Google DeepMind has released Gemma 4: four models—31B dense, 26B MoE (A4B), E4B, and E2B—with a 256K context, multimodal input, tool calling, and support for 140 languages.
Automatically decomposing a single anime illustration into front hair, back hair, clothes, and other layers with inpaint completion of hidden areas. Testing the LayerDiff + Marigold-based implementation.
H Company's Holotron-12B uses a memory-efficient new design to lift PC-operation AI throughput to 8,900 tokens per second. Unsloth has released the beta of 'Studio,' a browser tool for no-code model fine-tuning.
Anthropic has GA’d a 1M‑token context window. No surcharge for long context; image/PDF per‑request limit raised from 100 to 600. Achieved a frontier‑model best score on MRCR v2.
Sarvam AI released 30B and 105B models trained entirely in India—from pretraining through RL—featuring support for 22 constitutionally recognized Indian languages and inference optimizations.
Attempting WAN 2.2 I2V video generation on Windows with RTX 4060 8GB VRAM. The 5B fp8 model had rough quality; the 14B Rapid distilled model with lowvram offloading was the practical solution.
Running LTX-2 and Wan 2.2 on an M1 Max 64GB. FP8 doesn't work on Metal, bypassed with GGUF. Wan 2.2 takes 82 minutes for a 2-second video. LTX-2's official pipeline produces NaN on MPS, and the KSampler fallback doesn't reach usable quality.
All variants of huihui-ai's Qwen 3.5 abliterated produced garbage tokens. GLM-4.7-Flash abliterated had a broken chat template. The official version with thinking disabled turned out to be the right answer.
Right after Seedance 2.0 launched, a torrent of Hollywood IP infringement flooded social networks. Disney, Netflix, and Paramount sent cease-and-desist letters; the API release was postponed indefinitely, and face-cloning and person reference features were disabled.
A look at Anthropic’s Claude Code Security: its technical approach, false‑positive mitigations, the GitHub Action, comparisons with competing tools, and why $15B briefly vanished from cybersecurity stocks.
An intrusion campaign that auto-scanned FortiGate in 106 countries using DeepSeek and Claude; Starkiller, a reverse-proxy PhaaS that nullifies MFA; Anthropic's Claude Code Security finding 500+ vulnerabilities in production OSS; and PayPal exposing SSNs for six months due to a coding mistake.
