The Shocking Truth About How AI Quietly Took Over My Hustle—and Why It Might Do the Same to Yours

I gotta admit, it’s been a bit of a slow week on the posting front—blame it on the whirlwind of work stuff and a side hustle of vibing with some personal coding projects. Meanwhile, I totally missed riding the semiconductor wave, which is kinda ironic given how much AI is shaking up the entire game right now. So here I am, finally putting some thoughts out there on AI—not as a guru or certified expert (far from it, actually)—just a regular investor and tech enthusiast who’s been poking at ChatGPT, Claude, and Gemini, wondering what the fuss is really about. The big question I keep circling is: Are we ready to rethink how we actually do things—with AI no longer just a tool but almost a teammate? Spoiler alert: I found myself stunned when a bot started cranking out functional code, faster than I could have dreamed. What does this mean for the future of work, investing, and those grunt tasks that we once thought were irreplaceable? Buckle up—this rollercoaster goes deep into how AI isn’t just changing the scoreboard, but the entire playbook. LEARN MORE

img#mv-trellis-img-1::before{padding-top:75%; }img#mv-trellis-img-1{display:block;}img#mv-trellis-img-2::before{padding-top:216.49048625793%; }img#mv-trellis-img-2{display:block;}img#mv-trellis-img-3::before{padding-top:24.043715846995%; }img#mv-trellis-img-3{display:block;}

I have to apologize that my posting have been a little slow this week.

There was a few things at work to get through and if it is not that, I am also vibe coding something personally. I missed a lot of things in the markets, including the semi-conductor run.

I never put out anything about my thoughts about all these artificial intelligence stuff and perhaps it makes sense to put out a short note.

Firstly, I am not the expert here. I did not take up some course on artificial intelligence or large language model (LLM). I also did not do what ChatGPT, Claude or Gemini certification.

Like many of you, I just use the chat function to see if they can help with my work lah.

As an investor, I do observe different stocks benefiting but also disrupted. Initially it was those mega caps such as Microsoft, Nvidia, Meta that show that they might see meaningful effect to near term revenue and profits. Then when ChatGPT starts being more shitty, relative to Alphabet’s Gemini, Google start doing very well, and Microsoft stop doing well and starts shitting.

After this, people realize… oh shit, we need more data center. Then… oh shit we need more memory and storage. Then… oh shit we need more power!

When Claude came up with a certain version, it could actually help people vibe code a good looking, functional application in such a short time, all the software as a service stocks start diving.

The data provider stocks such as SPGI, ICE, LSEG, Factset, Thomson Reuters alsod dived because… their value is not the data but upselling premium analytical.

If you and me can do these premium analytical (we still need your data thank you), then wouldn’t a percentage of their future revenue, and fat margins be cut?

Yeah and that is why there are being repriced downwards.

Our head of tech wanted to sign up for a Claude plan and a pretty optimal one comes with 5 seats. That is too much for them and he graciously asked if it would be helpful for our Solutions team. Between my boss Bryan and myself, he decide to gave to me. Our portfolio manager Glenn got another one.

I just thought that I would “try” when I have the time. I have to get started on a research project. I went to find Glenn a few hours later, and Glenn basically told me Claude didn’t have the struggles of his ChatGPT and Gemini experience.

I decide… why not just asked it do research for a part of the research that I been doing since it is tedious enough. Then it starts fxxking coding out an Interactive HTML.

I was like WOW what the shit.

4 days into using Claude, I just paid 1-year of Claude personally mainly because I can add up in my mind the personal return on investment (ROI).

That Wednesday night, I threw the C# WPF Windows codes of Gilgamesh into a Visual Studio Code project.

Six hours later, I have Gilgamesh on Investment Moats. But it still have errors that are not glaring but critical (more on this later). If we calculate the cost in terms of tokens I spent on this, it cannot be more than $10-$15.

As a child who let the entire Web 2.0 development period bypass me, it will take mentally and physical enormous effort to do this. The reason why Gilgamesh originally exist as a C# WPF project is because with Visual Studio, it allows you to create the user interface by dragging and dropping and then changing in code. We never have this in web development. It is not the back-end part that I fear and its hard for me but the front end and routing.

Even if I can do the backend, I still have to think about how to code it, whether it is possible. With Claude, if it exist, it does it.

And with that introduction here ae some thoughts.

We have to Re-program How We do Things

What is mad about the first ChatGPT was not only the answers.

It’s to introduce us that we don’t need to present things on the screen but the user interface is a chat.

This was what I realized later. They realize a chat can solved many things.

I think it is the same when someone first got acquainted with a spreadsheet like Lotus 123.

You will see people telling us “You need to write your Prompts well!” “That is why Prompt Engineers get paid a lot of money!”. And here I was like… what is a prompt?

It is only later that I figured its the question that you type into the chat. There are always jargons when dealing with new things it seems.

In the short months up to this point, I realized that I have to mentally forced myself to make small meaningful changes to the way I do things.

In the past, its tapping in my brain if I know how to do something. And if I know, how would I do it. If I don’t know I will check with someone else.

It was my head of tech Zheng Rong that kept telling me, you don’t know.. ask it.

I slowly ventured and try it. Instead of immediately telling myself it can be done or cannot be done, I shifted to.. look at this thing as your junior engineer, analyst and asked how can it be done and could you do it.

In fact, I think that is how agents end up.

A agent is like your fellow digital analyst and it does something much better than you. It will do work fast but it may not fulfill all the business requirements because it does not process the context better than you.

And so both of you work together.

Entry level People Have No Value if The Work is Not Physical.

Last time, a lot of us started out doing our level of grunt work in our own field.

Through our experience, being in the environment, we slowly gain experience by absorbing.

A lot of the grunt work is meaningfully reduced.

If we are paying for something, it is really the minimal physical work that needs to be done.

If your work process still involves needing some data entry that cannot be get rid of (because your vendor still does not have an API or a process to connect to), then we still need that.

But the entry level of Kyith’s work is threatened.

Where the Human Level Value is (Currently at this Point)

I have to caveat the last part because… the earlier version can be more stupid.. and now it isn’t.

In your job, if you are what I call a Context Architect or Integrator, you would have some value (at this point).

You are someone who really knows your stuff. In the realm of your domain, you can discern the pros and cons of each stuff, and be able to add them up and structured them well.

For example, you understand that fundamentally what drives market returns, that everything will have to fall back to economic basis, and what drives value. You know what is bullshit but also why something does not matter so much for your strategy.

You will still have value compare to relying on a junior staff with an AI.

The junior analyst can just look at the past investment situations, the data. But he or she is limited by instructions and his or her own experience.

You cannot download the context of things from your mind, developed over so long into their brain. Sometimes not knowing that may be good because they don’t have your bias, but there is a lot of advantages.

In another example, a junior analyst also won’t know why modelling this financial model is useful.

I told enough people that I or Chye Hsern (our head of investments) could not come up with the models that Glenn came up with. Our exposure is different, and Glenn understands why these models would be significant enough for us to consider into the investment review process.

There was a point when I realized Gilgamesh is not producing mathematically the right result so I asked Claude to read back how its programming this PHP version. Then I asked Claude to read how was the C# version is done. Claude basically say they are doing the same thing. Then I have to asked it again and I reflect upon the decision tree. And then I asked again.

The more I asked it, the more I realized “Isn’t the two code decision tree different?”

So I decide to fxxk Claude like how I will fxxk a junior analyst.

Claude basically tell me it was thinking too much now its thinking in a loop. And that I am right.

In a way, I kind of know what should be the right output and was fortunate enough.

The people with more value will be the ones who have enough experience to contextualize across a few sub-verticals AND know how to use AI to value add.

These people can be downright scary now.

Many Problems that usually ends with “If only I could..” or “If only I know…” Can be Solved. That Should Make You Happier.

I already shared about Gilgamesh.

If Gilgamesh is supposed to be a software-as-a-service, it should be a web application. If only I don’t have problems understanding web frameworks, I don’t hate Javascript and I have time to learn them.

I paid for Claude the moment I can add in my mind how possible what I wish to achieve and whether it can help.

But there are a lot of little things.

I always imagined that while the banks became more backwards by closing their API (yes you DBS, OCBC and UOB who often touts your technology sophistication and to this date don’t have an API to get read-only data of our accounts), it will be alright if we can just download CSV credit card transactions and then we have an application to parse it.

By right Claude can help “listen” into the folder and parse it for you.

But you heard of these AI deleting their vital files. No laughing matter if it is work files. One of a community member told us that an AI deleted his work log files.

Now it is possible to code something that resides on your computer, listen to a folder and if there are new files not processed, processed them in a certain way. And you can push how far you want the output.

All you need to do is to put in these folders.

Would solving these “If only I could..” or “If only I know..” make you happy?

Yes!

Getting rid of problems should make you happy and that is what drives my excitement.

It makes me tremendously happy when I consider how viable AI can help solved some potential problem given where it is now.

Gilgamesh

I was damn happy for Gilgamesh even if no one else appreciates it. I think sometimes people have to look from another person perspective why this is such an important topic. Gilgamesh vastly advance how I look at the spectrum of income outcome with different questions such as can a bonds-only portfolio be viable over 20-30 years. Can we inflation-adjust it? In the past, I would be more conservative and say better don’t advise inflation adjustment.

If you triangulate with many different equity indexes, you would realize the magic number is around 2.5-3% SWR and a lot of things will work across different market.

My thinking and therefore my advice cannot be so robust without something like this to help.

{Yet Unnamed Personal Project}

It is not that I don’t have a name but I only told my boss Bryan the name of this personal project.

I hope this can help me productively in a few ways but also answers some questions at work.

I want to build a more enterprise ready web application platform. You can add users and better manage who can access what and who can’t access. A personal project doesn’t have to be this crazy and in the past, scaffolding all these login, permission matrix takes a lot of the work and that may not be where the value (and exciting things!) reside at. Learning how to configure it is another thing.

With Claude this takes place in within an hour.

Having a platform is like a blank canvas that I can decide what to put into it.

You could always ask Claude to think from an Enterprise Solution Architect who has worked in the finance field but then an experience solution architect in the finance field can also deliver shitty processes.

And so the blank canvas is for me to figure out some of the workflow.

The first was translate my FREE Google Stock Portfolio Tracker to app based. You can make so much errors with a spreadsheet and you cannot control the interface. I was able to successfully create the same function, pulling data from Financial Times, Stooq and Yahoo Finance have price data redundancy. It would also grab the forex data daily.

It removes some vigilance needed from me to ensure the Google Spreadsheet is correct.

This week I am reflecting and thinking how we can architect Assets in a Assets – Liability = Net Wealth part. I want to consider how we can keep track the value of an asset, not always the most diligently. And also if we are able to calculate the money-weighted-return (MWRR) on this asset.

How do we link this to an investment portfolio like Daedalus that I just told you is aggregated by transactions. How do I link it to a cash value insurance policy? What happens if I am a lazy adviser and don’t wish to update it? Would the design still work pretty well?

This is essentially a similar concept from this easy way to calculate the value of your dividend portfolio just by looking at cash into the group of accounts and cash out of the group of accounts. (A Very Simple Way to Track Your Portfolio XIRR Investment Performance with My FREE Google Sheet.)

I was happy with this week’s work.

Last time, I would depend on a paid platform and the functionality is always missing something. Then I also hope people start an opensource project to do something similar. This is why I like Actual Budget so much.

Now i have no excuses of making life easier.

Small Problems

Then there are small problems like this. Every time my colleague Yong Cheng give me a cutting (not this one), I always have this mixed feeling of whether I am going to fxxking kill another plant again:

Since the main stuff fallen off (most likely because yours truly was busy with something and did not water it), I wonder how can make it branch off the other side.

Last time, searching Google may not yield the right results. With a smarter AI, it might give me some direction:

Making the plant branch out may not be the most important thing in Kyith’s life. But fxxk it is something I think about sometimes and instead of just keep the problem there, at least i can do something (or maybe it will tell me this is normal).

I can Kind of Get The Fear over these SAAS Businesses.

I can remember discussing with Glenn when these software companies go down before and after we used Claude. We kind of agree upon that what will make these SAAS businesses good is that they can still make you expense these functions instead of setting up yourself and treating this as an expense. You would still want people to take away the maintenance piece of the work.

These business also work so well due to how software sales work, the account management and distribution.

After Claude, we kind of realize that the premium services.. such as analytics offered by any SAAS platform, probably becomes very competitive because their clients may be able to do it on their own. Which means the vendor is competing with the customer.

The customer now won’t do such a bad job.

And so that affects the margins.

Not just that, you cannot charge by per seat anymore, or by per user like Netflix. In the recent ServiceNow earnings call, they say there is a higher percentage of their business is by usage not user.

What this does is… it makes analyst harder to estimate your future cash flow.

We are willing to pay higher valuation for SAAS business because of this ability to map out the charging, annual recurring revenue, what are the cost and be able to systematize the valuation. Now you will ask how is this different than supply cocoa?

Rose Celine Investments have a pretty long but good thread on ServiceNow that explains better than me:

Final thoughts on $NOW

This whole situation around $NOW was a lot to say the least. I posted that I bought it, got hit with a tsunami of toxicity, and honestly it left a bad taste in my mouth. Enough where I said I’d probably never share a buy or sell again.

But then something… https://t.co/sUtxo5PoEC

— Rose Celine Investments 🌹 (@realroseceline) April 26, 2026

I think in this short time, analyst like myself have shown our vulnerability because of the lack of domain knowledge (and I used to work in the IT industry). I used to think that all these AI stuff will be a tailwind for cybersecurity. You would want to be in more control with your cybersecurity.

Now even this has its uncertainty.

There are counter balancing forces in place:

In order to do the really premium work, you may need a very smart AI, and to do that the cost-benefit analysis may break down.
This is an evolving space and what is important is staying ahead in the field of monitoring, security and the service that they provide. Would a small medium company wish for that distraction? Do they have the resources?

Which brings me to a framing about this to help others.

On-premise to Cloud to On-Premise again?

As I said before, we used to need servers in our own office. We get an IT vendor or systems integrator (much like my old company) with the expertise to list to your business requirements, then propose the software and hardware infrastructure to setup in your office.

That comes with potentially teething issues during setup, configuration, and then during maintenance.

When we realize that we don’t have to maintain these, and there is a vendor that can constantly improve their software service, and don’t have to negotiate new contracts and licenses, get people to come down and upgrade, facing anxiety whether shit will cock up again, software as a service looked so appealing.

In a way, being able to code an application, according to your own specification, just like what Kyith did is being hybrid on-premise.

You might still host it online to use compute, so that you don’t have to manage the servers, but the application is within your control.

So you take over the maintenance of your application.

Would you want to do that?

I think not everyone would.

Which is why systems integrator such as my old company, perhaps Accenture, Multichem might benefit, depending on your budget.

Less Spreadsheet and More Little Apps?

I am not sure if this is a trend but one of the reasons spreadsheets are so popular is because they can calculate, can be readily accessible since Google sheets and Microsoft excel can be online, and you can take notes and calculate at the same time.

But they might not always be the most efficient storing data and working with them.

How many times did you feel in the past when using my Google sheet stock portfolio tracker that you wish it could… suggest today’s currency. Or that when you click on Ascendas REIT, it shows you only the past transactions of Ascendas REIT and not others, but also show more info on that?

You make do with my Google sheet, because it is free, and you don’t want the hassle of figuring out those formulas again.

But you would wish for something better.

I think this wave allows you to do that.

This may be good business for hosts like WIX, GoDaddy and Digital Ocean. But in a way, I realize Digital Ocean’s share price is the opposite of WIX and GoDaddy.

AI is Constantly Underestimated by Those who Lack Familiarity.

This is how I feel at this point.

I looked into people’s eyes, when we discussed the topic and you can tell. It is the same feeling when I see people talk and emote about certain stocks, or different verticals of finance. You can feel the depth they have went into.

I think if you are not hovering between being excited and fearful, you are not far down enough.

Our CEO Chris would say that the potential is to automate a lot of parts of the financial planning and leaving the conversation and guiding clients to the human adviser.

As I went down deep enough, I am not so sure.

My colleague Lena gave a lesson/sharing/don’t-know-what-the-fxxk-you-call-it about the psychology of clients to our advisers. I did not attend it but since Lena have always shown interest in what I shared, and I always shown interest in what she shared (perhaps that is the key ingredient to being a solutions specialist since she used to be the solutions specialist at MoneyOwl), I took a look at the slides and I was rather intrigued by the 4 personas of clients:

The Passive Preserver: Driven by a deep need for safety and family security, the Passive Preserver sees money as a shield, not a tool for growth. Their wealth is often inherited or accumulated through steady employment — making them cautious custodians rather than active builders. They would rather leave money on the table than risk losing any of it.
The Friendly Follower: The Friendly Follower is socially intelligent and deeply relationship-driven — but this strength becomes a financial liability when investment decisions are outsourced to social consensus. They want to belong, not be left behind. Trend-chasing and FOMO quietly erode their portfolio while they remain convinced they’re being prudent.
The Independent Individualist: Intellectually rigorous, self-reliant, and quietly proud of being contrarian, the Independent Individualist has done the work — and they know it. The challenge is not their intelligence; it is their resistance to information that challenges their existing conclusions. They are at once the easiest and hardest client: easy to engage intellectually, hard to actually influence.
Active Accumulator: The Active Accumulator has built wealth by betting on themselves — and won. This creates a dangerous cognitive shortcut: the belief that the same bold instinct that built the business will reliably build the portfolio. They are fast, decisive, high-energy, and chronically overexposed to their own conviction. Their greatest risk is not the market — it is themselves.

Not sure who do you pigeon hole yourself closer to.

I think the Independent Individualist describes me well.

As I was reading, a thought triggered within me: Why not build on this and ask an LLM:

What would a person with this profile often say that pigeon them more to this group?
What are the deep mindsets that plague them, making them this way, and what are the root causes that create these deep minsets?
How does an adviser help untangle this for the client?
What are the right approach or influence levers?
What will make these people zoned out?
How do we handle disagreement on investment planning?
If we revisit their plan in 3-5 years (if they are not fully onboard), what do we tell them?
What do most people don’t know about this psyche?

And what the LLM came up with… actually describes my persona much better than anyone could.

For example in what makes Independent Individualist zoned out:

Generic presentations, boilerplate investment frameworks, and any conversation that feels like it’s been given to a thousand other clients. They will mentally exit the moment they feel like a standardized portfolio is being sold to them.

Also: excessive deference. If you agree with everything they say too quickly, they lose respect for you as an intellectual sparring partner. They want a worthy counterpart — not a yes-person.

What you should avoid when talking to such a person:

Generic advice or standard templates
Appealing to authority (“most experts say…”)
Directly challenging their research in front of them
Over-praising their approach (they’ll see through it)
Trying to change their mind in one meeting

What you should do when talking to such a person:

Ask them to walk you through their thesis — then probe with precise questions
Introduce disconfirming data as “something you found interesting”
Frame portfolio construction as “optimizing their thesis expression”
Position diversification as risk management for their best ideas, not dilution
Appeal to intellectual curiosity — “Have you looked at X? It’s counterintuitive…”

The deep mindset that make them this way:

Intellectual identity: Their self-concept is built on being a sharp, independent thinker. Investing is one arena where they express this identity. Accepting advice threatens that self-image.

Contrarianism as a strategy: They have often been rewarded for being different. This makes them susceptible to the belief that any popular view is automatically suspect — even when the popular view happens to be correct.

Conservatism bias: Once they’ve committed to a position, new contradicting evidence is processed as noise rather than signal. They hold on not just to investments, but to the thesis that justified the investment.

Selective research: Their research process, while genuine, tends to surface information that confirms existing hypotheses. They are not lazy — they are strategically thorough in one direction.

This whole profile basically describes me in a way that even I would not be able to. And if you talk about how to handle a “client” like, this is the way. You praise me too much, I start feeling very negative instead about it. If you present a well thought out opposing view, I leave the meeting with you in disagreement but respecting you more (and that is important for building relationship)

I sent it internally to a few people. Seems they weren’t too interested. I guess sometimes it is how close you are from needing things like this in your work.

Experiences such as this makes me excited about improving things but also make me consider how useful is the human adviser.

In my structured mind, almost everything can be decision-treed or derive from a framework. You think about it, where does all these planning if-else comes from? How do I answered all the variation questions? It is a massive and complex if-else right?

If so why can’t an AI handle that?

Can we have a camera to process doubts of those clients in the meeting? I think the machine can detect it better than humans.

If you cannot connect the dots between application, planning structures, you cannot really see the real picture.

We Felt a Sense of Relieved that We are More Financially Secured

I think the more I showed people what can be done, everyone eventually thought of whether their kids will have jobs or what kind of jobs.

None of the technological disruption is as mad as this one.

If you ask me, I also have no idea.

It used to be that if you preserved and build a domain, you are valuable and people will pay a good compensation.

Now we don’t know already.

There will be Other Jobs

I would always say that many of us plan as if the current world won’t change.

And so you could always Barista FI.

But now would there be so many people looking to be baristas that it is not so easy for you to find a job?

If there is mass unemployment, what will happen to your Barista dream?

The Big Hit this Time Might be the High Compensation Jobs instead of Low Value Jobs.

The struggle used to be that your skill set is not relevant anymore and you would struggle to find a job.

Or that your job “does not add so much value” and can be easily replaced, especially if you are older and you are paid more.

Well this time, what if the pool of people needed to do high level jobs are lesser and lesser.

It is not just offshoring, ageism, but its your slower and stupider than your competition.

Would there Be a Mortgage Crisis?

If those that currently earn a high compensation are paying a lot on private condos, and while TDSR have made it more prudent, what happens if many lose their job and cannot find as high compensation job anymore?

I think it is the degree of the unemployment in this space and how all these AI thing evolved.

Epilogue – Financial Independent Life is More Interesting

So that’s how I look at AI now.

One of my Telegram group members commented that people won’t be so bored because there are a lot of things to play with.

In a way its right.

Earlier, I said to frame how you communicate with AI as like an intern.

In a way, its like having an experience companion at home.

It depends on what you are interested in. If you find that you wish to automate some of your life, your LLM is like a season solutions architect.

If you are interested in making your garden look better, then the LLM is a more experienced person that you can “talked” to about your progress and improve.

So do we need human interaction?

That is a question for you and me to find out.

Do Like Me on Facebook. I share some tidbits that are not on the blog post there often. You can also choose to subscribe to my content via the email below.

I break down my resources according to these topics:

All my personal notes about how my philosophy behind my own money and how I manage it.
Building Your Wealth Foundation – If you know and apply these simple financial concepts, your long term wealth should be pretty well managed. Find out what they are
Active Investing – For active stock investors. My deeper thoughts from my stock investing experience
Learning about REITs – My Free “Course” on REIT Investing for Beginners and Seasoned Investors
Dividend Stock Tracker – Track all the common 4-10% yielding dividend stocks in SG
Free Stock Portfolio Tracking Google Sheets that many love
Retirement Planning, Financial Independence and Spending down money – My deep dive into how much you need to achieve these, and the different ways you can be financially free
Providend – Where I used to work doing research. Fee-Only Advisory. No Commissions. Financial Independence Advisers and Retirement Specialists. No charge for the first meeting to understand how it works
Havend – Where I currently work. We wish to deliver commission-based insurance advice in a better way.

Kyith is the Owner and Sole Writer behind Investment Moats. Readers tune in to Investment Moats to learn and build stronger, firmer wealth foundations, how to have a Passive investment strategy, know more about investing in REITs and the nuts and bolts of Active Investing.

Readers also follow Kyith to learn how to plan well for Financial Security and Financial Independence.

Kyith worked as an IT operations engineer from 2004 to 2019. Currently, he works as a Senior Solutions Specialist in Fee-only Wealth Advisory Firm Providend. All opinions on Investment Moats are his own and does not represent the views of Providend.

You can view Kyith’s current portfolio here, which uses his Free Google Stock Portfolio Tracker.

His investment broker of choice is Interactive Brokers, which allows him to invest in securities from different exchanges all over the world, at very low commission rates, without custodian fees, near spot currency rates.

You can read more about Kyith here.