The era of personal software is here
Why you should craft a personal AI software to streamline your work, automate the boring stuff, and level up your workflows.
Thanks to AI, coding is more accessible than ever. Before you had to be an engineer to write software that could increase your productivity or improve business inefficiencies.
Now if you have an idea you can get software written for you from plain English using tools like ChatGPT, Claude, and others. There’s coding environments that specialize in this now such as Cursor and Replit. These tools make building your own software simple.
There are lots of off-the-shelf AI solutions that solve all sorts of problems. Unfortunately, these AI agent SaaS products don’t always offer the best integrations into your everyday workflows. We all have our personalized workflows for how we like working. This is good. With coding being more accessible to everyone you can now take your workflows to the next level by crafting software in a way that you think is best.
In the future I believe companies will actually want to see just how efficient you are with your workflows. Personal software enhancing your workflows allows for you to stand out.
Why You Should Build Your Own AI
We all are unique in the way we work. Sure we use a lot of the same tools but we all have our own process for using these tools. Before you used to have to go find a SaaS that matched to your needs. This always involved changing up your workflow just a bit to accommodate that product. In the end most products fell short and were not exactly cheap. If you build your own you get to control your own data, cost, and integrate it directly into your process.
Another reason to consider is eliminating those repetitive tasks that plague your time. Before if you were not en engineer you didn’t really have a lot of options for introducing automation into your workflow. Now you can simply ask ChatGPT to write you a script to automate some annoying task. AI does this sort of thing really well.
In my opinion the most important reason is it helps you stay ahead of in the industry. The whole pitch of AI is it allows us to do more with less. Take advantage of that and explore how that’s true for you as a professional.
How to Get Started
You’re probably sold at this point and wondering how to get started? The good news is there are lots of tools that are really easy to get started with. The two I use the most are Ollama and LM Studio. LM Studio has a great Mac App that I used to work with all sorts of models. It has a great feature set and would be my recommended place to start playing with models.
Models! What models should you use? This really depends a lot on what you are trying to do. Here is a list of ones I have in rotation right now. All of these are free to use, Open Source, and private.
(IBM) Granite 3.1 8B Instruct
(Google) Gemma 2 9B Instruct
(Meta) Llama 3.2 3B
If you get to the point where you are ready to start leveraging AI within your personal software then you will need to make a decision. Do you integrate with the paid models or use the open source models? In my opinion, open source models can get you where you want to go. Especially if you are just building software that you plan to run on your own machine and not share with folks. If you decide you want to share it with folks then you will need to find a way to host the open source model for your users. This can be expensive but so is using the paid models. Pick your poison. The good news is that LM Studio and Ollama both have APIs that are OpenAI like and so you can actually build your model support in a way where you can swap out the model if desired. This is a common pattern in applications now.
Real-World Use Cases
Knowledge retrieval is a great use case for this. We all build up our own understanding of systems over time. Sometimes we forget things and need to search our notes to try and find some keyword. This works sometimes and when it does its magical! That said, we can do better now with AI. We can simply store everything and ask the AI where it is. It does this job well!
If you are an engineer you most likely are already using some sort of AI to help you code now. You can actually download open source models that are specifically crafted to help solve coding problems. I use these all the time for code generation, refactoring, or debugging.
We all have tasks that we do everyday for our jobs. Some tasks are prime for automation and AI to step in and take the wheel. With a little bit of effort you can probably automate those away. For example, every two weeks I host a guild at work. It’s my responsibility to update the meeting notes for this next session, add discussion topics, and let the other members of the guild know that the meeting is coming up and here are the meeting notes if they want to add anything. This process in total takes me ~5 minutes to do. Not long. Unfortunately, the worst part is it takes me out of the flow of what I’m working on when I have to scramble to do it that day. So I automated it! This is just a tiny example but hopefully gets some ideas going for you.
Challenges and Considerations
Depending on where you work they may be more or less open to using AI technologies. If you are at a place that is more strict around data security and privacy then using the popular paid models like ChatGPT or Claude might not be a good idea. In this case, I’d strongly recommend using LM Studio and downloading the open source models. These are completely secure and private because the models run on your machine. They are not hosted anywhere online and because of that you don’t need to worry about the data you are exchanging with them.
Unfortunately, with AI becoming more prevalent it means individuals will begin to rely on it too much. It’s important to remember that AI is simply a tool. This tool does a good job at generating text and now reasoning. This does not mean you should stop thinking. In fact, the opposite. You’re job is to read and analyze what is produced. Remember, these models are now a reflection of you and your thinking so it’s important to make sure they are in line with who you are.
Conclusion
In the end, I believe that with personal software and leveraging open source models we as individuals have incredible power that would have otherwise been impossible up until now. Use these tools to lever up who you are and what you can do. Invest in learning them and understanding them deeply. It will only help you!
Remember to start small here and iterate. Find that one pesky task and try and build some personal software to help you. It will be a bit frustrating but worth it.
Go download some of these tools and get started today!
Let’s connect!
Thanks so much for reading! If you liked this then please let me know down below. I’m also looking to collaborate more with others on Substack. So if that’s you then please reach out to me!