The Archive

It was a busy summer for AI lobbyists. Anthropic, OpenAI, and Nvidia each had their highest-spending lobbying quarter between July and September, federal disclosures show. Nvidia spent the most of the group at $2.2 million, followed by Anthropic's $2.1 million and OpenAI's $1.5 million. Compared to other Big Tech companies with long-established presences in Washington, the sums aren't out of the ordinary. Microsoft, for example, has spent at least $2 million a quarter going back to 2021, and Google regularly spends over $3 million, including this quarter. But it's a clear signal that the thre...

Multi-token prediction and alternative tokenization enable language models to learn shortest-path graph search at scale, overcoming prior scaling failures.

DIAL enables fine-grained numerical control over instruction-adherence in LLMs via distillation, scaling across diverse instructions efficiently.

Armin Ronacher comments that legacy software debt makes AI systems harder to secure, framing technical debt as a systemic risk.

Hackers were already probing the bug before defenders were told it existed.

Two vulnerabilities found in the programming language PHP put enormous swaths of the internet in danger. By some accounts, up to 70 percent of websites use PHP in some capacity. It powers Wikipedia, WordPress, Drupal, and a host of other popular sites and platforms, including DeviantArt and 4Chan, which were temporarily shut down after the vulnerabilities were disclosed. Even governments weren’t immune. But it’s not as simple as updating a few servers. The vulnerability exists in every version of PHP dating back roughly 20 years, many of which are no longer supported. That means some older si...

Spotify co-CEO Gustav Söderström explains why the rise of AI agents could be good news for the music service.

Google DeepMind releases global forest-type classification map at 10-meter resolution using satellite imagery and ML.

Moshi integrates market data into trader workflows via AI; Anekha Sokhal discusses product-workflow fit over raw capability.

Today, I’m talking to Bill Gates, cofounder and former CEO of Microsoft and chairman of the Gates Foundation. You may have noticed Bill in the news more lately — he’s been on something of a tour for the past few months, ever since publishing an essay in late August titled “The turbulent AI era is here. The choices we make now are critical.” In his piece, Gates ratcheted up his rhetoric around AI safety, writing that “even under the best circumstances, the transition to this new AI era will be one of the most turbulent times in human history.” In interviews since, Gates has called AI “the most...

Claude Duet is available through waitlist for Pro and Max plan users

Nirva will ship its $249 ambient AI pendant by the end of the year.

DeepSeek has reportedly raised $15 billion in one of China's largest funding rounds, just days after the US ordered federal agencies to remove its technology over national security concerns.

1Password is introducing a new feature that allows AI agents to sign into your accounts without actually seeing or handling your passwords. The aptly named 1Password for Agents doesn't give bots free reign to access all of your connected accounts - you'll still need to grant approval for sign-in requests - but should make it less cumbersome to use Agents for tasks that require them to log into various services. "Agents are starting to do real work on our behalf. The old model where you hand over a credential and hope for the best doesn't work," 1Password CEO Aashish Tripathi said in a press s...

T2Quant: task and time-aware mixed-precision quantization for diffusion language models using label-free sensitivity signals.

Inside-JEPA grounds relational channels in latent world models via inverse-dynamics loss without relation labels.

Study of metadata composition effects on LLM-generated museum descriptions reveals field-level grounding challenges.

RaSFNet: ray-aligned fusion network for LiDAR semantic segmentation in autonomous driving and robotics perception.

Evaluation of foundation models for particle track reconstruction and identification in high-energy physics experiments.

Using an AI agent to shop for you is a strange experience. Tech companies have marketed them as personal assistants - the democratizing of labor that ordinarily would only be available to the rich and powerful. In practice, at least for buying things, it's more like sending an anxious and apologetic teenager out to the store: They require an inordinate amount of instructions and specifications, run into technical problems regularly, and come back with solutions that are not exactly what you asked for. They did, technically speaking, complete a task, just not the specific one you asked them to...

Multilingual benchmark of 9 prompting strategies across 4 judges, 27 languages, 13 datasets reveals LLM-as-Judge robustness varies by model.

WholeBodyMem: perceptual memory module for humanoid loco-manipulation via object tracking and cross-attention action expert.

Network topology significantly affects consensus formation speed and stability in LLM agent populations across naming game tasks.

Two-layer calibration framework using conditional flow-matching to control predictive error in computer model simulators.

An investment research firm wanted to prove AI could revolutionize drug development. So they teamed up with Anthropic to see whether an unreleased version of Claude could make scorpion antivenom. It couldn't. It came up with some promising starting points, but nothing close to an actual drug, experts tell The Verge. Citrini Research, the firm behind the experiment, admitted that "none of us could claim to know anything about AI as it related to biology." That didn't stop it from trying to make a drug from scratch - or publishing a blog post titled "We Wanted to Use Claude to Create a New Drug...

Systematic leave-one-subject-out evaluation of wearable IMU systems for construction worker activity recognition reveals deployment gaps.

New benchmark for video-grounded research agents that synthesize multimodal evidence and generate interactive reports linked to video moments.

Study on prompt sensitivity in tool-augmented LLM comparisons, proposing matched protocol to isolate reasoning, tool-calling, and readout steps.

Causal inference approach for prescriptive maintenance in manufacturing using interventional modeling to recommend corrective actions.

CNN feature aggregation method using Fisher Vectors for scene retrieval with improved accuracy on standard image benchmarks.