Category: Tech

home

Baidu’s Open Source OCR Model Parses Hundred-Page Books In One Request: Here Is How SR-SWA Works

Posted on November 1, 2026November 4, 2026 by Clara Milicent

Unlimited OCR 2 is Baidu’s open source 3.3B OCR model that parses hundred-page documents in one request using SR-SWA attention. Benchmarks and limits.

Why Windows 11 Has So Many Web Apps: WinUI’s Missing Basics Are Finally Arriving

Posted on November 1, 2026November 4, 2026 by Clara Milicent

WinUI TableView, Chart and inking controls arrive in Windows App SDK experimental, closing the gaps that pushed Windows 11 apps to the web.

Auto router cost tiers: OpenRouter Auto router picks models by market token spend

OpenRouter’s New Auto Router Picks Models By Where The Market Spends Its Tokens

Posted on November 1, 2026November 1, 2026 by Clara Milicent

The OpenRouter Auto router picks models by where the market spends tokens, across five cost tiers. Benchmarks, costs, session stickiness and the caveats.

A flat vector illustration of a simplified insect-like geometric figure made of thin pale lines with three small amber dots along its body and a thin teal circle enclosing it like a sandbox boundary, representing the Google Mantis security review toolkit for coding agents

Google Open Sourced Mantis: Security Review Skills For Coding Agents, With Warnings Attached

Posted on October 28, 2026October 28, 2026 by Clara Milicent

Mantis is Google open-source security review skills plus an ADK harness that makes coding agents find, reproduce and patch bugs, with hard isolation warnings.

A flat vector illustration of a transparent glass box containing a small grey cylinder, with a thin amber line being scanned by a small teal magnifier above it and a red triangle warning marker in the corner, representing remote code risk when loading models in an AI supply chain

Inspecting A Model Ran Its Code: The Unsloth Studio Finding And The Hardening That Followed

Posted on October 28, 2026October 28, 2026 by Clara Milicent

Pillar Security found model inspection in Unsloth Studio ran repository code. The disclosure, the maintainer dispute, and the remote code hardening that followed.

A flat vector illustration of three simplified screen shapes in a row, a desktop monitor, a browser window and a phone, each with a thin pale teal outline, above a single small amber cube representing one container, illustrating CUA-Lite computer-use agents and sandboxes

Computer-Use Agents Got A Lighter Gym: Inside Berkeley’s CUA-Lite

Posted on October 28, 2026October 28, 2026 by Clara Milicent

CUA-Lite runs computer-use agent tasks in Docker at 0.9 GB instead of a 4.1 GB VM, with unified sandboxes, data and harnesses. What it changes.

A flat vector illustration of two parallel winding paths, one teal and one amber, converging into a single braided line ending in a small rounded badge, representing CopilotKit CLiF combining agent traces with user actions

Your Agent Sees Its Own Mistakes But Not Your Users’ Fixes: What CopilotKit CLiF Captures

Posted on October 28, 2026October 28, 2026 by Clara Milicent

CopilotKit CLiF turns full-stack trajectories into skills, evals and fine-tuning data. How the loop works, what it changes operationally, and its limits.

A flat vector illustration of a large grey speedometer gauge with a thin teal needle near the top and a small amber draft page shape beside it, representing Liquid AI LFM2.5-DSpark speculative decoding draft models for faster decoding

Liquid AI’s DSpark Draft Models Make LFM2.5 Up To 3.2x Faster Without Changing A Single Token

Posted on October 28, 2026October 28, 2026 by Clara Milicent

LFM2.5-DSpark adds exact speculative decoding to Liquid AI models: up to 3.18x faster on GPU, 2.9x on-device, with identical outputs. How it works and where it helps.

A flat vector illustration of one large white magnifying glass with six small thin-line magnifying glasses fanning outward and small amber dots scattered where they point, representing Exa Agent Ultra, the subagent swarm for deep research and entity enrichment

Exa’s Agent Ultra Sends A Swarm Of Subagents At Deep Research, And The Benchmarks Say It Works

Posted on October 28, 2026October 28, 2026 by Clara Milicent

Exa Agent Ultra runs a subagent swarm for deep research and entity enrichment, beating frontier agents on benchmarks at lower cost per task. The caveats.

A flat vector illustration of a simple amber speech bubble containing a white plus and equals sign made of thin bars, with soft teal sound waves radiating outward, representing Kyutai Voice of Reason solving spoken math with speech and reinforcement learning

Kyutai Taught A Speech Model To Do Math Out Loud, And The Method Matters More Than The Score

Posted on October 28, 2026October 28, 2026 by Clara Milicent

Kyutai Voice of Reason lifts spoken GSM8K from 27.3 to 77.1 percent with open speech-to-speech models. The training method and the trade-offs explained.