2025-06-21
The illusion of competency
reading now kept times
Forced to use AI at work, I crafted an end-to-end solution: a boost of confidence, and the caution that it may be an illusion of competency.
log2 min read
I was forced to use AI at work recently and throughout the whole week, I ruthlessly crafted a business solution for our company's consolidation of contract negotiation records into a single spreadsheet as a "working database," while also leveraging and standardizing future documentation of client negotiations. As of now, the main problem was in consolidating these negotiation records as learning tools for our inhouse recruiters. This was a task that was requested to me by our company's recruitment manager.
I meant that I was "forced" because I was immediately thrown into working with our Business Improvement Director since he was already building a custom GPT for our recruiters, which will act both as a learning tool and an assistant for anything inhouse recruitment-related questions. To be honest, I find counterintuitive, reducing my confidence in our company's hiring strategies, where it seems like we have incompetent recruiters. On the other hand, I would also argue that our company's business model does not operate similarly with your common recruitment firms. But to go back to the topic at hand, besides the task request that were asked of me, I was also involved in training the GPT by creating a file that can be used to train it.
By Friday, I presented my solution to my immediate supervisor and she admitted that she was impressed that I was able to devise an end-to-end solution that merges both of my seemingly unrelated task that just came about last week. I think this experience gave me a different perspective of the use of AI tools in allowing me to create a functional and practical solution to a problem despite my lack of knowledge and experience in creating datasets, workflows, and analyzing data to such an extent that would make it robust and withstand data rot using simple tools like a spreadsheet, a text file, and a vibe-coded script to automate the values in the spreadsheet including metadata tags and automated keyword search.
This made me reflect because it gave me a boost of confidence in my capabilities. However, I am also cautious of the fact that it is giving me an illusion of competency in areas that I have no working knowledge of. I guess, I can give myself the credit for this one because of accomplishing a relatively complicated series of task-based actions that I was prompted to use AI. All I did was to think of a solution, assess the possibilities and its practical use, look for limitations and ease of use, and determine its stability for future-proofing. I would say that I have achieved all of that on top of creating a consolidated file of negotiation that is fit for training the GPT.
What I will do is since I'm already here, I'll try to use these tools within the context of cognitive scaffolding to learn more things that can help me get out of this hellhole.
On Becoming a self / how the world is continually desiring to deny its own humanitySelf-dehumanization without our consent On Work, therapy, and recovery / What I Learned in Therapy
From here
Kael Espino. “The illusion of competency.” enjambments, Jun 21, 2025. https://enjambments.com/logs/2025/06-june/2025-06-21/
@online{espino2025the,
author = {Kael Espino},
title = {The illusion of competency},
year = {2025},
date = {2025-06-21},
url = {https://enjambments.com/logs/2025/06-june/2025-06-21/}
}
All 10 connections
- 2025-07-09 · Automating oneself out of work
- 2025-07-11 · Freedom transforming into agency
- 2025-06-10 · The burden of difficult conversations
- 2025-01-20 · On communicating with strangers
- 2025-06-05 · The trap of the unfinished
- 2025-06-25 · On the pace of eventual knowing
- 2025-01-04 · On machines that lie
- 2025-05-26 · The boundary of public speech
- 2025-01-06 · On creating without words
Select any words to keep them
One letter a month
Reader
- Essays a few days early (every essay stays public)
- Your name in the acknowledgements, if you want it
Interlocutor
- Everything in Reader
- First claim on prompt requests, one each quarter
- A word from me when the piece is done
You have read the whole path