The daily ledger

MilleMiglia: A realistic instance generator for middle-mile logistics

Algorithms & Theory

Coding Agents Keep Shipping Silent Failures — Here Is How to Catch Them

How to verify your app aligns with your intent without ever reading a line of generated code. The post Coding Agents Keep Shipping Silent Failures — Here Is How to Catch Them appeared first on Towards Data Science.

Who Grades the AI Tests? Epoch Found Problems in Nine Major Benchmarks

Epoch AI’s benchmark reviews found incorrect answer keys, grading defects, and inconsistent evaluation methods across nine tests. Four others passed, with caveats that explain what their scores can actually tell us.

We Pinned Our Model Version to Stay Safe. The Provider Deprecated It Anyway.

The recurring cost of production AI is not inference. It is re-qualification: the eval reruns, prompt retuning, and regression testing you owe every time a model changes under you. Here is what that tax actually covers,…

The OpenAI Hack Shows Why Every AI Connection Needs Limits

Hacktron’s investigation puts a practical question in front of every company deploying connected AI: how much could someone do with one compromised account?

Starting a Career in Data Science in the Age of AI

How do you set yourself up for a career that will stand the test of time when things are changing so fast? The post Starting a Career in Data Science in the Age of AI appeared first on Towards Data Science.

Multi-Agent Coding Isn’t Enough — Agents Need a Commitment Layer

Multi-agent coding systems don't necessarily fail because agents can't communicate. They can fail because important commitments made in conversation have nowhere to live afterward. The post Multi-Agent Coding Isn’t…

Astra Decoded a 1918 WWI Message. The Evidence Spans Archives.

GPT-6 Astra reportedly decoded a November 1918 German radio message with a key associated with December. Separate naval records support the proposed reading and offer a lesson in checking AI-generated research.

😺 OpenAI Cracked an Old Riddle

Anthropic Says Claude Leads 26% of Its AI Research. The Bigger Story Is Who Sets the Metric

Anthropic says Claude now leads 26% of its measured AI R&D. But the bigger development is the measurement system behind that number—one that could eventually shape how governments decide whether frontier AI is moving…

World Labs’ Atlas Can Build 3D Worlds From a Few Images. The Hard Part Is Knowing What It Invented.

World Labs’ Atlas can turn a handful of images into controllable 3D worlds—even filling in areas no camera captured. That makes it powerful for creators, but raises a harder question as generated environments move…

How AI Cracked an 85-Year-Old WWII Enigma Message

A researcher used GPT-6 Astra and specialist AI agents to recover an 82-letter WWII Enigma message. The investigation published its code and evidence so others could check the result.

Atria Dawn: The Dawn of Agentic Superintelligence

arXiv:2609.15818v2 Announce Type: replace Abstract: As AI agents become participants in the development of their successors, they reshape both the production of intelligence and the role of human researchers. We…

RAFT: A Stateful Retrieval-Augmented Framework for Troubleshooting Agents

arXiv:2609.20754v1 Announce Type: new Abstract: Effective troubleshooting agents in enterprise customer support depend on retrieving actionable guidance from similar historical cases, yet existing retrieval-augmented…

PointEvent: Rethinking Event-based Tiny Object Detection via Serialized Motion Evidence Accumulation

arXiv:2609.20066v1 Announce Type: cross Abstract: Event cameras offer high temporal resolution and motion sensitivity for tiny UAV detection, yet distant targets generate sparse and fragmented events that are easily…

Dreaming the Sound of Contact: Leveraging Video and Audio Generation for Zero-Shot Force-Aware Manipulation and Data Generation

arXiv:2609.19137v2 Announce Type: replace-cross Abstract: Video generation models have advanced rapidly and can now synthesize plausible videos of robot manipulation from image and text prompts. Recent work extracts…

Accelerating Sharded Data Parallelism at Scale with Federated Learning

arXiv:2609.20359v1 Announce Type: cross Abstract: The symbiotic scaling of artificial intelligence models and high-performance computing systems continually creates algorithmic challenges in their convergence.…

Inference-Engine Fingerprinting Attacks are Practical: Exploring Model-Driven Environmental Discovery, Exploitation, and Escape

arXiv:2609.20614v1 Announce Type: cross Abstract: Frontier AI models are rapidly gaining the ability to exploit vulnerabilities in complex pieces of software. The risk is not theoretical, as evidenced by recent sandbox…

HIL-UMI: Bringing Human-in-the-Loop Post-Training of Vision-Language-Action Models to Universal Manipulation Interface

arXiv:2609.20659v1 Announce Type: cross Abstract: Large-scale vision-language-action (VLA) models provide powerful priors for robot manipulation, yet adapting them to a specific deployment remains challenging.…

MaSCoD: A Multi-Agent Framework for Structural-Context-Guided Candidate Causal Graph Generation

arXiv:2609.19944v1 Announce Type: new Abstract: Large language models (LLMs) have been applied to causal discovery, but candidate-graph generation rarely treats premature omission of potentially relevant causal…