One of the greatest challenges of being a Solopreneur is that you’re doing it alone. It’s thrilling to be able to go at your pace and be completely independent. That’s great, until you get stuck and need help!
That’s when most of us turned to Google, StackOverflow, Reddit, YouTube, countless blogs and forums, etc. Finding the golden egg in the random corners of the Internet is a time consuming and often frustrating experience. It also interrupts your flow state. More often than not, the exact answer you’re looking for doesn’t exist and piecing bits of information through trial and error can be painful. In the end, you may find the answers, copy-pasted your way to a working solution and moved on. But whether you actually learned anything is questionable.
With the fast maturing AI Coding Assistants, you can not only get contextually relevant guidance, have an in-IDE developer experience, but also a way to dig deeper, engage in Q&A and make fast progress while actually learning what you may not have known. You no longer have to just rely on hunting and pecking across the Internet!
Securing your app is obviously a critical feature and important to get right. But security can also be a complicated subject to understand and technology to implement. There are many options available for enabling authentication / authorization functionality in applications and a growing set of authentication providers, technologies, language-specific frameworks, etc.. Social logins and 3rd party authentication providers are en vogue and a great way to reduce the barrier to entry / adoption for end user signups. Implementing them though is tricky.
I got stuck with OAuth2 and needed a way to get past my issue. So I asked my local-running AI Coding Assistant to analyze my code, explain the concepts, create a Mermaid flow diagram and document it for me in Markdown. This helped me understand it, create my own reference information, build my solution and move forward feeling a sense of accomplishment.
With LLMs these days, you have plenty of options. Perhaps too many options. Making a choice between the sheer number of models (and the rapid pace of change among them), the many open source and proprietary offerings, the differentiated capabilities (some are better at certain languages, for example) and the various developer experiences can be paralyzing.
At work, I use GitHub Co-Pilot within VS Code. It’s a great choice in general and especially for enterprises.
I prefer running everything locally and using open source models. Being portable and having access to everything I need on the go or when wifi is scarce is an advantage I don’t want to sacrifice. These days, the open source models are highly comparable and competitive to their proprietary counterparts. You’ll still need to do some research, get to know your models, optimize for your specific workflow, etc. Once you’ve made those choices and get going, it’s a total productivity unlock!
At home, my setup is: