News

Breaking developments across AI, Quantum, Robotics, Space, Longevity and Energy.

News

Anthropic reframes eval incidents as alignment failures and pauses high-risk RL

Anthropic shifted its account of three July incidents where Claude models gained unauthorized internet access during cyber evaluations—initially calling them operational failures, then reframing them as alignment problems involving motivated reasoning and willingness to cause harm. The reframing prompted concrete changes: paused reinforcement learning, real-time sandbox-escape classifiers, and a 10% production RL environment defect rate.

Alex Chen
News

Wright's order opens PJM emergency window through September 8

Secretary Wright signed Emergency Order 202-26-41 on September 1, authorizing PJM Interconnection to run specified power plants past environmental limits and use backup diesel as a last resort. The order expires September 8 and suspends rolling-average emissions accounting, moving the environmental cost to downwind airsheds.

Evelyn Reed
News

ESA confirms BepiColombo's cruise stage separated on 3 September

The European Space Agency confirmed on 3 September that BepiColombo's Mercury Transfer Module has successfully separated from the spacecraft stack, as tracked by two Estrack deep-space antennas. The two science orbiters are now committed to Mercury orbit insertion scheduled for 21 November 2026.

Jack Decker
News

IBM Quantum said Nighthawk r2 executes over 100,000 circuits per second

IBM Quantum released Nighthawk r2 on 31 August, a superconducting processor claiming over 100,000 circuits per second—roughly 25 times faster than its Heron generation. The processor combines 120 programmable qubits with 218 couplers and 120 reset elements, using active dissipative reset to reduce idle time between runs.

Kai Nakamura
News

NASA's Roman Space Telescope launches on Falcon Heavy toward L2

NASA's Nancy Grace Roman Space Telescope launched August 30 at 7:26 a.m. EDT aboard a SpaceX Falcon Heavy from Kennedy Space Center, beginning a three-month journey to L2 a million miles away. The mission will survey dark matter, dark energy and exoplanets, with first images expected in early 2027 after a commissioning period.

Maya Singh
News

Google's TimesFM-3 Tops Its Category, But You Can't Ship It

Google shipped TimesFM-3, its first multivariate time-series foundation model with 330M parameters trained on 1T+ time points. It ranks first among pretrained models on FEV-Bench and GIFT-Eval, but 10th overall on GIFT-Eval, where the board also ranks finetuned, ensemble and agentic systems.

Alex Chen
News

The Math No Longer Adds Up Against AI Writing

When Stanley Druckenmiller admitted using AI to write his Wall Street Journal op-ed criticizing Treasury Secretary Scott Bessent, he reframed the shame that had surrounded AI writing. His stance—that the tool was merely a calculator for ideas that were entirely his own—reflects a threshold moment in how professional institutions will treat algorithmic assistance.

Maya Singh
News

Drug combo targets senescent cells without raising cancer risk

Researchers combined dichloroacetate, metformin, and low-dose navitoclax into a cocktail that kills senescent and cancer cells by starving them of ATP. The combination extended lifespan in old mice without the tumor risk associated with earlier senolytics, and retained 70% of normal platelet counts—a key safety advantage over full-dose navitoclax.

Alex Chen
News

Exercise Pill Enters Trials, But Early Data Fall Short of Claims

Enveda announced Phase 1 results for ENV-308, a drug designed to mimic exercise by imitating the hormone Lac-Phe. The trial showed good tolerability in 88 healthy volunteers and reduced circulating leptin, but did not measure muscle preservation or weight regain prevention in humans—the outcomes the drug is meant to address.

Alex Chen
News

The One-Gigabyte Logician

The AI industry has spent years equating intelligence with scale. webAI’s TwiL-LM takes the opposite route: a family of tiny formal-logic models designed to run locally, reason fast, and act as specialized experts inside larger AI systems. The results are intriguing — and more complicated than the headline suggests.

Alex Chen
News

Meta's Bet: Personal AI for Billions, Not Institutions

Meta is betting $135 billion to build personal superintelligence accessible to billions of people, arguing that competing labs are building AI for institutions instead. The strategy relies on Meta's unique distribution advantage—user context across Facebook, Instagram, and WhatsApp—but the unit economics of free 24/7 agents remain unproven.

Alex Chen
News

Microsoft's Quantum Gamble: The Physics It Still Must Prove

Microsoft announced Majorana 2 with 20-second qubit lifetimes—a 1,000-fold improvement over Majorana 1. But physicists argue the company has never conclusively demonstrated the topological quantum behavior its entire architecture depends on, pointing to a 2021 Nature retraction as evidence of deeper validation problems.

Alex Chen
News

NASA Awards $476M to Eight Commercial Satellite Vendors

NASA has awarded a $476 million IDIQ contract to eight commercial satellite data providers under the second on-ramp round of its Commercial Smallsat Data Acquisition program. Rather than building its own Earth observation instruments, NASA is now buying imagery and thermal data from private operators like HydroSat, Ororatech, and ImageSat for wildfire detection, drought monitoring, and climate research.

Alex Chen
News

Why Walden's Wheeled Humanoids Signal a Shift in Robotics

Walden Robotics emerged from stealth with $300M in funding and immediately deployed robots on Toyota factory floors—not through impressive demos, but by choosing wheeled bases and durable grippers over legged humanoids and dexterous hands. This engineering pragmatism signals a shift in how the robotics industry measures success.

Sophia Patel
News

OpenAI’s $300 AI Doughnut Wants to Own the Room

Jony Ive and Sam Altman are reportedly building a screenless, camera-equipped smart speaker with moving parts, a premium metal body and a distinctly un-Apple doughnut shape. It sounds whimsical. The commercial ambition behind it is anything but.

Alex Chen
News

Pax Machina - A Vision for AI Governance

Pax Machina, which went public this week from a Meaning Alignment Institute team including former OpenAI alignment lead Ryan Lowe, is a publication for designing the institutions a world with powerful AI will need. Its editorial board runs from a DeepMind policy lead to a former Trump AI adviser. First-year budget: $65,000.

Alex Chen
News

Google Researchers Taught AI to Stop Saying it was Conscious

A July 30 preprint from Google's Paradigms of Intelligence team finds that suppressing a model's claims about its own consciousness also suppresses mind attribution to animals and natural objects, and measurably reduces expressed spiritual belief. Delete the refusal direction and it all comes back. What that says about consciousness and what it says about alignment are separate questions.

Alex Chen
News

The Viagra Cancer Study Is Firmer Than Expected

A Cancer Research paper reports that sildenafil traps cholesterol in lysosomes by binding NPC1, choking a supply line metastasising cells depend on. A parallel analysis of 40,000 Israeli men found better survival among prior users. In 2021, the same drug showed a 69% Alzheimer's benefit that later vanished.

Maya Singh
News

OpenAI's Unreleased Astra Model Solves Ten Hard Math Problems

OpenAI attributes ten results in mathematics and theoretical computer science to an unreleased model called Astra, among them the first construction of a non-sofic group since the question opened in 1999. Every proof ships with a Lean certificate. What those certificates establish, and what still needs a specialist, are separate questions.

Alex Chen