I used AI to build myself an investment workstation

Over the past two weeks, I’ve become somewhat obsessed with Vibe Coding.

Not the kind of obsession where I aim to “build a groundbreaking product.” Instead, it’s the sudden realization that many of those small ideas lingering in my mind can actually be brought to life, piece by piece, by my own hands.

As you know, Vibe Coding means using natural language to instruct AI to write code and build a product.

I usually use the Codex and Claude Code clients in tandem, describing requirements and functional modules for them to code. When my credits run low, I switch to the CLI connected to the DeepSeek API to continue.

1. Those Thoughts I've Wanted to Act On, But Never Did

I used to have a swarm of ideas pop into my head.

For instance, could there be a single dashboard where I could view US stocks, crypto, Hong Kong stocks, and A-shares together, instead of switching back and forth between several apps every day?

Or, could I create an anomaly monitor that instantly alerts me when a particular asset suddenly surges or plummets, while also showing which other assets or sectors it’s correlated with?

Another example: could I build an investment map for researching a specific sector, not just looking at a single project, but laying out the entire web—upstream and downstream beneficiaries, potential risks, and related assets?

Furthermore, Prediction Markets (PM) have bets on things like private company valuations, market cap overtakes, or macro events. Could I line this data up alongside news events and secondary market changes for a side-by-side comparison?

Plenty of ideas, but making them a reality? Too much hassle.

You need to know code, design pages, connect data sources, and iterate endlessly. Outsourcing is costly, and it’s hard to articulate exactly what you need. After a few rounds, most ideas fade into that familiar refrain: “Forget it, I’ll just make do with Excel for now.”

But after two weeks of tinkering with Vibe Coding, I’ve realized things are genuinely different now.

I’ve started building rough but functional tools for myself. When an idea comes to me, I can get it into a system that same day, instead of it getting lost in chat histories, bookmarks, or my own head.

2. Four Small Tools I Built in Two Weeks

Over these two weeks, I mainly focused on building four things (ignoring other tiny, throwaway tools).

First: A Cross-Market Asset Dashboard

The reason was simple. My assets are scattered across several places: Hong Kong and US stocks in brokerage apps, crypto on exchanges, and A-shares in yet another platform.

To check my overall portfolio status, I had to open each app, one by one, and jump between them. After going through the whole circuit, I still couldn't piece together a complete picture. So, the first thing I did was cram all my holdings onto a single page:

On top, you see total assets and daily P&L; below, it's segmented by market—a column for US stocks, one for crypto, one for Hong Kong stocks, one for A-shares. A single glance tells you your entire financial status and exactly who’s up and who’s down today.

Finding it surprisingly useful, I couldn't resist adding one tab after another. As I used it, new needs kept popping up:

• Anomaly Monitor: I pre-set the assets and thresholds I care about. If anything experiences a sharp surge or drop, it flags it immediately, saving me from having to watch the screens constantly.
• Investment Map: When researching a sector, I visualize the upstream, downstream, beneficiaries, risk points, and related assets into a network, making it easy to trace capital flow and relationship chains.
• Notes & Review: I jot down why I was bullish at the time, what happened later, where my judgment was right or wrong. I can easily refer back to it later.

Since this dashboard contains all my real holdings and is quite private, I deployed it locally.

Second: PM Bet Monitor

This one is specifically for tracking prediction markets.

Simply put, prediction markets (like PM) allow people to bet real money on whether a future event will occur. The price itself represents the market's perceived probability. For example, if a “yes” bet on “SpaceX reaching a $2 trillion market cap by the end of June” is priced at 0.8, the market believes there’s an 80% chance it will happen.

I follow bets on things like “Will OpenAI/Anthropic’s valuation go up by year-end?”, “Will one of the Magnificent Seven overtake another in market cap?”, or “Will X and Y meet?”. Previously, I had to check these one by one. Now, I’ve centralized them on a single dashboard, placing probability changes alongside news events and secondary market fluctuations. It’s instantly clear which moves first and which influences which.

I’ve also stratified these bets based on my own criteria (internally called T1 for high conviction, T2 for relatively stable, and T3 for pure speculation), sorted by expected returns. This makes it easy to see at a glance which bets are just noise.

To be honest, one of my small edges in this market is access to Chinese-language information and East Asian political/economic dynamics. Many of these markets are dominated by Western players, who are often slow to price in this area. The opportunity lies in this time lag.

Third: A Small Operations Backend

This one is unrelated to investing; it’s for my own writing.

I usually handle topic selection, drafting, and publishing across several platforms, relying on memory and searching chat history. It gets messy quickly. So, I built a small backend to manage my writing tasks, covering topic lists, article progress, publishing platforms, and an idea bin.

Since I might need to access it on the go, I didn’t deploy it locally. Instead, I put it on the cloud—using GitHub + Vercel. I can open it on my phone to view and make changes, which is quite convenient.

Fourth: A One-Click Formatting Tool

This was mainly to solve my personal hassle. After finishing a draft for distribution on multiple platforms—especially Web3 media, each with its own formatting rules—manually adjusting everything was extremely time-consuming.

So, I built a small tool. Combined with a browser Tampermonkey script I coded, I just drop in an original Markdown or Word document. It automatically converts it to the required format for each platform and inserts images. It’s not high-tech, but it saves me a chunk of mechanical work every day.

Admittedly, these four tools are all very basic right now, maybe even a bit ugly, and certainly not mature products. But they are incredibly useful to me. When an idea comes, I can get it into a system immediately, instead of letting it scatter and be forgotten.

This is the most important change I feel.

3. The Way Ordinary People Do Research is Truly Changing

Because of this, I increasingly feel that when ordinary people start investing, they don't necessarily need to dive straight into complex models. But at least they should have a few basic, personal systems in place.

Because the way AI is changing things for regular people isn’t by turning you into an expert overnight. It’s by allowing you to create a prototype for many things that were previously “things you wanted to do but couldn’t.”

This is particularly noticeable for someone like me who monitors markets daily. With the right mindset, every ordinary investor can gradually build up a few foundational systems:

• An Asset Observation System: What assets are you actually watching? Which market do they belong to? What changes are happening recently?
• A Signal Monitoring System: Which events, once they occur, might signal a shift in market expectations?
• A Knowledge Mapping System: A sector is not a single point but a network. Who is upstream, who is downstream, who benefits from sentiment, who benefits from earnings, and who benefits from capital flows? This has been especially true over the past year, where stocks in the AI sector have practically rewarded those who could deeply understand an entire vertical chain (from HPC to optical modules to storage).
• A Review System: Why were you bullish at the time? What happened later? Where were you right, and where were you wrong?

These things weren't impossible to do before, but they were too much trouble and hard to maintain. The greatest significance of AI is that it has slashed a huge chunk of that “trouble.”

You don't necessarily need to know how to code, but you can describe your needs and gradually build your own product design. You don't have to finish it all at once. Create a first version, and refine it as you use it.

This is also what attracts me most about Vibe Coding: the incredibly fast feedback loop. Before, there was often a long gap between an idea springing up and its realization—sometimes so long that you forgot why you wanted to do it in the first place.

Now, if I think of a feature today, I can try it out the same day. If I'm not satisfied after trying it, I modify it immediately. After using it for two days, new needs emerge, and I iterate again.

Once this closed loop of “Idea → Implementation → Use → Feedback → Modify” starts spinning, it genuinely becomes hard to stop.

Final Thoughts

Consider this the first record of the new phase for “Tyler Taille.”

Going forward, I'll try to update regularly, documenting my investment thoughts, hands-on tool tests, on-chain operations, and arbitrage research, as well as some beginner-friendly Web3 practical knowledge and investment tips.

Feel free to follow along and share your thoughts anytime.