Are We Becoming Architects or Butlers to LLMs?
12 points by metadata
12 points by metadata
In a recent viral post, Matt Shumer declares dramatically that we've crossed an irreversible threshold…
Leading with a reference to a “viral post” by Matt Shumer doesn’t do a lot for the credibility of the article.
This is one of the highest quality bloggers on the internet, and this post is also great.
Referencing something popular doesn’t take away credibility at all.
The post is fine, I don't disagree with it. Leading in with a popular post written by a someone generally associated with boarderline fraud at best, not great. But sure, fine post.
What do you mean by borderline fraud? I'm not familiar with the details, I'd love a pointer to what he did.
I've had a couple people breathlessly send me this Matt Shumer post, but I've taken an internal pledge not to read anything about AI written by someone with a financial motivation as to the tech.
When people claim AI's can't develop software, they're dismissed as trying to cope with a threat by denying it. This ... just reads like a different way to cope with the same thing.
"Yes, programmers will never be employed at the same scale they are today, but I will retroactively declare programming to be a tedious grind. The real work will still be there, requiring my unique genius and prompting skills."
It's AI-positive. It'll undoubtedly be true for some (if you're a research scientist at a major company, your chances are probably better than our chances). But it feels like just another way to deny reality.
The last industrial revolution ushered in a century of horrible labor relationships, inequality so bad it caused people to romanticize the lives of manorial peasants, and it enabled the growth of vast oppressive world empires.
If there is a new one, we can hope it won't be as bad. But it's okay to be scared.
He’s an academic and researcher. It makes sense that programming was never the main thing for him
Even if things come to that, I know I'll keep thinking, keep learning, keep striving to build things. As I reflected in an earlier post on finding one's true calling, this pursuit of knowledge and creation is my dharma. That basic human drive to understand things and build things is not something an LLM can automate away. This I believe.
This is one of the best paragraphs I’ve read in many years. This is my personal belief system as well. I still play chess against other humans, even though a machine can beat both of us 10 times over. Because curiosity and a desire to build skills has always been what motivates me.
Whether or not I make money from exactly what I’ve been doing over the last 15 ish years, I’m going to continue being curious and trying to get to the essence of how complex things work.
We do the real thinking, and then we make the model grind.
This is it. When we break free from the grind, the higher-level thinking kicks in. This isn't just true for LLMs, but life in general. I think we are going to see an acceleration in innovation in the coming years powered by the fact that a lot of creative thinkers are no longer bogged down by the minutia of tedious implementation.
Software architects even back in the day was mostly a refuge for programmer rejects and now with programming gone that space is getting really crowded. At the same time there's diminishing need for your "high-level thinking" as models get better. In two years they'll be delivering passable solutions even to a drooling idiot.
I don't think we mean the same thing by high-level thinking. And when spaces get crowded, that's when new spaces are created.