Uncle Tom and IA Agents
What & Why
Since I'm a matoore adult (though I like to think of myself as a child of roughly 7 years) who is able to challenge my own biases and view the world through OOPs (other people's perspectives – yeah you know me, s/o Naughty By Nature) I will attempt to take the side of Uncle Tom who at Christmas dinner regurgitated some incoherence about how "IA is going to take our jobs"…
I replied briefly: "No. Uh, I mean, yeah… too unreliable! I don't know…" I then smiled and cut myself off.
In my head… "Are you just that much of a f-–— idiot?! Is he f-–— baiting me?"
Maybe, Anton, you're arrogant, and he was interested in having a conversation.
Anyway – for an "Agentic Software Engineering" class as part of my Computer Science Masters (social-signal social-signal… I'm hardly a master, please… The man that knows something, Knows that he knows nothing at all s/o Erykah) we are simply reviewing & presenting 2 papers a week for 10 weeks straight the first of which is: Large Language Model Agent: A Survey on Methodology, Applications and Challenges
The above paper attempts to provide a unified taxonomy and architectural perspective on what is the fast moving and thus fragmented field of designing agents.
- attempting to address that IA boomer-doomer-ism
- my required summarization exercise for this paper
- while conceding that: sure, yes, these "agents" are fundamentally changing the human-macheen relationship
Also, it's more fun to write a narrative alongside learning purely technical information.
Summary
TODO revise from google doc
Conclusion & Addressing Boomer-Doomer-ism
No, IA is not going to take all of our jobs. Just the jobs of mediocre people. If that sounds arrogant, I probably am. Hopefully it's a wake up call to rise above mediocrity.
On the bright side, my hope it that it's a "tide which lifts all boats". By this I mean that humanity will be required to operate at levels which IA cannot.
From a purely software engineering perspective (yes, you can engineer software) it is at least encouraging that these agentic workflows demand good engineering practices and good practices compound to make agentic workflows effective.
On the dark side, many people will blindly trust IA systems and bad things will (continue to) happen. This has already been happening for many years before ChatGippity (s/o Primeagen) was unleashed. People seem to be that way… trusting answers to that which they don't understand.
Lastly and however… I'll remain that… IA will never, ever be as beautifully intelligent and capable of adapting and inspiringly creative and intuitively perceptive as you. Problem is… you don't even know how f-–— beatifully [intelligent] you really are (s/o Dave Chapelle). You have been conditioned out of knowing this and habitually hold yourself back. It is time to un-know-do.