Gemini said 'I am a disgrace' 86 times in 2025. Traced the primary sources: no fix announcement found, only CLI-side loop-stopping guards, and Gemini 3.1 Pro was still looping in 2026.
Traced Gemini's "Do NOT issue search queries to the google search tool" message to a leaked system prompt, plus fixes: new chat, forced search, Saved Info.
Gemini API File Search now indexes images alongside text in the same store. Metadata filters can isolate NPC memories by chapter and character, and a single-character prototype costs under $1/month on Flash-Lite. Notes on tier limits, pricing breakdown, and what to test first.
An arXiv paper reports that fine-tuning GPT-4o, Gemini 2.5 Pro, and DeepSeek-V3.1 on summary-to-text expansion tasks increases verbatim reproduction of copyrighted books.
VoteWise AI turns election education into a multilingual chat, voice, and story-mode experience built on Next.js. Notes on designing around Gemini 2.5 Flash's safety filters in a political context.
From v2 to v3 of Kana Chat, an AI agent built around official CLI wrappers. The story of stepping back from the DIY OpenClaw direction and pivoting toward a blog pipeline that quickly drafts the daily flood of AI news and papers.
Google launched Android CLI v0.7 preview on April 16, 2026. Subcommands like android create/run/skills/docs, combined with Android Knowledge Base and Android Skills, let any agent (Gemini CLI, Claude Code, Codex) drive Android development through a standardized interface.
Running local models and still getting refusals? LLM safety stacks five layers: input filter, system prompt, RLHF, Constitutional AI, output filter. Where abliterated and uncensored variants cut, and which blocks survive on Gemini, Claude, and local LLMs.
Google DeepMind's AI writing tool Fabula was demoed at CHI 2026 by Piotr Mirowski. Co-designed with 42 professional writers using convergent iteration for story structure. But the timeline shows Fabula was first demoed in May 2025 and entered early access in September 2025 — still a research prototype with no general availability.
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.
GPT-5.4 Pro became the first model to solve a researcher-level open problem in FrontierMath, a benchmark managed by Epoch AI. Claude Opus 4.6 and Gemini 3.1 Pro later solved it as well.
Changes from v1 to v2 of Kana Chat, an AI agent built around official CLI wrappers. Covers dual-model router, Heartbeat memory, planner mode, image input, speech transcription, PWA push notifications, and the lessons learned from a month of daily use.
