Learning to Code in the age of AI

In 2026 the majority of programming content is focused on the existential nature of the software developer. In fact we’ve seen headlines from major companies that the developer will be no more in the near future. But the second I started prototyping apps with Claude, it was not enough, there was an intense itch to understand how Claude had produced it. I decided I wanted to learn how to code, but could not foresee the unique mental battles that would come with it in 2026.

This week I finally had a breakthrough with understanding hooks in REACT. These aha moments are killer as they feel like you unlock a new weapon you can wield in an RPG. These moments are incredibly positive, but for me they are usually immediately overshadowed by thoughts of the impending AI jobpocalypse. There is an ongoing battle in my head over whether or not it is worth it to keep learning programming. The wins stoke the fire, but the negative sentiment on the future of software development can suffocate my interest.

I tried completely staying off media platforms to shelter my self from this negativity but that proved to only isolate my self from reality. Especially since I gain so much from interacting with other developers at work or online, the topic of AI induced job loss will always come up eventually, and it needs to be dealt with.

In fact, this is a great opportunity to be a light in the darkness. Regardless of what the AI corporate powers say, you as a software developer are not obsolete, quite the contrary. Programmers need to hear this message on a daily basis because the AI CEOs are not giving us the truth. Whether they are doing it on purpose or not, the general message of AI impact is overblown.

Why can I say this confidently? Since I manage IT projects, and interact with “non-technical” people on a daily basis, I see the fear in the eyes of those are asked to use AI for the first time to do something. People who have never seen a <div> in the wild will not build a website, they are not going to prototype something, in fact they will indeed stick to their lane using AI as a search engine.

We are also told programmers are more prolific with AI tools and therefore we need less of them to do the same work, and this is absolutely true. However, the effect of that was supposed to be increasing a team’s productivity across the board, not slashing a quarter of your workforce to maintain current levels of productivity with fewer people.

As a business owner, this behaviour looks familiar, downsizing when times are tough is necessary, otherwise the company dies. But this is not being sold as downsizing, every layoff is being preceded with the CEO citing AI productivity as the reasoning, not financial difficulties or over-hiring.

You would think investors would understand that the messaging is not inline with reality, companies leading the charge with layoffs are not revolutionizing anything, in fact in many cases their products and services are getting worse. Take the recent CoinBase layoffs as an example, they had the confidence to say “non-technical teams are shipping code” and then in comedic fashion experience a devastating outage three days later, even their status page went down. The cited reason was an infrastructure failure at an AWS center, but the proximity of this event to the CEOs tweet creates unfavourable optics to say the least.

I could be completely off-base with all this, but what is driving this line of thinking is how much better my AI experience gets the more technical I become, not the opposite. And what is crazy about this is I am trying hard to limit my AI interactivity while learning! If I am absolutely stuck I use AI to act as a teacher by not giving me the answer, but giving me hints instead to guide me to the right place. But as a rule I’m figuring out that using less AI to learn, makes AI more useful to me, and me more useful to the world.

Once I fully understand a topic, and have put in a significant amount of practice, that is when AI works magic for me because I can give it proper context to return something clean. Otherwise the guesswork the AI has to perform with a non-technical prompt leads to a result that does not fit the use case.

I am trying SO hard not to use autocomplete in VSCode, it temps my flesh constantly, but the more I force myself to physically type out the code, the better my retention is. My next step is to turn autocomplete off entirely, but baby steps my friends…

Regardless of a major market correction or an AI jobpocalypse, what stands out to me is the need for knowledgable programmers. I have been amazed with the level of complexity in this field, and what AI usage shows me is that the more this tooling progresses, the more knowledgeable professionals we need to implement it safely and efficiently. For the record I am officially labeling myself as bullish on software development. If I have to retract this later it will be humbling, but worth it.

Also, I published a song to Youtube recently. I had my brother produce it and add some gritty guitar riffs. I really like infusing industrial rock with electronic music. You can check it out here.

I would normally post the track at the top for you to listen while reading, but this track has vocals, so it may be distracting.

Thanks bros, God bless. Jesus loves you. Keep learning!