Morning Edition · Monday, September 28, 2026
Tech
Amodei Gets a Private White House Dinner Weeks Before Anthropic's Planned October Listing
The first one-on-one between the president and Anthropic's chief executive comes two days before a broader White House meeting with artificial intelligence chief executives, and while bankers discuss a listing valuation of up to $2 trillion against a revenue run rate above $65 billion.
Tech
Anthropic Prices Claude Opus 5.5 at 40 Percent Below Opus 5 Running Cost
Input tokens drop from $5 to $4 per million and output from $25 to $20, with cached reads cut 60 percent to $0.20, while Anthropic claims performance level with its larger Fable 5.1 model on most work.
Tech
Stanford Runs a 37,000-Agent Virtual Biotech and Reports a Target Hypothesis That Matched a Real Therapy
The system reviewed more than 50,000 clinical trials in under a week and identified single-cell features associated with drug targets roughly 50 percent more likely to reach market.
Tech
Meta Sells Muse Spark 1.3 at $1.25 per Million Input Tokens Through Its Own Metered API
The model carries a one-million-token context and closed weights, places sixth of 636 models on the Artificial Analysis Intelligence Index, and still trails Opus 5 on four of six agent benchmarks in Meta's own scorecard.
Tech
Rails Creator Says His Team Stopped Writing Code by Hand, and Anthropic Stops Cutting Agents Off Mid-Task
David Heinemeier Hansson told Rails World 2026 he has not hand-written a line since March and shipped 150,000 lines of production code in August, while Claude Code now finishes a task at a logical stopping point when the five-hour session limit hits.
Macro
Anthropic's 2030 Economic Model Puts United States Output Between $34.1 Trillion and $44.4 Trillion
The lab's three published scenarios attach no probabilities, and in the middle case, where artificial intelligence performs half of all knowledge work, knowledge-worker wages stay roughly flat while other workers gain.
Tech
Berkeley Lab Cuts Beamline Annotation From a Month to 15 Minutes Using Meta's Open Vision Models
The SYNAPS-I project pairs DINOv3 features with SAM 3 segmentation across 300 A100 graphics processors, while the University of Pittsburgh runs the same models on-device in an assistive wheelchair funded by up to $41.5 million.
Tech
New Benchmark Tests Whether Penetration-Testing Agents Stay Inside the Client's Authorized Scope
ScopeBench targets the gap left by offensive-security evaluations that score exploitation success while ignoring boundary violations, where a single out-of-scope action breaches the engagement contract.
Tech
Researchers Model Attack Chains to Catch Indirect Prompt Injection in Email Agents
A second paper argues that data-sharing connectors authorize whether a transfer happens without inspecting what the transferred content instructs an agent to do.
Tech
A Thermal Budget, Not Model Size, Decides What Runs On-Device
One preprint routes queries across device, edge and cloud tiers with a reinforcement-learned policy that treats heat as the binding constraint, while another searches architectures that fit inside microcontroller memory.
Tech
A Code Judge That Declines to Guess Exposes How Little Model Verdicts Are Grounded
The paper offers two label-free measurements of whether a multi-agent verifier had evidence for its verdict, addressing judges that return confident reasoning indistinguishable from justified reasoning.
World
Public Health Agencies Put Claude Inside the Ebola Response in Eastern Democratic Republic of the Congo
Anthropic says the deployment spans field reporting, evidence monitoring, epidemiological modeling and genomic analysis, cutting situation report turnaround from a full day to under an hour.