Theo Nga

Tech Lead during the day 💼, Founder at night 🔥

[Theo Nga @theo_nga]

Hey X vibecoders, how do I save my Claude code tokens?

Genuinely, without degrading the quality of the output.

[Theo Nga @theo_nga]

@13_narcissus Exactly, that's the point.

[Theo Nga @theo_nga]

The most lethal developer combination..

gh, aws cli, gcloud cli, claude code

[Theo Nga @theo_nga]

@anupamrjp What are you talking about? I gave AI access to my AWS and Google Cloud via their CLI clients. 🎤

[Theo Nga @theo_nga]

@shub0414 Comrade Opus 5 in Claude Code.

[Theo Nga @theo_nga]

Startup Metrics Every Founder Should Track

[Theo Nga @theo_nga]

I gave Claude agents full access to my AWS CLI, and Google Cloud CLI.

What can possibly go wrong?.

[Theo Nga @theo_nga]

Today I'll be coding by hand, wish me well 😀

[Theo Nga @theo_nga]

Is this true?

There are no BAD ideas, just poor execution and marketing.

[Theo Nga @theo_nga]

Startup Metrics Every Founder Should Track 🔥🔥

For startup founders, these are some of the most important numbers to constantly watch:

- Customer Acquisition Cost: How much does it cost to acquire a customer?

- Customer Acquisition Cost Payback Period: How quickly do you recover that cost?

- Customer Lifetime Value: How much profit does a customer generate over their relationship with you?

- Conversion Rate: How well are you turning traffic into customers?

- Time To Value: How quickly does a customer experience your product's value?

- Average Revenue Per User: How much revenue does each customer generate?

- Logo Churn & Revenue Churn: How many customers and how much revenue are you losing?

The metrics are connected:

- Better Conversion Rate → Lower Customer Acquisition Cost → Faster Payback → Higher Customer Lifetime Value → More Sustainable Growth.

On the other side:

- High Time To Value → Higher Churn → Lower Customer Lifetime Value → Worse Unit Economics.

Don't just track growth. Track whether your growth is healthy and sustainable.

[Theo Nga @theo_nga]

How will AI change e-commerce?

Understanding Buyer Intent, Simplifying Product Discoverability
AI will make it much easier to find the right product.

For example, a buyer could simply say:

“I need a laptop that is good for AutoCAD. My budget is $450.”

AI can understand what the buyer needs and recommend the best options based on their budget and requirements.

Finding the right product is still a hassle on most e-commerce platforms today.

You search, click through different pages, apply filters, and keep changing your search. All of this is just to try to tell the platform what you want.

And the worst part is that buyers don't always know what they want.

Yet, e-commerce platforms expect them to figure it out.

For example, if you are buying a phone, you might have to choose the processor, RAM, battery size, storage, and so on.

But not every customer understands what these things mean.

You might know that you want a fast phone with good battery life, but you may not know which processor or how much RAM you actually need.

AI changes this.

AI will become the shopping interface. People will shop the same way they interact with AI chatbots.

They will simply tell AI what they need, and AI will help them make the right choice.

What will happen to traditional e-commerce platforms?
Traditional e-commerce platforms will mainly become data sources. Unless they evolve and start integrating AI that promotes good discoverability of products.

For example, imagine you want to buy a laptop. You know what you want to use it for, but you don't know exactly which laptop you should buy.

You can go to Claude or ChatGPT and explain what you need and why you need it.

AI can then look at the available options, recommend the laptop that best fits your needs, and explain why.

You then go and buy that laptop.

In developed countries, I think this will eventually happen directly through AI. The user may simply click a button, enter their payment details, and that's it.

However, for us in developing countries, there is still a major barrier.

The local e-commerce market still has an important role to play. Especially if they have the whole e-commerce stack. Stock management, Distribution, Warehousing. etc. Now, that's something AI can't do. But if we only have listing platforms, with no delivery facilities. Then definitely, AI can dilute the role and take over the market.

Also, the biggest barrier in Zimbabwe is TRUST.

Zimbabweans don't yet fully trust buying things online. This is a changing trend; however, people are starting to trust buying online. However, the majority don't like PAYING for goods online. They prefer CASH ON DELIVERY. Unless they truly trust you.

And by buying, I mean ordering something, paying for it, and trusting it will actually be delivered to you.

AI cannot solve that on its own.

[Theo Nga @theo_nga]

@fullstacklover Don't get me started. With Bucky Roberts

[Theo Nga @theo_nga]

Used to tell AI what to do.

Now I ask it what to do next 🤡🤡

[Theo Nga @theo_nga]

@saranshhx AI & Robotics Together. Tesla Optimus. Then Quantum computers.

[Theo Nga @theo_nga]

If RAM is so fast, why not store everything in RAM and do away with storage? 😌

[Theo Nga @theo_nga]

@Tech_p001 Marketing and public speaking.

[Theo Nga @theo_nga]

If you were to build an AI first e-commerce platform, what would it look like.

[Theo Nga @theo_nga]

With the way people are vibe coding stuff and moving too fast 🔥🔥

Don't you think CyberSecurity is going to be a very high demand skill ?.

[Theo Nga @theo_nga]

@EOEboh And taste

[Theo Nga @theo_nga]

I don't know if this is the best approach with AI coming in hot. No human will be a better coder than AI. However, humans will always be better engineers, at least for now. So the question is how to be a better engineer.

My answer is: know the basics: different types of programming languages, databases, Data structures, Servers, Linux. Understand infrastructure, servers, serverless stuff, databases, common scaling issues, security measures, etc.

And most importantly, getting out of the "I'm a developer" mentality, and talking to real people, customers, leaders, etc and understanding problems deeply before the actual coding begins.

And yes, I agree with you. DOING and BUILDING is the only way of learning all those things.