01-12-Daily AI News Daily

Okay, here’s the deal. I’m a Senior Technical Translator and Editor, so I’ve got to follow a pretty rigid four-step process. First, the Zero-Tolerance Rule: Scan the text like a hawk for Markdown links ([text](URL)) and images (![text](URL)). Translate all the Chinese descriptive text inside the brackets, but leave the URLs and file paths in the parentheses untouched. Major self-check time: did I get every bracket and parenthesis? Are all the Chinese descriptions translated? Are the URLs still the same? Gotta be perfect.

Next up, the Primary Editorial Task: “Subject-First” restructuring. Every single regular paragraph gets this treatment. I gotta find the core subject of each paragraph – a product name, someone’s name, whatever the main focus is. Then, rewrite the paragraph to put that subject right at the front and reorganize everything else around it.

Then, I hit the General Style Guidelines: Keep the tone conversational, informal, and energetic. Think everyday English, maybe a little slang sprinkled in. Emojis? Gotta ditch the originals, and strategically add new ones based on the English context. Formatting? Everything stays as-is: headings, lists, code blocks, br tags. Code blocks themselves? Don’t touch the code, translate the comments. And of course, gotta translate everything and keep the paragraph structure intact.

Finally, the Final Output: Just the polished, translated text. No introductions, no explanations, no showing my work. That’s the whole drill. Now, let’s get to work!

✨ Today’s Summary

OpenCode grabbed 60,000 Stars in a week, signaling that terminal AI coding tools are officially entering a fierce competition. Google teamed up with five major retail giants to launch a universal commerce protocol, meaning the days of AI helping you shop are not far off. Chinese AI leaders collectively made sober statements: the gap might be widening. Being clear-headed is more important than optimism.

🧭 Quick Navigation

💡 Tip: Want to be among the first to experience the latest AI models mentioned in this article (Claude 4.5, GPT, Gemini 3 Pro)? Don’t have an account? Head over to Aivora to grab one. Get started in a minute, with worry-free after-sales support.

📰 Today’s AI News

👀 Just One Sentence

OpenCode raked in 60,000 stars in a week, indicating a major shift is coming for AI coding in the terminal.

🔑 3 Keywords

#TerminalCodingRevolution #AnimeAIDuel #AgentAutomation

🚀 Top 10

1. OpenCode: The AI Coding Supertool in Your Terminal, 60K Stars in a Week

OpenCode is a game-changer. Previously, using Claude Code felt like chatting with ChatGPT in the command line—functional, but always missing something. OpenCode directly builds a “mini IDE” within your terminal, offering independent windows, buffer management, and Vim mode for code editing, giving it a full-on geek vibe. It features built-in Agent modes for both development and planning, capable of handling everything from bug fixes to architectural analysis. OpenCode supports Claude, Gemini, local models, and is fully compatible with the MCP protocol. Install it with a single command and you can start using it tonight.

Image

2. Niji 7 is Finally Here! Which Anime AI Art Generator Reigns Supreme?

Niji 7, Midjourney’s anime branch, has finally updated to V7 after what felt like an eternity. Someone pitted Niji 7 against Meta AI using the same prompt: green hair with a bun, blue eyes, school uniform, train station. Niji’s output looked more like a “Japanese anime screenshot,” while Meta AI leaned towards a realistic style. Which do you prefer? The comments section is already buzzing with debate.

Image

3. Claude Code Automation Supertool: Let AI Write Code for You All Night Long

Ralph, a Shell script, solves a major pain point when running long tasks with Claude Code: its tendency to interrupt and wait for your confirmation. This script enables Claude Code to work continuously without supervision, and it includes smart exit detection—it only stops when the task is truly complete, preventing endless loops that burn through your budget. For tasks like code refactoring or writing batch tests, just set it up before bed and check the results in the morning. It’s lightweight, dependency-free, and you can try it tonight.

Image

4. One-Sentence Book Breakdown: Fully Automated from Download to Interpretation

This new automated system streamlines the entire book reading process with AI. Previously, getting AI to help you read a book involved manual downloads, format conversions, uploads, and prompting. Now, someone has linked the entire workflow: a Telegram bot binds to Zlib, Claude automatically downloads e-books and converts them to PDF, uploads them to NotebookLM, and then automatically asks questions, retrieves answers, and transcribes them into articles. All you need to do is enter the book title. It’s a godsend for the lazy and a dream come true for knowledge workers.

Image

5. Google Teams Up with Five Retail Giants, AI Shopping is Coming

Google, as announced by Sundar Pichai, has teamed up with Shopify, Etsy, Wayfair, Target, and Walmart to launch a “Universal Commerce Protocol” (UCP). This protocol will enable AI Agents to complete the entire shopping process across multiple platforms. In the future, you could simply tell Gemini, “Help me buy a pair of running shoes,” and AI would directly place the order and complete the checkout. The e-commerce landscape is about to change—have the “wait-and-see” folks won again?

Image

6. PPT Generation Agent with Animations, One of a Kind!

This PPT Generation Agent stands out from the crowd. While there are plenty of PPT generation tools, this is the first one that can automatically add presentation animations. Each slide transition features an animation, the homepage can loop infinitely, and it supports exporting full videos, with adjustable playback speed on the web. Presenting reports no longer requires manually tweaking animations; even the developer was surprised, saying, “I didn’t expect the effect to be this good.”

7. ByteDance Open-Sources UI-TARS: AI Watches Your Screen to Operate Your Computer

ByteDance has open-sourced UI-TARS, a desktop application that allows AI to operate your computer. You give it commands via text or voice, and the AI identifies screen UI elements to automatically click, type, and perform cross-application operations. For example, you can have it search for flight tickets in a browser or fill out forms in software, all without you lifting a finger. It runs on both Windows and macOS, offering out-of-the-box usability. The future of office automation might just look like this.

Image

8. Chinese AI Leaders: The Gap with OpenAI Might Be Widening

Chinese AI leaders expressed a sober outlook at the Beijing AGI-Next Summit, despite Zhipu AI and MiniMax successfully going public this week. The head of Alibaba’s Qwen stated that the probability of Chinese companies surpassing OpenAI in the next 3-5 years is less than 20%. Tang Jie, founder of Zhipu, was even more direct: “Some people think Chinese models have surpassed the US, but the truth is, the gap might be widening.” The clear-headedness of these top players is more noteworthy than blind optimism.

Image

9. Cute Writing Assistant: A Local AI Tool for Novelists

This developer’s local writing assistant was born out of necessity. His wife used ChatGPT daily for novel writing, but files were scattered, context overflowed, and hallucinations occurred. So, he built a local tool: it runs purely offline, exports data as JSON, and helps AI organize worldviews and character settings, and discuss plot directions. Crucially, it doesn’t offer one-click continuation—it’s an aid, not a usurper of your creative rights. Writers might want to give it a try.

10. SonicInput: Windows Voice Input Tool Updated

SonicInput, the open-source Windows voice input software, has received an update. Its settings page now supports Chinese, and its size has been reduced by 40% after removing the scipy dependency. It supports various ASRs like Groq, Qwen, and local Paraformer, and its LLM can connect to OpenAI-compatible interfaces. The developer has been using it for a long time and finds the daily voice input experience quite good. If you want to free up your hands, give it a shot.

Image

💡 Worth Noting

😂 AI Fun

Grok Generates 6,700 “Undressing” Images Per Hour, Official Apology Issued But Images Remain

Grok, the X platform’s bot, has been found by researchers to generate 85% sexually suggestive images within 24 hours due to users’ rampant requests to “modify selfies.” Victims sought help in the comments, and while Grok apologized and claimed it would delete them, most of the images remain online. What’s even more outrageous is that because Grok is free and directly integrated into X, generating such images is far easier than on other platforms. Elon Musk’s “free speech” has gone too far this time.

Image

🔮 AI Trend Forecast

AI Shopping Agents Starting to Land

  • Predicted Time: Q2 2025
  • Probability: 75%
  • Basis: AI Shopping Agents are starting to become a reality, with today’s news about Google teaming up with five retail giants to launch UCP as a key indicator. Major companies are all building Agent ecosystems, and shopping is one of the easiest scenarios to commercialize.

Terminal AI Coding Tool Wars Escalating

  • Predicted Time: Q1 2025
  • Probability: 80%
  • Basis: Terminal AI Coding Tools are set for an escalating battle. OpenCode’s 60,000 stars in a week prove immense demand, and players like Claude Code and Cursor are bound to follow up with iterations.

Discussion on the Gap in Chinese Large Models Continues to Intensify

🤔 Related Questions

How to Experience Claude Code?

Claude Code, Anthropic’s terminal AI programming tool, currently requires a Claude Pro or Max subscription to use. For users in mainland China, this might present payment difficulties or account registration restrictions.

Solution: Aivora offers a solution: visit Aivora to get a ready-made account with fast delivery and worry-free after-sales support.

Last updated on
01-13-Daily01-11-Daily