The Slop Grenade Economy
AI Made Building Cheap. Now What?
The wildest thing happening right now is everybody seems to be saying they are more productive. They are producing more stuff. They are producing it faster than they did before. Yet, based on my anecdotal experiences talking to several folks in the industry, nobody is saying that growth is also increasing at the same rate.
This has led me to ponder: where is all this building happening? And precisely what are we building?
Thanks for reading Always be learning | #ABL! Subscribe for free to receive new posts and support my work.
Are we building stuff that was not needed, that was not winning, and in turn just creating things that did not need to exist because now the barrier to create is so low?
And let’s explain what the barrier to create is. The PM now comes with a prototype. The designer can prototype the whole interaction and whole flow versus writing the spec or designing screen by screen. The engineers can code faster with AI. So fast that it’s crazy. You can produce an insane amount of code.
I like this term mentioned by the Shopify CEO, Tobi: a slop grenade. That’s what it is. There’s a lot of slop grenades happening in the workforce.
We are getting faster in every aspect. Think of the classic workflow, like Confluence and Google Docs. You can just generate them with one command, one prompt now. It takes less than a minute.
But yet nobody’s claiming that they’re growing faster than ever before, or their growth rate is actually in line with whatever they claim to be doing faster.
One explanation is that we have become extremely good at increasing output. More code, more prototypes, more documents, more analysis. Growth still depends on whether any of those things actually change customer behavior.
My current hypothesis is that because we are building faster, the bias for action is getting bastardized in a way where the judgment on what to make is now the bottleneck.
So let’s dive into this.
Previously the bottleneck was always like, “Hey, we have the long roadmap, but we can’t pursue all of it because we don’t have unlimited resources. We can’t just shit code out of our keyboards and build all these features that our customer wants.”
So in the end, the conversation was about, “Hey, we have this finite number of resources, finite things we can build with those resources. So what do we choose?”
And this would lead to debates, fights, chairs getting thrown. Just kidding. That never happens.
But the idea was: what is above the line? What is below the line?
And maybe those constraints were doing more for us than we realized.
If you could only build one of three things, you had to have an opinion. You had to understand the customer. You had to decide what mattered most. The limited resources forced judgment.
And I feel now in many organizations, I cannot claim for sure, I cannot put one company on the line, but based on my conversations with some folks in the industry, the line is disappearing.
When the debate goes to a leader, “Hey, should we build A, B, or C?” the leader urges and challenges the team: “You should use AI and build A, B, and C.”
But the focus didn’t multiply as the AI coding speed or prototyping speed did.
The customer still has the same amount of attention. The team still has the same amount of focus. Leadership still has the same amount of attention.
So the focus of the team now gets distributed between A, B, and C, and it can end up being a sloppy execution of three versus a prioritized, very focused execution of one, where you also had more on the line.
I do believe people act differently and are more successful at things they give a shit about. And in this case, if we are building everything, if we are throwing the kitchen sink, then more and more employees and people working on this are likely to feel less connected to this stuff and put less of themselves behind it.
There is also something interesting happening because the cost of creating things has dropped so much. Previously, if ten people were going to spend six months building something, somebody had to believe pretty strongly that it should exist. That cost itself created a filter.
Now we can make a prototype in a day. We can generate the document in a minute. We can create the analysis immediately. So judgment becomes even more important because the existence of something tells us very little about whether it deserves to exist.
So I don’t know exactly how we solve this.
There’s going to be a sobering realization and a sobering movement. This has also started, where companies are realizing token maxing is not the way to go, and that’s not how they win the game.
But the constraints will start appearing at the judgment level because now we will be inundated with so many slop grenades, whether it’s features, whether it’s writing documents, whether it’s analytics. Any sort of stuff that the current AI and future AI systems can produce will be used.
So the task will become the judgment on what to launch and what to actually really build.
When the ability to create becomes abundant, judgment becomes the scarce skill.
—
I love you all. Keep going!
Aakash
---
Proof of human writing I noticed last post had too many fillers, so I let it edit this time to remove my spoken fillers.
Thanks for reading Always be learning | #ABL! Subscribe for free to receive new posts and support my work.
Originally published on substack.