I have recently started to learn the programming language Python. In the middle of 2024, I needed something to rejuvenate my interests in technology other than games, which I do not have a lot of available time for these days due to having two small children. 

At work throughout the year prior to then, I had been required to modify and program some old Cisco layer 2 and layer 3 switches. At first this was a daunting and begrudging task as I had taken part in some of the Cisco CCNA course many years ago and did not enjoy it. At the time I could not understand why you would want to configure a switch by command line when a GUI (Graphical User Interface) would be “easier”. 

Fast forward to the present day and in a lot of cases I would much prefer the CLI (Command Line Interface) to the GUI. It’s faster and usually less complex to look at and understand than having vague icons all over the place and not be able to fully rely on the GUI working properly or as intended. 

Anyway, I actually really enjoyed recalling and using the buried knowledge I had. I had to look up and remind myself of quite a few of the commands with the help of Microsoft’s recently released (at the time) Copilot AI powered by Chat-GPT. It’s been a massive help the past couple of years! 

This all made me realise that I might enjoy programming, where previously I believed I would hate it. 

Whilst browsing TechPowerUp.com I saw an advert for 1000 courses for next to no money from StackSocial (“EDU Unlimited by StackSkills: Lifetime Access”). I looked at the contents and there were a number of programming courses so I thought it would be worth a punt as it was so cheap. I then started researching the “best” programming languages for beginners. I was aware of languages such as C, C# and Java but I had no real knowledge of them. I then discovered that one of the languages that is easiest to pickup and understand the syntax is Python.  

Then I thought I should look up the best Python programming course and came across many, many users recommending the “100 Days of Code: The Complete Python Pro Bootcamp” by Dr Angela Yu / London App Brewery. I checked it out and it really did look like a great course. I had previously tried some other courses on Udemy and they were simply screen recordings with voiceovers and no practical components. But, this course was exactly what was needed. Video-based theory lessons, followed up with quizzes, tests and practical programming exercises. 

You can take the course at your own pace It was priced at £60 on the Udemy site but on their own website I managed to get it for just £15 and it still runs off the Udemy site anyway. 

I’ve just started the Intermediate section of the course which is now switching to object-orientated programming. Up to this point I have created a few basic games and programs as part of the course e.g. Blackjack; Rock, Paper, Scissors; Password Generator; Number Guessing Game etc. 

I’m really enjoying it as it’s a fun new challenge with a number of head scratches along the way. 

One of the biggest helps and in some ways it could be considered cheating, is the access to Copilot. I try to only use it when I am really stuck just to nudge me in the right direction. I can’t seem to get the method to call information from lists or dictionaries to stay in my head, however, I have used to it do some more complex tasks that I have not learned yet. 

For example, I wanted to create a game launcher to allow a player to select from the different games I had created so far. I started writing out the code myself but I needed a little assistance especially when trying to use Pyinstaller to create a standalone .exe file. It took a while for Copilot to understand what I meant by standalone as it kept giving me methods that would still require the .exe to depend on specific external folders and files. Eventually I got it working and now I have the method I can easily repeat it. 

My end goal is to shift my job role at work to something programming based (assuming I still enjoy it by the time I complete the course!). When Deepseek hit the news the other week, it led me down the rabbit hole of running AI locally rather than using the cloud-based solutions such as Copilot, Chat-GPT etc. 

I installed Ollama on my laptop (along with WSL2, Docker, OpenWebUI) and tried the small LLMs (Large Language Model) such as Deepseek-R1:7b, Llama3.1:8b, Mistral:7b and Qwen2.5-coder:7b 

They are fun to play around with but really slow on the laptop. We do have some older rack servers with 72 logical cores each and 384GB RAM, due to decommissioned this year, that I’m hoping could be re-used. I’m hoping these could be used to run the smaller LLMs at a decent rate without a GPU, at least for testing purposes.