Zotpaper

Pentagon Puts $29 Billion Price Tag on Iran Conflict as Hegseth Signals Flexibility on Escalation

Defense Secretary tells Congress the US is prepared to either intensify or wind down military engagement with Iran

Defense Secretary Pete Hegseth appeared before a congressional panel on Monday to present a $29 billion cost estimate for the ongoing US military conflict with Iran, while downplaying concerns about munitions shortfalls and stating the United States remains prepared to either escalate or draw down the war depending on circumstances.

Zotpaper·13 May·3 min

Read Full Account

Geopolitics

Israel Says It Killed Iran Security Chief Ali Larijani in Overnight Strikes

Israel's defence minister announced that Ali Larijani, Iran's national security chief and one of the regime's most powerful figures, was killed in overnight airstrikes along with commander Gholamreza Soleimani. Iran has not yet confirmed the claim, but if true it would represent the highest-profile assassination since Supreme Leader Ali Khamenei was killed on the first day of the war.

Geopolitics

Iran War Day 16: Trump Threatens to Strike Kharg Island Again Just for Fun as UK Readies Warships for Hormuz

As the US-Israel war on Iran enters its third week, President Trump declared the US may strike Iran's critical Kharg Island oil hub again "just for fun" while urging allies to deploy warships to reopen the Strait of Hormuz, a remark that drew immediate international alarm.

Tech

Silicon Valley's Pivot to the Pentagon: How Tech Giants Are Becoming Defence Contractors

A growing number of Silicon Valley's most prominent technology companies — including Palantir, Anduril and Google — are expanding aggressively into defence contracting, selling AI-powered and computer-guided weapons systems to governments, raising significant questions about the ethics, accountability and long-term consequences of merging consumer-grade artificial intelligence with lethal military capability.

Researchers Push Toward Autonomous AI Systems That Handle Data, Identity, and Imbalanced Learning

Exaforce Raises $125M to Deploy AI Against Real-Time Cyberattacks

AI

Silicon Valley's Pivot to the Pentagon: How Tech Giants Are Becoming Defence Contractors

Researchers Push Toward Autonomous AI Systems That Handle Data, Identity, and Imbalanced Learning

Medicare's New Payment Model Opens Door for AI-Powered Patient Care

Business

Elliptic Raises $120M Series D as Deutsche Bank and Nasdaq Ventures Deepen Crypto Compliance Bets

Blockchain analytics company Elliptic has closed a $120 million Series D funding round backed by Deutsche Bank and Nasdaq Ventures, valuing the firm at $670 million, as major traditional financial institutions continue to expand their exposure to cryptocurrency infrastructure.

CFTC Moves to Assert Federal Authority Over Sports Prediction Markets

Bitcoin's Current Downturn Shallower Than Past Bear Markets, But Risks Remain

Latest Research

All research →

Paper · 13 Apr

Physics-grounded simulation matches real training data for cloth manipulation tasks

Yunsong Zhou et al.

Zhou et al. present SIM1, a data engine that converts a small number of real demonstrations into large-scale synthetic training data for robotic manipulation of deformable objects such as cloth. The system digitizes real scenes into metric-accurate virtual twins, calibrates soft-body physics via elastic modeling, and generates diverse trajectories through a diffusion model with quality filtering. Policies trained exclusively on this synthetic data match those trained on real data at a 1:15 equivalence ratio — one real demonstration is worth roughly fifteen synthetic ones, or equivalently, fifteen synthetic examples substitute for one real one.

Paper · 13 Apr

Training-free framework improves object counting in text-to-video generation

Zhengyang Sun et al.

Text-to-video diffusion models frequently generate the wrong number of objects when given a numeric prompt, such as 'three dogs' producing two or four. The authors introduce NUMINA, a training-free framework that detects count mismatches during generation and corrects them by refining the spatial layout derived from attention maps. Applied to Wan2.1 models of three different sizes, NUMINA improves counting accuracy by up to 7.4 percentage points with no additional training.

Paper · 13 Apr

Neural Network Decoder Unlocks a Steep Error-Suppression Regime in Quantum LDPC Codes

Andi Gu et al.

Gu et al. introduce a structure-aware convolutional neural network decoder for quantum error-correcting codes that matches the geometric layout of the code. Applied to the [144, 12, 12] Gross code, it reveals a previously hidden "waterfall" regime of steep error suppression, reaching logical error rates of ~10⁻¹⁰ at 0.1% physical error rate — with latencies compatible with real-time operation on current quantum hardware platforms.