Hex2077's AI Daily | AI News | AI Tools | AI Insights – 何夕2077的博客

11-19-Daily AI News Daily

Wed, 19 Nov 2025 15:32:00 +0800

Google’s latest demo video totally just blew my mind. I’d been collecting a few things to talk about, planning to save them for later, but I simply couldn’t hold back. Had to jot these thoughts down ASAP before they slipped away.

Google Antigravity, something DeepMind quietly unveiled a couple of days ago, has made quite a stir.

Antigravity initially comes off as an IDE (Integrated Development Environment), but the more I checked it out, the more I realized something was different. This tool doesn’t quite feel like the Cursor or Copilot we’re using now. Instead of just handing me a faster ‘gun,’ it feels like it’s given me a capable ‘spotter’ to back me up.

So, I’ll walk you through my thought process from watching the demo and share what’s so cool about this thing.

The UI is Kinda Wild, But I’m Game

I was honestly a bit confused when I first opened the video.

Antigravity totally breaks the mold of what we expect from an IDE. Instead of the usual left-side file tree and right-side code editor, it’s rolled out three core interfaces, which it calls ‘Three Surfaces’.

Beyond the familiar Editor, Antigravity also shoves in an Agent Manager and a Browser.

Chrome is directly embedded in the IDE? Seriously? Isn’t that just asking for memory to explode? 🤯… Then again, Chrome is Google’s baby, so a deep integration totally makes sense.

The Agent Manager is pretty neat. You can actually see several AI agents doing their thing – not just one, but a whole crew. You can tell one to crank out backend code and another to whip up some designs.

The built-in Browser isn’t just for you to preview web pages. Here’s the wild part: it’s for the AI.

Once the AI’s code is up and running, the AI itself takes control of that browser, clicking buttons, typing text, and scrolling pages just like a human tester. Honestly, watching the AI interact with its own self-written webpages gives you that heartwarming ‘my kid finally learned to dress themselves’ kind of feeling.

Don’t Ask, It’s “Fully Automatic”

Most current AI programming tools still rely on that back-and-forth chat: ‘help me write a function,’ ‘fix this bug for me,’ and so on.

Antigravity, get this, has an Auto mode.

With this mode on, the AI just goes wild. If it figures it needs to install dependencies, it’ll run terminal commands on its own. Needs to create a file? Done. It won’t stop at every single step to ask you, ‘Hey boss, mind if I run npm install?’

Of course, the video did mention it’ll still ping you for sensitive operations.

Right now, I’m at the point where I’m sometimes too lazy to even type a prompt. This ‘I’ll just watch and not lift a finger’ vibe is seriously kinda sweet. As long as it doesn’t go deleting my libraries, alright?

Antigravity also supports multi-task parallelism. In the demo, the presenter was having the AI write a backend API while simultaneously firing up a new conversation to get it to design a logo. Both tasks happened at the same time, without skipping a beat. Now that’s what a multi-core CPU is made for!

No More “Black Box”: It Actually Writes Reports

What’s everyone’s biggest fear when using AI to code? It’s getting a bunch of gibberish that won’t run and is a total nightmare to debug.

Antigravity introduces a system called Artifacts here, and I think this idea is seriously on point.

Before getting to work, Antigravity will first lay out a Task List, telling you its game plan, step-by-step.

Then, listen up, it’ll whip up an Implementation Plan. It’s like it hands you a technical proposal doc first, and only when you give the green light does it actually get its hands dirty.

The coolest thing is that after it’s done working, Antigravity generates a Walkthrough report.

My Leader used to constantly hound me for daily reports. Now, get this, the AI writes my daily reports for me. What a time to be alive!

This report doesn’t just show what code it changed; it even includes screen recordings! Remember that built-in browser I mentioned? It records its own testing process for you to watch: ‘See? I typed in the flight number, clicked search, and boom, results popped up, all good.’

This totally puts your mind at ease.

Paired with Gemini 3.0, It’s Seriously Smart

Gemini 3.0, Google’s latest model, is the brains behind all this.

The ‘Flight Tracker’ demo in the video really shows off what it can do.

To get the data sorted, the AI actually went and hunted down the “Aviation Stack API” docs all by itself. It didn’t just pull a URL out of thin air; it genuinely read the documentation, and even ran a curl command in the terminal to verify the API Key and check out the data structure. How wild is that?

It even generated the logo right there using the Nano Banana model and casually tossed it into the project. And get this: it even integrated the flight info into Google Calendar at the end.

This isn’t just code completion, folks. This is a full-blown ‘digital employee’ that’s got the chops of a junior full-stack engineer, a test engineer, and a UI designer all rolled into one.

It Even Handles My Commit Messages

Finally, here’s another feature that totally hit a nerve for me: Context-Aware Commit.

When I usually commit code, my Commit Messages are often those super unhelpful words like update, fix, or wip.

Because Antigravity watches everything you (and it) do from start to finish, it can automatically whip up super detailed and accurate Git commit messages based on the context.

This feature is an absolute lifesaver for anyone with OCD. Though, I gotta wonder, will it feel a bit strange seeing Git Logs full of AI’s ‘voice’ down the line? 🤔

Anyway, after checking out that entire demo, my takeaway is that Google totally came prepared this time.

Antigravity teamed up with Gemini 3.0 — the vision it’s pushing isn’t just about ‘helping you type characters’ anymore. It’s about ‘getting stuff done for you.’ From understanding requirements, digging through docs, writing code, whipping up assets, running tests, and finally committing code — this end-to-end service really has some serious chops.

Alright, since you’ve made it this far, if you found this tech share pretty interesting, why not hit that follow and like button? Let me know there are plenty of fellow enthusiasts out here who love messing around with new toys! 👋

10-28-Daily AI News Daily

Tue, 28 Oct 2025 23:53:00 +0800

This guide details the process for opening a bank account online within Hong Kong. 🇭🇰

This guide covers five banks: HSBC, Bank of China (Hong Kong), ZA Bank, Airstar Bank, and Ant Bank. Following these steps, you can complete all online applications within a single day.

Part One: Getting Ready

Before opening an account, you’ll need to gather these materials. Missing any of them means you won’t be able to complete the online process. ⚠️

Identification Documents
- Mainland Chinese Resident Identity Card.
- Exit-Entry Permit for Hong Kong and Macau.
- Ensure both documents are valid for more than six months.
Mobile Phone Number with International Roaming Enabled This mobile phone number is solely for receiving SMS verification codes from Hong Kong banks, not for internet access. You can handle internet traffic via Alipay’s international data packages or your carrier’s roaming service.
Exit-Entry Record Certificate (PDF) This certificate is a crucial document.
- On WeChat, search for the “Immigration Administration 12367” mini-program.
- Once inside, download the “Exit-Entry Record Query Certificate”.
- Save the generated PDF file to your phone.
- Just one exit record from a mainland port of entry is sufficient. ✅
Physical Card Mailing Address You’ll need to prepare a mainland address for receiving your physical card. This address must be provided in both Chinese and English versions. For lengthy addresses, feel free to use standard abbreviations, such as ‘Building’ abbreviated to ‘BLDG’. 🏠
Bank Applications (Apps) You’ll need to download the following applications in advance:
- HSBC HK
- BOCHK
- ZA Bank
- AIRSTAR Bank
- For Ant Bank, you’ll operate via its Alipay mini-program, so no separate app is needed.

Part Two: Operation Process

The suggested order for applications is to first process HSBC and Bank of China, then move on to the three virtual banks. 🚀

1. HSBC Bank (HSBC)

The entire application process for HSBC is completed online.

Open the HSBC HK app and select “I don’t have any HSBC account.”
During identity verification, choose “I am in Hong Kong, but I don’t have a Hong Kong ID card.”
For account type, select “HSBC One”; this account has no minimum deposit requirement.
Fill in the information:
- For account purpose, it’s recommended to select “Savings/Investment”.
- For tax identification number, fill in your Mainland Chinese ID number.
- Enter the prepared Chinese and English mailing addresses.
Upload documents: As prompted, take photos of your identification. Afterward, you’ll need to use your phone’s NFC function to read the chip information from your Exit-Entry Permit for Hong Kong and Macau. Just hold the permit close to your phone’s back sensing area.
After setting up login credentials, submit your application. Approval results are typically notified via email within a few minutes.

Regarding Card Receipt: Your physical card will be dispatched by the card manufacturer IDEMIA via EMS from Shenzhen. We recommend following the “EMS China Post Express Logistics” WeChat service account and binding your phone number to track its delivery. If you haven’t received it after two weeks, you’ll need to contact bank customer service. 📞

2. Bank of China (Hong Kong) (BOCHK)

The BOCHK application process is similar to HSBC’s.

Open the BOCHK app and select “Open Account.”
For identity, select “Chinese Resident Identity Card” and confirm “I am in Hong Kong.”
For account opening method, choose “I am not at a branch,” and for account type, select “Easy Banking” (自在理财); this account also has no deposit requirement.
Upload documents: As prompted, upload your Identity Card, Exit-Entry Permit for Hong Kong and Macau, and the prepared Exit-Entry Record PDF file.
Complete facial recognition and personal information entry. For account opening reason, select “Investment & Wealth Management” or “Savings.”
Submit the application.

3. Virtual Banks: ZA, Airstar, Ant Bank

These three — ZA, Airstar, and Ant Bank — are virtual banks, and their application processes are largely identical: fill in information, upload documents, and perform facial recognition.

Do note the following crucial points: 👇

Address: You must provide a detailed address, precise down to the house number.
Tax Status: Select “Only a tax resident of Mainland China.”
Tax ID Number: Fill in your Mainland Chinese ID number.
Exit-Entry Record: This is required for all of them.

Part Three: General Principles

Data Authenticity: All information you submit must be completely truthful. Banks will verify everything, and providing false information will lead to application failure and potentially being blacklisted. 💯
Information Consistency: Make sure that details like your name’s Pinyin, address, and other information are absolutely consistent across all bank applications. ✨
Account Opening Purpose: When asked, choose straightforward options like “Personal Financial Management,” “Savings,” or “Hong Kong Stock Investment.”
Operation Time: We highly recommend submitting your application during working hours on weekdays for faster review speeds. ⏰
Account Activation: After successfully opening your account, you should immediately log into the mobile banking app and set up fingerprint or facial recognition login. This is the crucial first step to proving your account is active. 👍

10-09-Daily AI News Daily

Wed, 08 Oct 2025 06:06:00 +0800

19 Days to Launch a Website, Plugin, and Desktop App Suite: An Unconventional AI Coding Practice Report

So, remember my AI Coding invitation? Well, after all these days, the project’s finally live, and my head’s clear! Forget the sentimentality and the personal journey for today. Let’s get hardcore and talk frankly about how I used AI as my ‘pair programming’ buddy to turn the PromptHub project from just an idea into a stack of runnable code in just 19 days. ✨

This report? Yeah, it’s probably not what you’re expecting from those ‘Vibe Coding’ videos. There’s zero magic here—just pure engineering, tough trade-offs, and a whole lot of ‘aha!’ moments after tripping over every single pitfall imaginable. 🚧

Architecture Selection: AI as the “Scaffolding,” Me as the “Decision-Maker”

PromptHub? From the jump, I hit it with a pretty wild goal: build for Web, Chrome extension, and Electron desktop—all at the same damn time! Backend? Next.js API Routes. Database? SQLite to start, but ready to swap to a production setup whenever. 🚀

This tech stack, honestly, used to be a nightmare for me. Back in the day, just getting package.json balanced and tsconfig.json squared away for all those different environments? That alone would’ve sent me spiraling. 🫠

My approach? Simple: I treat AI like a super-advanced scaffolding generator. I’m not asking it ‘what tech should I use?’ Nah, I’m just giving it direct orders: 👇

“I need a Next.js project, using TypeScript. Integrate Drizzle ORM, with SQLite for the database. Add JWT authentication, implement Google and GitHub OAuth login. Then set up the Stripe billing framework and leave me the interfaces ready.”

On September 17th, AI spent roughly an afternoon on a basic multi-language template project and, voilà, it spat out a fully functional backend framework for me! This wasn’t just a few code snippets; this was architecture made real. It crushed all the super tedious, repetitive ‘glue code,’ letting me dive headfirst into the core business logic. 🤯

My first takeaway? Easy: In a project’s early days, AI’s ultimate superpower is just wiping out all that ‘startup friction.’

Development Methodology: Ditch “Vibe Coding,” Embrace “Atomic Tasks”

Those ‘Vibe Coding’ (one-liner development) videos you see online? Honestly, just watch ’em for kicks, ‘cause anyone who buys into that is seriously naive. 🤪 Real, enterprise-level project development? That’s all about rigorous engineering. The pattern I’ve stumbled upon—I like to call it ‘Atomic Tasks’ or ‘Building Blocks’—is the real deal.

When I’m cooking up a new feature, I break the whole process down to its absolute bare bones. Then, I pop open a bunch of AI chat windows and get them all chugging along in parallel, like so: 👇

• Window A (Database Expert): “Based on my requirements, design the prompts table structure and write it out using Drizzle ORM syntax.”
• Window B (Backend Expert): “Here’s the table structure. Write me the corresponding CRUD APIs, implement them using Next.js API Routes, and ensure proper permission checks.”
• Window C (Frontend Expert): “Here are the API interfaces. Using React and Tailwind, write me a management page component that can call these interfaces.”

So, the perks of this pattern are pretty sweet: 👇

1. Context Isolation: Each AI window focuses on just one thing, so it doesn’t get “mentally confused” by an overly long context.
2. Single Responsibility: Code decoupling is super clean, meaning AI rarely spits out spaghetti code that mixes frontend and backend.
3. Parallel Efficiency: While I’m waiting for the backend APIs to be written, I can already start brainstorming frontend components.

My second big takeaway? Don’t even think about treating AI like some all-knowing, all-powerful deity. Instead, view it as a small, specialized team made up of multiple ‘domain experts.’ 🧠

Hardcore Pitfalls: Moments Even AI Couldn’t Save My Butt

AI Coding? It’s no silver bullet, trust me. 🚫 In certain areas, especially when it comes to low-level stuff or configurations, AI somehow manages to mess up even worse than I do. Seriously.

1. Database Selection: Turso vs. Supabase

Initially, just for kicks, I gave the distributed Turso database a shot. Sounds cool, right? But man, the data synchronization delay was beyond ridiculous—a user would create a prompt, and it just wouldn’t show up even after refreshing a gazillion times. 😵‍💫 I even tried slapping on the consistency=strong parameter, but it did absolutely nothing. Zero.

So, I straight-up ditched it and swapped back to PostgreSQL-based Supabase. 🔄 That kind of call? AI can’t make it for you. You gotta deeply understand a database’s consistency model to grasp why Turso’s async vibe fundamentally clashed with my whole business scenario.

2. Next.js’s `useEffect` Infinite Loop

This? Oh man, this is a classic problem. 🤦‍♂️ On the management page, the API was getting hammered in an infinite loop. I tossed the code to Qwen3, and it futzed with it for ages, but nope, still no fix.

Ultimately, I had to get my hands dirty and fix it myself. I dug into the useEffect dependencies and realized there were just too many dynamic states mixed in, creating a nasty chain reaction. So, I manually refactored it, keeping only the absolute core user?.personalSpaceId as a dependency. Boom! Problem solved. ✅

After that, I made sure to feed the correct solution right back to the AI, telling it: ‘Hey, next time you hit a snag like this, this is how you fix it.’ Essentially, I was reverse-training the AI, schooling it on my best practices. 🧑‍🏫

3. Chrome Extension’s Permission Black Hole

When it came to extension development, AI was basically a total noob. content.js flat-out refused to load, localStorage data just wouldn’t communicate… every single answer AI spit out for these issues? Dead wrong. 😤

Ultimately, I just had to humble myself, crack open the Chrome developer docs, and actually figure out the difference between host_permissions and scripting permissions. Only then did I finally squash that bug. 🤓

My third crucial takeaway? AI’s a whiz at ‘implementation,’ no doubt, but it seriously struggles with ‘decision-making’ and ‘debugging.’ Especially when you’re deep in the weeds with underlying principles, platform quirks, or performance bottlenecks—the final call and the real debugging? That’s still all you, buddy. 🛠️

My Model “Toolbox”

Me? I never blindly put all my faith in just one model. My game plan is dynamic switching: use the right tool for the right job, every single time. 🧰

• Architecture Design & Complex Bug Fixing: Gemini 2.5 Flash is my go-to. It’s free, and it works wonders for tricky issues like Next.js hydration errors.
• UI/UX Code Implementation: Claude 4.1 is the undisputed champion here. Its CSS aesthetics and code implementation skills are top-notch, though it is the priciest, so I only bust it out for critical pages.
• Daily CRUD and Component Development: Qwen3 Coder Plus offers the best bang for your buck, a true workhorse that never complains about the grind.
• Data Processing and Script Generation: When reverse-engineering Google AI Studio’s API for data migration, I used Kilo paired with Gemini. It analyzed JSON structures and automatically whipped up Python scripts with insane efficiency.

So, all in all, these 19 days of development? Less ‘AI programming,’ more like ’extreme human-machine collaborative programming.’ Think of AI as that blazing-fast intern cranking out code, and me? I’m the architect, constantly steering the ship, making the big calls, and always ready to jump in and save the day when things go sideways. 🚀👨‍💻

Under this model, a developer’s core value totally shifts. It’s less about ‘writing code’ and more about ‘asking the right questions,’ ‘making smart decisions,’ and ‘crushing system design.’

This, folks, might just be what our future as developers is all about. Wild, right? 🤔

Here’s the website link, feel free to give it a try: https://prompt.hubtoday.app/

Oh, and you can also hit me up on WeChat to join the discussion group: justlikemaki