The Archive

Sparse autoencoders reveal LLMs isolate time-aware vs. look-ahead-biased features; feature steering improves cross-domain forecasting generalization.

Process harness wraps legacy workflows with agentic reasoning layer via Task-Decision-Flow model, enabling Agentic BPM without engine replacement.

HarmVideoBench evaluates LVLMs on multi-layered harmful video detection beyond binary classification, requiring explicit explanatory rationales.

VLM-PBRS uses vision-language model guidance to learn potential-shaping heuristics for sparse-reward RL, avoiding naive reward hacking.

BERT-based multi-task framework for FDA medical device recall triage, severity assessment, and root-cause analysis on 54K historical records.

Model-assisted sampling framework reduces variance in stochastic gradient estimation for deep learning without excessive computational overhead.

Vision-language-action policy with RL achieves 1st place (online) and 2nd place (offline) in LeHome Challenge 2026 bimanual garment folding task.

TOPS framework prunes visual tokens in MLLMs via principled optimization, balancing instruction-relevance and diversity for faster inference.

OpenRCA 2.0 introduces PAVE labeling protocol with causal propagation paths to evaluate LLM agent reasoning, long-context understanding, and tool use.

DMuon enables efficient distributed training of matrix-orthogonalization optimizers with overhead near Adam, improving convergence on modern architectures.

Self-improving iterative codebook refinement for autoregressive image generation improves safety and text-conditional quality in unified multimodal models.

fTNN tensor neural network for fractional PDEs using geometry-adapted integration split on bounded domains.

Joint learning framework combines natural-language experiential rules and parameter updates to improve LLM agent policies in sparse-reward multi-step environments.

Empirical comparison of Kolmogorov Arnold Networks against MLPs and GNNs for aerodynamic prediction tasks.

Netris provides software that runs on network switches, and offers a platform that helps neocloud operators reduce the time it takes to go live.

Artificial intelligence is rapidly reshaping retail, but not in the ways consumers might immediately notice. The biggest transformation may not be flashy virtual try-ons or chatbot shopping assistants, but in how decisions are made behind the scenes: how products surface in search results, how inventory moves through supply chains, how engineers ship code faster, and…

Adobe said that it will integrate Topaz Labs' tools across its apps.

Amazon’s latest India investment comes as global tech companies race to expand AI infrastructure in the country.

To celebrate its new status as No. 1 in JD Power's initial quality ranking among mainstream automakers, Ford is opening up about the challenges it has faced in recent years, especially around its reliance on automated systems in production and design. It turns out that those automated systems were not as robust as previously assumed, requiring Ford to hire experienced technicians - sometimes bringing back former employees - to correct errors made by the company's robots. In Ford's view, AI is both powerful and prone to pitfalls. Its effectiveness depends entirely on the quality of the data us...

OpenAI explores how organizational structure and decision-making speed limit AI adoption in enterprises, framing latency as a critical bottleneck.

IBM’s nanostack transistors could boost chip performance or energy efficiency.

An interview with Figma CEO Dylan Field about building Figma, and why he believes AI gives the company a tailwind.

This is what the new Creator Studio app looks like, after the original service was shuttered in 2023. | Image: Meta Meta is bringing back the Facebook Creator Studio page manager, now "reimagined" as a standalone AI companion app. The new app aims to make it easier for creators to connect with their audiences, and show them "exactly how to grow on Facebook," according to Meta's announcement. Meta's AI Creator Assistant is a central focus of the newly reimagined app - users can ask the chatbot to provide performance tracking insights and tailored recommendations for improving engagement. It ca...

Move over, Harness Engineering, it is time for the harness of harnesses!

A new OpenAI research paper shows how AI agents are transforming work, enabling longer, more complex tasks and expanding productivity across roles.

As ASML CEO Christophe Fouquet told TechCrunch in May, what China can currently buy are older-generation deep ultraviolet tools — gear first shipped about a decade ago — the same machines the MATCH Act would now put off limits.

Learn how the Cohere team uses North, Wiz, and a custom MCP server to automate incident response workflows with AI.

AI agents automate vLLM fork maintenance, cutting upstream sync time from weeks to days by auto-rebasing, testing, and fixing conflicts.

simonw/browser-compat-db Inspired by Mozilla's new MDN MCP service - source code here - I decided to try converting their comprehensive mdn/browser-compat-data repository full of browser compatibility data into a SQLite database. This new GitHub Repo includes a Claude Code for web (Opus 4.8) generated script for doing that using sqlite-utils . I wanted the resulting ~66MB SQLite database to be available via the GitHub CDN with open CORS headers. GitHub releases don't have those, but any file stored in a regular GitHub repository does - so I had Codex Desktop (GPT-5.5) build a GitHub Actions w...

Backed by Mayfield and Aramco Ventures, Vishal Sikka’s new venture brings together veterans from SAP, Infosys, and VianAI.