A research project reverse-engineered Google DeepMind's SynthID image watermark using FFT-based spectral analysis. The V3 bypass achieves 91% phase removal while maintaining SSIM 0.997. Is removing an invisible watermark copyright infringement? Analysis from DMCA, EU AI Act, and Japanese law perspectives.
Sentence Transformers v5.4 adds multimodal support. Eight embedding models and four rerankers including Qwen3-VL and NVIDIA Nemotron can now be used through a unified API.
The full picture of Operation Masquerade as disclosed by the FBI and NCSC. APT28 tampered with SOHO router DNS to intercept Outlook auth tokens — the techniques and countermeasures.
Union types land in C# 15 with .NET 11 Preview 2. A look at the design, exhaustive pattern matching, internal implementation, comparison with unions in other languages, and practical use in Blazor and Minimal API.
An Adobe Reader/Acrobat zero-day actively exploited since November 2025. A two-bug chain achieves sandbox bypass and RCE, affecting all versions including the latest. No patch available.
MegaTrain flips the GPU-centric paradigm by treating CPU memory as primary storage and the GPU as a transient compute device, enabling full-precision training of 100B+ LLMs on a single GPU with up to 12.2x throughput over DeepSpeed ZeRO-3.
Meta unveils Muse Spark, the first model from its new Meta Superintelligence Labs. The Scale AI acquisition, the shift from open-weight to proprietary, multi-agent reasoning via Contemplating mode, and the evaluation awareness problem.
Anthropic's unreleased Claude Mythos Preview discovered thousands of zero-day vulnerabilities including a 27-year OpenBSD bug and a 16-year FFmpeg bug. Deemed too dangerous for public release, it ships exclusively through Project Glasswing to 12 founding partners.
Zhipu AI's GLM-5.1 is a 744B MoE (40B active, 200K context, MIT) targeting long-horizon agent tasks. Hits 58.4% SOTA on SWE-Bench Pro (edging out GPT-5.4 and Claude Opus 4.6) and sustains performance across 8-hour sessions with 6,000+ tool calls without degradation.
9 Japanese-specialized LLMs as of April 2026 — LLM-jp-4 (11.7T tokens from scratch), PLaMo, Nemotron Nano 9B JP (#1 sub-10B on Nejumi 4), Swallow 30B-A3B, Namazu — broken down by whether they were scratch-trained, continued pre-trained, or post-trained, with size, license, benchmark scores.
WordPress staple plugin ACF 6.8 adds Abilities API integration, automatic Schema.org structured data, and WP-CLI commands. How AI agents can now discover and manipulate WordPress content models.
