There’s been a substantial amount of discussion lately surrounding the capabilities of Large Language Models (LLMs), with debate swinging between praise and criticism. Navigating through this raft of information to identify the genuinely useful applications can be somewhat challenging.
I’m certainly not championing the use of LLMs to create content with minimal input other than a brief prompt. Asking an LLM to “write me a blog post on X” is a low-value request that serves little purpose, and it’s relatively easy to detect. As I’ve mentioned in the past, I sometimes employ LLMs to enhance the spelling, grammar, and structure of some of my posts.
Another beneficial use I’ve found for these models is their assistance with coding brief snippets in C#. My work often requires minor adjustments within my demo environment, and our demo tools facilitate the addition of functionalities through C# scripts. While I can write C# code, it’s not my speciality, given that I’m more of a ‘citizen developer’ than a dedicated programmer. Due to my infrequent coding activities, I often need to refresh my memory on certain coding procedures.
Recently, I had to identify potential dates embedded within a larger body of text. My prompt to the LLM was straightforward: ‘write me C# code to find any date from a string. The date format should be ‘6 Apr 2023′, and an example string would be ‘NexusBond Stickystik 643X Epoxy Adhesive Required By: 6 Apr 2023.’
The LLM provided a method and a usage example, which I could implement within minutes.
Without the LLM’s assistance, the task would have consumed significantly more time.