Why Shell, Regex and ASN.1 Are Suddenly Sexy Again
Some technologies never died.
They were merely displaced.
Shell. Regex. ASN.1.
For decades they were treated as relics of an era when developers still knew what a byte was. Cryptic. Unreadable. Nerdy. Not “state of the art”. If you wanted a career, you talked about frameworks, not pipes.
And now?
Now millions of people sit in front of an AI and type: “Write me a regex that …”
The Truth: They Were Never Bad. Just Inconvenient.
Shell is not a toy. It is a deterministic, transparent orchestration system. Regex is not a monster. It is formal language theory in work clothes. ASN.1 is not a bureaucratic format. It is precise type definition with binary efficiency.
What made them unattractive was not their quality. It was their cognitive barrier to entry.
Anyone who has spent half a day wrestling with escaping, quoting and subshell hell knows: the tools are powerful. But they punish carelessness immediately.
The human was the bottleneck.
AI Shifts the Bottleneck
Suddenly syntax is no longer the problem.
You describe semantically what you want. The AI produces the syntactically correct form.
Example – regex:
Before: You leafed through documentation, built lookaheads, tested by trial and error.
Today: “Generate a regex that captures all log lines containing ERROR and an IPv4 address, and explain it.”
Done. With explanation. With tests.
The same with shell:
“Write me a robust Bash script with error handling, retries and logging.”
Ten years ago, the advice would have been: “Just use Python.”
Today? Why, actually?
The Quiet Renaissance of the Formal
ASN.1 is the best example.
Strictly typed. Efficiently encodable (BER, DER, PER). Ideal for low-bandwidth networks, embedded systems, military applications, edge devices.
And yet it was always a niche discipline. Bit-level thinking. Encoding tables. Manual labour.
With AI, the mechanical drudgery disappears.
What remains is the actual value:
- formality
- determinism
- efficiency
- interoperability
These are properties that modern JSON-over-HTTP worlds often lack.
While Everyone Talks About “Vibe Coding” …
… engineering is quietly returning in the background.
I have argued elsewhere that Vibe Coding confuses producing with understanding. Here we see the other side: the same AI that generates prototypes without system understanding can also serve to make formal tools accessible again.
Vibe Coding produces prototypes. Shell produces reproducible workflows. Regex filters billions of log lines with zero runtime overhead. ASN.1 saves bandwidth in networks where every byte counts.
The key point: AI democratises complex tools.
Regex used to be gatekeeper knowledge. Today it is a prompt. Shell used to be an art. Today it is a generated building block. ASN.1 used to be a speciality. Today you can have the encoding explained to you – bit layout included.
But Isn’t That Dangerous?
Yes.
When nobody understands what is being generated any more, a new dependency emerges. Formal rigour without formal understanding can be deceptive.
Blindly generated low-level code is not progress. It is merely faster error.
The difference does not lie in the AI. It lies in systems thinking.
Those who understand architecture win. Those who merely generate lose.
The Real Punchline
Perhaps what we are witnessing is not a revolution of new tools.
But a rehabilitation of old ones.
AI works like a complexity prosthesis. It takes over the syntactic heavy lifting and returns our gaze to what matters.
And what matters has always been:
- clean interfaces
- clear typing
- deterministic workflows
- minimal dependencies
Shell, regex and ASN.1 are not sexy because they are new. They are sexy because they are robust.
And suddenly they are usable again – without having to devote your life to escaping.
Perhaps the Future Is Not No-Code.
Perhaps it is: AI-assisted return to formal, solid architectures.
And that would be one of the quietest yet most profound shifts of recent years.
The framework hype no longer wins. What wins is what always worked.
Only this time with an interpreter.