HOT
Advertise

Research

Latest News

2 min read Editorial composition with names and marks from the official websites of Microsoft 703
Research

Microsoft’s AI Beat Generalist Doctors on Written Tests. Patient Trials Are Next

It scored 83.2% on complex cases, but the study did not test treatment decisions or patient outcomes.

2 min read A fictional blonde adult scientist examines a sample plate beside a laboratory robot. 792
Research

Google’s AI Agents Raised a Lab Hit Rate to 42.1%, but the Work Is Early

Google’s AI agents raised an aptamer hit rate to 42.1% across three robotic lab rounds, but the result is research, not a treatment.

2 min read A researcher prepares bacterial samples for tests of a possible antimicrobial peptide. 931
Research

AI Flags a 97-Amino-Acid Antimicrobial Candidate. Lab Tests Come Next

The bacteriocin has computational and genomic clues behind it, but only laboratory tests can establish whether it works against bacteria and is safe.

2 min read A researcher studies printed number sequences beside an open laptop. 878
Research

Claude Traced Five NIST Curve Seeds to a 1990s Random Generator

Anthropic says Claude matched five NIST curve seeds to timestamp-seeded generator output. The finding eases one concern, but leaves other design choices unexplained.

2 min read Physicists study particle-collision visualizations at computer workstations. 643
Research

CERN and Mistral Plan AI for Billions of Simulated Collisions

Their proposed physics model could support several analysis tasks, but its budget, schedule and eventual performance remain unknown.

2 min read A farmer checks crop conditions in a field beneath an overcast sky. 525
Research

AI Weather Correction Doubled Forecast Skill, but Not Local Certainty

A Nature study reports stronger forecasts two to six weeks ahead. The gains improve odds of broad conditions, not promises about a town’s weather.

2 min read A technician inspects several small TV boxes connected on a workbench. 591
Research

Anthropic Tested AI on 15 TV Boxes. It Did Not Build a Worm

The experiment reached 0.39 tokens per second, but left a crucial gap between running AI and proving a self-spreading infection.

2 min read Researchers compare machine-learning experiment results at a workstation. 693
Research

GPT-5.6 Sol’s Reported Gains Fell Under Stricter Scoring, Epoch AI Says

Epoch AI’s October 7 evaluation put Sol at 15% of a human method’s gains under strict scoring, exposing how selected runs complicate comparisons.

3 min read Public-health researchers discuss geographic data visualizations beside printed maps. 580
Research

Google Earth AI Shows Promise, but Its Health Forecasts Still Vary

Five studies across four countries found gains in select tasks, but no evidence yet of a live public-health warning system.

2 min read An adult learner works through a fraction worksheet beside an open laptop. 608
Research

AI Lifted Test Performance, Then Participants Struggled Without It

In short online tests, AI users answered fewer follow-up questions correctly and skipped more. The researchers did not measure lasting effects.

2 min read A specialised AI accelerator chip mounted on a circuit board. 527
Research

Could AI Chips Make a 10-Year Training Pause Enforceable?

A 26-researcher paper says inference-only chips could help verify a decade-long pause, but only if rival governments agree on rules and inspections.

2 min read A mathematician studies equations beside a laptop displaying a proof-checking interface. 695
Research

OpenAI’s 722 AI Math Manuscripts Are Public, but Many Proofs Still Need Checking

The manuscripts trace back to about 4,000 problems. OpenAI says formal checking is incomplete, so 722 papers does not mean 722 verified breakthroughs.