All In on AI

2026-03-19

Why I’m All-In on AI Now

Over the past month, something has shifted.

Since being an early adopter, I've found AI useful and a key component of daily life. From brainstorming, Teacher, internet searching, coding mentor, writing editor and artist its been everything.

I've always held something back not wanting to jump in the deep end without something to float on.

But now I’ve gone from experimenting with AI…

To its core to everything, I do.

I think we've hit a point of no return, and if you don't get on board your going to be left behind.

It's The kind of shift where you stop asking “is this useful?” and start asking “how far can this actually go?”

The Shift

For the longest time, AI was an enhancement layer. It made things faster. Helpful, impressive and occasionaly game-changing. It was something you used but did not change how you worked.

That's no longer the case.

I held off from using AI in my coding space, whilst I could not decide which to engage learning in, I was also very concious on building my skills and not skipping this very important stage of becoming a developer. AI helped me learn on the job of developing, not doing everything (just almost everything).

After experimenting with Lovable and Codex, I'm now ready to hand off coding to the AI full time.

The shift Im seeing with the rise of agents is:

Tools → Systems Prompts → Processes Assistance → Execution

The leverage now is:

• Structured workflows
• Defined responsibilities
• Repeatable systems

AI is not just interesting anymore, it's useful.

Reality Check

Last weekend, I attended AgentCamp in Christchurch. My first tech conference.

I expected to feel behind.

Instead, I realised something surprising:

Everyone is still figuring it out.

Even the people building in this space every day are experimenting, iterating, and learning in real time.

We are super early, but things are about to change at break neck speed.

There has been a lot of doom and gloom about AI taking jobs, and this is starting to exponentially ramp up, especially in Software Development. There isn't a week go by without a big tech company firing in the thousands.

However, we are already seeing change it's great at writing code but it needs management, it needs oversight and this requires new skillsets.

The small bubble tech exists in, everyone is fighting for scraps but that bubble is going to expand far outside technology into every industry.

Business has and always will succumb to the most effective strategy. We will see a new economy and new ways of working. Its important to be a part of that change, help shape and mold it.

The game changer: OpenClaw

A lot has been said about OpenClaw but there is no denying its revolutionary impact.

Following this development for a couple of months, allowed me time to process and once Nat Eliason developed FelixCraft I saw the winds had truly shifted.

I've been experienting with OpenClaw, not just prompting (in fact I've still to really have a conversation with an LLM yet) I'm trying to build a system.

No Demos, No Outputs. Actual Work.

It is very frustating and not an easy learn especially something so early in its infancy and changing daily.

Its been a great learning experience and understanding of:

• structure
• memory
• defined steps
• clear intent

Because without those it falls apart very quickly.

Where are we going

AI is not slowing down, the bubble is not about to burst, and it's not staying in the tech bubble.

Anywhere there is:

• repeatable work
• admin overhead
• decision making
• money to be made

AI will fit.

A common use case I'm seeing it being applied is Sales + Marketing. A major bottleneck for most businesses is simple: How many leads can we interact with at once

That's limited by team size held back by how big their team is. When you can move at the scale and speed agents can, those that move fastest will win.

Closing

I'm not chasing trends I always think practically how can this solve a problem.

Right now I'm focused on:

• Getting my openclaw useful day-to-day, enhancing my ability to perform multiple tasks at once allowing me time to focus on ideas and shipping end product.
• Building workflow where AI handles clearly defined tasks
• A real environment applying it to real projects not fancy mvps

More trial, iteration, real usage because thats the only way to figure out what actually works.

Businesses will require people who understand agents, the skillsets are not naturally occuring. Whilst development jobs will be replaced management of agents will become a new opportunity.

And people who understand it will move faster, build better systems and operate differently.

So for now, I'm all in.

Not because I have it figured out.

But because I already have a head start in the race, and I don't want to catch up later.