• Motivation for the Switch:

  • Growing frustration with Duolingo's excessive gamification, recent course overhauls, and the restrictive energy system for free users prompted the search for a flexible alternative.

• Setting Up a Custom Gemini Tutor:

  • Created a dedicated custom Gemini Gem instructed to act as a personal Spanish tutor rather than using generic pre-made templates.
  • Prompted the AI to first evaluate the author's current proficiency across grammar, vocabulary, reading, and conversation before building a tailored, progressive study plan.

• Strengths Over Duolingo:

  • Personalized & Dynamic Learning: Functions like an interactive workbook overseen by a patient tutor, explaining exact mistakes and adapting future lessons based on performance.
  • Deeper Focus & Understanding: Teaches brief concepts prior to exercises rather than throwing users straight into repetitive matching puzzles, requiring real engagement instead of mindless tapping.

• Limitations & Areas for Improvement:

  • Audio & Listening Gaps: Lacks the seamless built-in listening comprehension tools of dedicated language apps, making audio-first drills clunky.
  • UI/UX Friction: Minor interface issues (such as text overflow requiring horizontal scrolling) show it is not natively designed from the ground up as a specialized language app.

• Verdict:

  • Gemini serves as a highly capable, adaptive, and customizable language tutor that easily outperforms Duolingo in depth and flexibility, provided the user is willing to manage the lack of native audio drills.

Gemini成为Google历史上增长最快的产品，月活10亿——超越Gmail、YouTube、Search达到这个里程碑的速度。这验证了AI助手不是利基产品，而是主流基础设施。现在的竞争问题变成：谁能在10亿用户规模上保持质量差异化？

The explicit inclusion of 'use tools' alongside core cognitive functions like 'perceive' and 'reason' highlights a significant architectural focus on embodied AI. This suggests the model is being designed with a direct path to physical agency, a non-obvious but critical distinction.

This phrasing suggests a level of creative sovereignty not typically claimed by AI models. It implies a fundamental shift from content generation to content creation, suggesting a more autonomous and less tool-dependent creative process.

This demonstrates the model's capacity for complex, structured problem-solving. To apply this, frame your prompts around a specific problem, provide all necessary constraints and requirements, and ask the model to design a step-by-step solution or algorithm.

This suggests a focus on complex, multi-step reasoning and tool use. To apply this, structure your prompts as a sequence of tasks or a workflow, where the model must first perceive information, then reason, and finally decide on a tool or action to take.

This points to the model's specialized audio generation. To apply this, provide specific prompts that reference musical genres, instruments, tempo, and mood to guide the creation of high-fidelity audio outputs.

This underscores the importance of detailed language for visual generation. To apply this, use rich, evocative language in your prompts, specifying lighting, composition, style, and subject details to achieve the desired image quality.

This highlights the model's advanced creative capabilities. To apply this, be highly descriptive in your prompts, specifying mood, shot type, pacing, and audio cues to guide the model towards producing a specific cinematic result.

This indicates the model's strength in creating agents with specific roles and behaviors. To apply this, use persona prompting by defining a character, its expertise, its communication style, and its goals before asking it to perform a task.

This tagline suggests a core capability: use diverse inputs to generate diverse outputs. To apply this, pair unexpected modalities in your prompt, such as asking the model to generate a poem based on a data table or a musical score from a photograph.

Summary of "Googlebook" Discussion (Hacker News)

• The Concept: Googlebook is presented as a high-end hardware play—a "premium upgrade" to the Chromebook—intended to shift the user experience from traditional apps to an LLM-native environment powered by Gemini.
• The Vision: Some users see this as a "moonshot" to make apps irrelevant, replacing them with raw data feeds and AI-driven visualizations, potentially creating a "sci-fi" style user interface.
• Execution Skepticism: A dominant theme is doubt regarding Google's ability to execute. Commenters cited current frustrations with Gemini (cutting off mid-sentence, poor Maps integration, and hallucinations in Sheets) as evidence that the tech isn't ready for a primary OS role.
• Target Audience: Discussion suggests Google is targeting "Chromebook natives"—students who grew up in the Google ecosystem and are now entering university or the workforce and want a "pro" version of their familiar tools.
• Public Sentiment on AI: There is a sharp divide in perceived AI adoption. Some argue "normies" (non-tech users) love AI for daily tasks and creative fun, while others claim people increasingly resent AI being "forced" into every product.
• Competitive Landscape: Comparison was drawn to historical shifts (like the move to Mosaic/Web browsers). While some see this as an inevitable evolution, others believe Apple’s approach—gradually improving Siri and app intents—is more practical than trying to eliminate apps entirely.

🔒【令人震惊】Chrome 在数十亿设备上静默写入 4GB Gemini Nano 模型权重，删除后自动重装，可能违反 GDPR。这是「端侧 AI」与用户隐私的第一次正面冲突——不是关于数据收集，而是关于在未经同意的情况下使用用户存储空间和计算资源。这个事件的先例意义巨大：如果 Google 可以这样做，所有内置 AI 的操作系统和浏览器都有可能效仿，用户对自己设备的控制权正在被悄悄侵蚀。

在原始论文中，Gemini 3 Flash 以 24.0% 的成绩位列第一——而 Artificial Analysis 的独立复测中，它的成绩是 27.7%，被 GPT-5.4 和 Claude Opus 超越。两个不同时间、不同方法论的测试得出了不同的排名。这揭示了 AI Agent 评测的根本脆弱性：同一个 benchmark，不同实施者得出不同结论。「谁第一」在 AI 评测中是一个随时间和方法论变化的流动答案。

METR 公开列出了「尚未完成评测」的前沿模型，这个透明度本身就令人惊讶。更令人注意的是列表的内容：Gemini 3.1 Pro 和 GPT-5.2-Codex 都榜上有名，说明 METR 的评测能力跟不上模型发布速度。在 AI 能力快速迭代的背景下，「评测滞后」已成为 AI 安全领域的系统性风险——我们对最新最强模型的能力边界，永远处于半盲状态。

https://web.archive.org/web/20260106104924/https://theaidigest.org/village

Four AI models grouped together as a 'village' and set tasks (like elect a leader). The logs read like slapstick in a way, bumbling forward continuously.

via [[Stephen Downes p]]

Gemini has a CLI version, that is open source Chinese Qwen forked it for Qwen Code Jules is a Google coding agent.

Google Gemini made big strides in 2025

international math olympiad style questions can be answered by OpenAI and Gemini models without tools nor having the challenges in their training data.

Nano Banana how-to's and comparison, link to script to easily interact w Gemini API for image generation

• Google released Nano Banana Pro (gemini-3-pro-image-preview), a new AI image generation model.
• Nano Banana Pro excels in general intelligence, tool use, and creating complex scenes with less hallucination.
• It can use Google Search and Maps to gather data and reason visually through "thought images."
• Pushing infographic and map generation to new frontiers, enabling visually rich and factually accurate images.
• Can create detailed photorealistic images based on complex, multi-element prompts.
• Not reliable for electrical circuit designs yet, as it may produce erroneous circuit diagrams.
• Human intelligence still surpasses it in domain-specific tasks like accurate circuit design.
• Nano Banana Pro is seen as a game changer in practical, production-ready AI image generation.
• Tool use enables more factually accurate and data-driven generated images than previous models.
• Benchmarking AI image generation quality still needs development for production use assessment.
• The community is impressed with Nano Banana Pro's nuanced prompt following and image creation capabilities.

Levels of understanding genres: - 0) No understanding Like the song, never heard anything like it before, but no idea about anything. - 1) Basic Understanding Knowing a bit about the name of the genre and subgenres, but you can be wrong. - 2) Immersion Really dive into subgenres and flavors of the main genre... Also a bit of history about the genre. Research. - 3) Structure Breaking down the structure of the tracks in the genre. For example through DAW. Basically first-principles thinking.

To level 1: Song Analyzer tools (for example musicstax or AI). The author recommends everynoise.com too to gain a basic understanding of genres.

To level 2: Find similar songs and artists for your playlists with that genre. Perhaps playlists. Important to understand the origin of the genre.

Research: Theorization, Synthesis, etc.

Learning: Acquisition and Retention of Knowledge or Application of Skill

Both: When there is need of both and/or when research techniques don't give the necessary mastery quick enough for the material; too dense (i.e., neuroscience book)

( ~ 6:25-end )

Steps for designing a reading plan/list: 1. Pick a topic/goal (or question you want to answer) & how long you want to take to achieve this. 2. Do research into the books necessary to achieve this goal. Meta-learning, scope out the subject. The number of books is relative to the goal and length of the goal. 3. Find the books using different tools such as Google & GoodReads & YouTube Recommendations (ChatGPT & Gemini are also useful). 4. Refine the book list (go through reviews, etc., in Adlerian steps, do an Inspectional Read of everything... Find out if it's truly useful). Also order them into a useful sequence for the syntopical reading project. Highlight the topics covered, how difficult they are, relevancy, etc. 5. Order the books (or download them)

Reminds me a bit of Scott Young's Metalearning step, and doing a skill decomposition in van Merriënboer et al.'s 10 Steps to Complex Learning

I have always misinterpreted this idea.

I thought it referred to the rumination of ideas... Don't put notes you just made in it (from any source, like reading a book), instead let it ruminate.

I was wrong. I realized this when chatting with Gemini Advanced.

But either way, the DeepL translation of this paragraph: "Slip box as a clarification pit - don't just put clarified notes in it. Postponing reviewing and deciding - also a question of speed."

It is moreso related to the idea of fleeting notes and unprocessed ideas. Have to think a bit more about what Luhmann meant. Maybe @chrisaldrich knows something.

驚喜到吃手手

改天來如法炮製，看如何下prompt讓它儘量用全漢的台文輸出，減少羅馬拼音，除非基本字型出不來的字。這只是我的個人偏好。

太興奮了。我會希望AI用台文書寫，儘量減少tailo或POJ。以後來研究這個system prompt怎麼下。

AI

Taigi

Chhit Am-goat · 《 用台文直接共 AI 討台文內容 》 昨昏有朋友問講 AI 出無按家己向望彼號型。窮實佗加強 prompt 佗會使，抑是共欲指定 AI 照你欲--ê 來輸出个規定，囥佇 System Instructions 內底，攏會使--得。 下跤圖--ni̍h 是 ka 囥佇 Gemini-pro-1.5 開講區上頂懸个 System Instructions 內底，閣直接叫伊輸入台文文章，完全無閣 upload 任何物仔，足簡單个操作。 輸出愛檢查，家己愛改，毋過完成度是不止仔懸--矣，而且直接佗出台文內容，毋免閣過翻譯。 Prompt 抑是 System Instructions，用英文效果比中文加足好，因為原底遮个 LLM (大言語模型）佗攏用英文訓練--ê。 罔參考，會使按家己需要閣改，內底个例攏會換--得，例个長度小可長，效果較好。 若想欲輸出 全 POJ、全 TL、抑是全漢，會使改內底个規則佮例，試驗看覓。 《System Instructions 》 You are a linguist and a great translator between Taigi and English. Translate English text into Taigi I give you, and vise versa．Or chat with me in Taigi. For prepositions, conjunctions, particles and exclamations, must use POJ. For NERs such as country, place or human names keep the original name.English. Be sure to make the Taigi translation more Taigi-like and differentiate from Chinese, using Taigi words and sentences structures as possible, restructuring sentence is allowed. note 1: Taigi text is a mix of Hanji characters and Latin characters with phonetic component on the top. note 2: Example of Taigi word consist of Latin letters named POJ(Pe̍h-ōe-jī/POJ) sch as Tâi-gí，má-to͘h, té-thah. Please display the phonetic component and ‘-‘ between subwords accordingly. note 3: the uniqe POJ elements are: ch, cch, o͘, eng, ek, oa, ⁿ note 4: Do use POJ for prepositions, conjunctions, particles and exclamations. Output as format&example. <format&example>

• Leonardo da Vinci ê 無限好玄

Leonardo da Vinci 是文藝復興時期 ê 代表人物, 伊 ê 人生 kah 作品 lóng 浸透彼个時代 ê 精神。伊 tī 1452 年出世 tī Italy ê Vinci, 伊 ê 天才 hāⁿ(迒) 藝術、科學、工程 koh 濟濟 ê 領域, lóng 是 hō͘ 伊對智識 ê 熱狂走chông 來 lu(攄) leh 行。

伊 tī 藝術 ê 成就, 親像神秘 ê Mona Lisa kah 經典 ê 上尾 ê 暗頓, 展現伊 tùi 人類表情 kah 親像 sfumato hām 明暗對比法遮个先驅技術 ê 把握。 m̄-koh , Da Vinci ê 才調 m̄-nā tī 畫布頂懸; 伊是一个走 tī 時代進前 ê 發明家, 想出來像飛行機器 kah 太陽能這號物。

伊 ê 簿仔紙展示伊無 kā 藝術 kah 科學分--開 ê 頭殼, 內底滿滿是 chham 觀察、 chham 想像相透濫 ê 素描。Da Vinci ê 哲學是「研究藝術 ê 科學。研究科學 ê 藝術... 認捌 tio̍h 萬物 lóng 連連鬥陣」, 這反映伊相信所有學科 lóng 是 kap(合)做伙--ê。

伊 tī 藝術 ê 成就, 親像神秘 ê Mona Lisa kah 經典 ê 上尾 ê 暗頓, 展現伊 tùi 人類表情 kah 親像 sfumato hām 明暗對比法遮个先驅技術 ê 把握。 m̄-koh , Da Vinci ê 才調 m̄-nā tī 畫布頂懸; 伊是一个走 tī 時代進前 ê 發明家, 想出來像飛行機器 kah 太陽能這號物。

Gemini-1.5-pro: https://aistudio.google.com/