Recently, I've been reading a lot of discussions about clawdbot, with endless talk about how to improve efficiency, how to help reply to messages, and how to take over the computer to get work done. It was only then that I truly understood one thing.
In the past two years, new technologies, new applications, and AI plugins have emerged one after another. Whether it's self-media hype or user slogans, the talk is always about disruption, revolution, and efficiency leaps. Regardless of whether it's marketing rhetoric or shared user experiences, I've noticed one word that keeps appearing: work.
Everyone is using these so-called "ever-evolving" tools to work. But the problem is, the more they hype them up, the more similar their outputs seem. I always felt something was off but couldn't pinpoint it. Now, I finally get it.
Because aside from work, these things offer almost no other value.
"Set up the tool and let it work for me 24/7."
This statement itself is highly alarming.
This narrative not only denies the value of individual labor but also negates the value of labor itself. If a task can be fully automated around the clock just by configuring a tool, it only means your labor has no "premium," and the job's value is negligible. In other words, it lacks scarcity and doesn't require the cost of judgment.
From a market logic perspective, any output that can be replicated at low cost, on a large scale, and automatically is unlikely to have a premium, let alone dignity. The obsession with automation is essentially admitting that you're engaged in labor that can be replaced at any time, just dressed up in a more polite term.
These hyped AI tools are highly homogenized copy-paste.
Ultimately, they point to an automated process with no room for premium. The commonalities far outweigh the differences, with product distinctions mostly limited to interfaces and packaging (even interfaces are strikingly similar). That's why it's hard to articulate what makes a particular "innovative" tool irreplaceable or what it can change.
When the design goals of various tools increasingly converge and their forms become similar, the tools themselves lose their unique identity in users' minds. Everyone uses the same logic for "fully automated work," resulting in flatter architectures, more average code, and less flexible systems.
So-called efficiency improvements often just postpone complexity without eliminating it from the system.
But the most ironic part is the programmers who hype all this.
In reality, most programmers are indeed moving a pile of shit from point A to point B. This repetitive labor often stems from organizational structures, decision-making methods, office politics, and performative overtime, making it hard for individuals to escape. Yet, oddly, they get excited about "using different tools to move the same pile of shit."
Hundreds of AI tools have emerged, specifically designed to accelerate moving that pile. Programmers complain about repetitive work while taking pride in switching to a more convenient tool. Instead of using tools to amplify their creativity, they stay put, using flashy laser shovels to dig themselves deeper into the hole.
Their split reminds me of the absurd myth of Sisyphus pushing a rock uphill, but the twist is they also play the role of punishing gods. Self-punishment—nothing is weirder than that.
When the software industry completely transforms into a traditional industry.
From the "incompetent revelry" among programmers, it's clear that the software industry now seems to have fully morphed into a traditional industry.
Most insiders know that many system issues stem from inadequate requirement communication or insufficient testing. These shortcomings often arise from unreasonable deadlines. Many so-called urgent tasks are just results of temporary decisions or a leader's whims—symptoms of project management failure. The real needs of a normal project are limited. What's unlimited are constantly changing opinions, political consumption in meetings, and procedural performances that drag on for hours over something that could be said in one sentence.
In such circumstances, a piecework logic centered on hours and delivery speed naturally emerges. This distortion of a brainwork industry into piecework handicraft is squeezing a field that should be centered on thinking and design.
When an industry starts pursuing undifferentiated speed, ignoring unique logic and judgment, its evolution has essentially stalled.
Our tools may be more advanced than ever, but if the work content still revolves around chaotic decisions and legacy issues, even the shiniest tools only improve the efficiency of self-consumption.
This kind of progress looks lively but is actually a refined form of self-comfort.