I built myself an investment workstation using AI

In the past half month, I've become somewhat addicted to Vibe Coding.

It's not the kind of addiction where "I want to build an incredibly impressive product." Instead, I suddenly realized that many small ideas that had been lingering in my mind can actually be brought to life, bit by bit, by myself.

As everyone knows, Vibe Coding means using natural language to command an AI to write code for you and build a product.

I mainly use a combination of Codex and Claude Code clients. I describe the requirements and functional modules, and they write the code for me. When I run out of quota, I switch to the CLI connecting to the DeepSeek API to continue.

1. Those "Wanted to Do, But Never Did" Ideas

I used to have a bunch of ideas popping into my head.

For example, wouldn't it be great to have a dashboard where I can view assets like US stocks, Crypto, Hong Kong stocks, and A-shares all in one place, instead of switching between several different apps every day?

For example, could I create a monitoring tool for unusual market movements? If an asset suddenly surges or plummets, I could see it immediately, and also know which other assets or sectors it's related to.

For example, could I build an investment map? When researching a specific track, instead of just focusing on one project, I could lay out the entire network: upstream and downstream companies, beneficiary assets, potential risks, and related investments.

Another example: Prediction Markets (PMs) have a lot of bets on the valuations of unlisted companies, market cap overtakes, and macroeconomic events. Could I cross-reference this data with news milestones and changes in the secondary market?

I had plenty of ideas, but actually making them happen was just too troublesome.

You need to know how to code, how to design pages, how to integrate data, and you have to revise things repeatedly. Outsourcing it is costly, and it's hard to articulate the requirements clearly. After going back and forth a few times, most ideas end up with that phrase – "Never mind, I'll just make do with Excel for now."

But after tinkering with Vibe Coding for these two weeks, I found that this time, it's truly different.

I started building some very crude but functional tools for myself. An idea pops up, and by the end of the day, it can be working in my system, instead of being scattered across chat logs, bookmarks, and my own memory.

2. In Half a Month, Four Small Tools I Built

In this half month, I mainly made four things (not counting other trivial little tools).

First, a Cross-Market Asset Dashboard

The reason was very simple. My assets are scattered in multiple places: Hong Kong and US stocks are in broker apps, Crypto is on exchanges, and A-shares are on another platform.

Every day, to get a sense of my overall situation, I had to open each one, switch back and forth, and after looking at them all, I still couldn't piece together the full picture. So the first thing I did was cram all my holdings onto a single page:

On top is total asset value and the day's profit/loss. Below, it's divided by market – a column for US stocks, one for Crypto, one for Hong Kong stocks, and one for A-shares. With one glance, I can clearly see the state of my entire portfolio, and who's up or down today.

After I finished it and found it quite useful, I couldn't resist adding one tab after another. As I used it, new needs kept popping up:

• Anomaly Monitoring: I pre-set the assets and thresholds I'm watching. If something suddenly surges or plummets, it highlights it directly, saving me from constantly staring at the charts.
• Investment Map: When researching a particular track, it draws the upstream, downstream, beneficiary assets, risk points, and related assets into a single network, making it easy to trace the capital flow chain and relationship network.
• Memo & Review: I jot down why I was bullish on something at the time, what happened later, and where I was right or wrong. I can look back on these notes anytime.

Because this dashboard contains all my actual holdings, it's quite private, so I deployed it locally.

Second: PM Bet Monitoring

This one is specifically for tracking prediction markets.

To briefly explain, a prediction market (like PM) is where people use real money to bet on whether a future event will happen. The price itself represents the probability the market assigns – for example, if a "Yes" contract for "SpaceX market cap reaches $2 trillion by end of June" is priced at 0.8, it means the market thinks there's an 80% chance it will happen.

For the bets I care about – like "Will OpenAI/Anthropic's valuation go up by year-end?", "Will a market cap overtake event among the Magnificent Seven come true?", "Will X and Y meet?" – I used to have to check them one by one. Now, I centralize them on a dashboard. By placing the probability changes alongside news milestones and secondary market fluctuations, I can see at a glance who moved first and who is driving whom.

I also layered these bets based on my own criteria (I internally call them T1 (High Conviction) / T2 (Relatively Stable) / T3 (Pure Speculation)), sorting them by expected return. This way, I can easily distinguish which ones are just noise.

To be honest, my small edge in this market is Chinese-language information and the political/economic dynamics of East Asia – a lot of this is dominated by Western players, and their pricing of this sector is often a step slow. The opportunity lies in this time lag.

Third: A Small Operations Backend

This one is unrelated to investing. I use it for my writing.

I usually handle topic selection, drafting, and publishing on multiple platforms, relying on my memory and scrolling through chat logs to track progress. It often gets messy, so I built a small backend to manage it. It includes a topic list, article progress, publishing platforms, and an idea box.

Because I might need to use this when I'm out, I didn't deploy it locally. Instead, I put it on the cloud – using GitHub + Vercel. I can open and edit it right from my phone, which is very convenient.

Fourth: A One-Click Formatting Tool

This was mainly to solve a personal need. After writing an article, I need to publish it on several platforms. Web3 media, especially, each have different formatting rules. Manually adjusting it for each one is very time-consuming.

So I created a small tool. Combined with a browser Tampermonkey script I coded up, I can drop in a Markdown or Word draft. It automatically converts it to the required format for each platform and directly inserts images. It's not super advanced, but it saves me some mechanical work every day.

Actually, these four things are still very basic right now. You could even say they're a bit ugly, and not really mature products. But for me, they are incredibly useful. Because as soon as an idea appears, I can get it into my system immediately, instead of letting it dissipate and be forgotten.

This is the most important change I feel.

3. The Way Ordinary People Do Research Has Truly Changed

Because of this, I increasingly feel that ordinary people don't necessarily need to start with complex models when investing. But at the very least, they should have a few basic systems of their own.

Because the impact of AI on ordinary people now isn't about suddenly turning you into an expert. It's about making those things you "wanted to do but couldn't" start as a prototype first.

This is especially clear for someone like me who looks at the markets every day. As long as you have an idea, every ordinary investor can actually slowly accumulate a few of their own core systems:

• Asset Observation System: What assets are you actually tracking? Which market do they belong to? What's been changing recently?
• Signal Monitoring System: What events, if they occur, might signal a shift in market expectations?
• Map Organization System: A track isn't a single point; it's a web. Who is upstream, who is downstream? Who thrives on sentiment, who on fundamentals, who on capital flow? Especially over the past year, stocks in the AI sector have almost exclusively rewarded those who could deeply understand a whole track (from HPC to optical modules to the storage chain).
• 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. They were just too troublesome and hard to stick with. The greatest significance of AI is that it has cut out a huge chunk of that "trouble."

You might not know how to code, but you can describe the requirements, then slowly build up your own product design. And you don't need to finish it all at once. Start with the first version, use it, and improve it as you go.

This is also what attracts me most to Vibe Coding. The feedback loop is so fast. Before, there could be a long gap between an idea popping up and its realization – so long that you might even forget why you wanted to do it in the first place.

Now, if I think of a feature today, I can try it out by the end of the day. If I'm not satisfied after trying, I change it immediately. After using it for two days, a new need emerges, and I iterate further.

This closed loop of "Idea – Realization – Use – Feedback – Re-modification," once it gets spinning, truly makes it hard for you to stop.

Final Thoughts

Consider this the first record of the new chapter for "Taile Tyler."

Going forward, I'll try to update regularly, documenting my investment thoughts, hands-on tool trials, on-chain practices, arbitrage research, as well as some educational/getting-started content on Web3 operations and investment knowledge points.

Feel free to follow and discuss anytime.