Bruce Chillaxes with "meta" - Happy Bootcamp Story 50 of 100

Bruce Chillaxes with "meta" - Happy Bootcamp Story 50 of 100

Bruce solved a mystery.

TL;DR: Learning the basics of "information about information" - or "meta" - can be a very happy day for anybody, especially a new coder.

"Oh, Of Course"

Bruce had been struggling with his project setups for several assignments in a row, but the day we showed him why, was surprisingly comforting.

The unexpected, for him, was that keeping his tooling organized was neither complex nor magical, it was just some dumb files sitting at the root of his project. Once he saw this pattern, he started seeing it everywhere.

Before, he was focusing on his own files, as if they were the last thing that he needed to focus on. But now he was seeing the pattern - information, then under that, information about information, then more on down, layer at a time.

Under his own files, there were files that organized how the project itself was structured. Hmmmm.

No Coursework in Meta

We don't teach any lessons in the abstraction of "meta" or "information about information" - instead it's a part of other lessons. For this primary information, we rely on other information to set the structure of the primary information with secondary information.

So it's always specific information about specific information - which might be unfortunate, because now Bruce is seeing it everywhere - just in different styles, different formats, different languages. But it's really all just variations on a theme, right?

You have to tell the stupid machine what kind of information it is about to process, so it can set itself up, to process the information accordingly. Now I'm spending frkn half my time just telling the machine what kind of information to expect, and when to expect it.

Hmm, Bruce thought he was going to learn to code. Is this all that learning to code means? Information about information?

Bruce, relieved.

"Either way," says Bruce. "It's a good day when I'm starting to think of this process as a lot dumber than I originally thought."

And it is dumber, definitely.

He thought it was just about languages and code. Instead, everything is just a bunch of made up lingo or API - and the real work is just figuring out how to talk to the stupid machines about the data and how it is structured, whatever that is.

"Now I'm just reduced to dealing with meta as a way to talk to machines. This, I can do."

LLMs

AI can't figure out what Bruce wants to get done and then do that. The job of a coder is still very real, and Bruce has to direct every process through every different step.

But Bruce doesn't mind at all when an LLM, or Large Language Model, helps him talk between the many different systems - each in their different language.

When Bruce graduates from our program, he feels ready. Really not so hard, this job of moving data around the enterprise.



was that files existed at the base of his project that


要查看或添加评论,请登录

Pete Carapetyan的更多文章

社区洞察

其他会员也浏览了