million lines of code, from a blank slate state. Kind of. Referring to the project as a “really interesting experiment,” Topoleski said that the goal was to test a long running…
As OpenAI releases GPT-Rosalind for life sciences research, we test out the new AlphaFold and PubMed plugins in Codex
GPT-Rosalind, named for the chemist Rosalind Franklin, is OpenAI’s first life-sciences model. Access is limited as a research preview in ChatGPT, Codex, and the API for qualified customers under OpenAI’s trusted-access program. Framed around evidence synthesis, hypothesis generation, experimental planning, and multi-step research tasks. OpenAI describes it as a “frontier reasoning model” and the first…
After 9,000 announced job cuts, Novo Nordisk taps OpenAI as it hires in China
What Novo Nordisk’s job openings say about its direction Snapshot of 342 open roles on novonordisk.com, April 15, 2026. Top 5 countries China: 96 United States: 75 Denmark: 46 India: 29 France: 28 Top 5 job categories (global) Sales: 111 Commercial Marketing: 29 Engineering & Technical: 23 Education: 17 Clinical Development: 17 China (96 roles…
Anthropic and industry leaders debut Project Glasswing to preempt AI-driven cyberattacks
Anthropic last week announced Project Glasswing, in partnership with Amazon Web Services, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks. Claude Mythos Preview revealed that “AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and…
Anthropic’s oversight trust just hit majority control. The tipping-point was adding Novartis CEO Vas Narasimhan to its board
Anthropic added Novartis CEO Vas Narasimhan to its board on April 14, 2026. With his appointment, directors chosen by the company’s Long-Term Benefit Trust now hold a majority of board seats. Narasimhan is a physician-scientist who has overseen the development and approval of more than 35 novel medicines and vaccines, and he joins the board…























