Plans for 2026 and thoughts on AI
Goals for 2026
- ✅ Launch a pro plan for my side project JoshBeeTCG.com (2026/01/03)
- ⬜ Learn more advanced algorithms for image processing and edge detection
- ⬜ Develop my own AI first workflow
- ⬜ Do a tech talk.
- ⬜ Finish a sprint triathlon in under 2 hrs.
- ⬜ Have an adventure and travel to a place I have never been to before.
Thoughts on AI
I have some really mixed thoughts on AI. On one hand it uses a ton of electricity, we give up control, and using it too much may degrade our technical and cognitive skills.
But on the other hand just ignoring it feels like it could be setting back my long term career as a software developer. As someone who has always struggled with giving up control and feeling like I need to be a perfectionist it has been hard to fully embrace AI development.
I have been trying to consume some more long term content on some of the more technical users of AI, and I think this video with Boris Cherny was really good on his overall career and his work on building Claude Code. (1) One of the biggest takeaways I had was when he mentioned coding with AI he still makes sure that the code quality is on the same level of a human developer. This is a really important point and its something I have really focused on as well since I started using Claude Code this year. I typically plan and work in smaller chunks, make sure to review and correct any code that looks out of place, and then commit it before continuing onto the next task.
There are small 'experiments' that I will sometimes vibe code without caring about code quality and then redo in a manner that is fitting for a production environment. Right now there is a lot of content on the internet on how you can vibe code a full project in a weekend or generate a fully UI in a single prompt and it had always made me feel like I wasn't properly using AI. This is something that I feel is more on the hype and marketing side of AI and not a true representation of building an excellent product. To that end I plan on pretty much eliminating almost all content related to vibe coding from any type of news feed I consume.
Another small word of wisdom from Boris was when he was asked about how he sees Claude Code compared to some of the other competitors in the agentic coding space, and his response was basically that he doesn't and his one goal is focusing on improving Claude Code regardless of what anyone else does. From this point of view he is:
- Focusing on what he can control
- Guiding product direction based on actual user feedback
- Not spending any energy thinking about what the competition is doing
- Coming up with more creative solutions than just copying some existing solutions just to hit feature parity.
While this applies to more than just the AI field, I think this does directly apply to my perception on what other developers are doing, and to not be looking at the greener grass on the other side. I consider vibe coding analgous to low code platforms, it works great for small things but for any serious project you will exceed the capabilities of what the AI models are capable of.
The noise is loud maybe its time to ignore it
If you have been on the internet for the last year there is tons of speculation around AI. We may be in a bubble, we may not. We may be right around the corner from AGI, we may not. AGI is going to replace all of our jobs, or it may not. Its honestly a bit exhausting listening to all the hype and content being pushed out. However if you watch interviews with some of the top AI pioneers like Ilya Sutskever or Demis Hassabis it sounds like we are still around 5-10 years away from a true AGI, and even then it will still take years for companies to fully adopt a next generation of AI. Compared to influencers who are generating slop, click bate headlines, and only posting hype, these leaders have a much more grounded approach when you listen to them speak. Yes, an AGI could fully upend the current economy and how we work, but today is not the day to be worrying about something we can't control that may be a decade away.
What about cognitive decline?
This is also another worry I have about AI and its potential detrimental effect on how we are able to think and reason about code. I can go back and forth on this, but I have come to my own personal conclusion that unless you want to be a code monkey for the rest of your career, you just need to embrace AI as a higher level language. The same could be argued about any existing higher level language, and yes we will lose some of our technical capabilities, but this happens when any new language comes out. If the new languages weren't overall superior we would still be coding in assembly (the same language used to send men to the moon).
When viewed in the context of just a higher level language there is no reason to not try and embrace it. While yes some technical skills will atrophy, this leads to opening up our time to focus on a more systems level thinking and how to review code. This is more like taking an a roll of a principal engineer who is more focused on system level thinking and managing a team when compared to someone who may just want to code at a junior to mid level. My one requirement of spending more time on AI development is not letting code quality decrease, so whatever workflow I develop has to maintain the highest level of code quality.
To this end I want to spend some time in the new year focusing on building up a more AI first workflow. One that can grow as the models improve over the next 6-12 months. This will give a chance to experiment and tinker with new tools to figure out what works and what doesn't.
Keeping the brain sharp
Outside of AI coding there are still areas that can be focused on to keep our brains sharp and growing. The first one on my list is blogging. Writing about what I have learned about using AI is no different than writing about tips and tricks about programming languages and design patterns, its just a different topic. Other things that I want to spend more time on:
- Focus on healthy eating and getting in better shape for my overall health and enjoyment or sports like rock climbing and triathlons.
- Getting out of my comfort zone in terms of events and activities that I wouldn't normally do.
- Spend time working on coding problems and optimizations that are more complicated that simple CRUD applications
- Listening and reading material coming from thought leaders in fields and hobbies I am interested in.
My overall goal for 2026 is to try and embrace how AI can improve my life, instead of worrying about the downsides of how it may be detrimental. The worst that happens is I have to go back to writing code by hand which doesn't sound too bad when I think about it. Outside of coding I don't really see myself heavily using AI.