Om
About
ProjectsExternal
Writings
Builds
Ships
Dev Notes
Connect
/writings/postgate

Code, prose, and periodic noise. A place to put things so I don't forget how they felt.

Newsletter

Quick Links

  • About
  • Projects
  • Writings
  • Research
  • Builds
  • Ships

Explore

  • Dev Notes
  • Poetry
  • Reading
  • Photography
  • RSS Feed
  • The Sandbox Podcast
  • System Status

Contact

  • Connect
  • Email
  • Report Security Issue

© 2026 DESIGNED & BUILT BY OM.

RRelative
SponsorFeedback
PDFPDF
Download resume
back
Essay

The Gate Moved

Aug 9, 2026, 12:04 PM·Om
Share
Table of Contents▼

Picture someone who has spent her career writing marketing campaigns, someone who has zero background in programming. One weekend she opens an AI coding tool for the first time and describes an idea for a small app that tracks freelance invoices. By Sunday night the app works. She shares a short thread about how she built it, and within a day, a few hundred people sign up to try it. The interesting part here has little to do with the app itself, since plenty of invoice trackers already exist. The interesting part is what changed for her. A few years ago, turning that idea into something real would have meant months of learning to code, or finding someone else to build it, or leaving the idea alone. This time she skipped that entire step. Coding stopped acting as a requirement for acting on a software idea, and that shift matters more than the app itself.

Something similar is happening across many fields at once, in video editing, music production, design, and writing. The honest way to describe this is that the cost of turning an idea into something people can actually use has fallen dramatically. That is a real shift, though it helps to stay precise about what it actually covers. Judgment about what to build, debugging, distribution, understanding what users actually want, and the slow work of iteration still take real skill and real time. What has changed is the starting cost. A person with an idea and some curiosity can now produce a rough, working version of that idea in an afternoon, something that used to take a team weeks or months to reach.

This kind of shift has a history, and looking at it helps put the current moment in context. The printing press let ordinary readers interpret texts that once required a trained scholar standing between them and the words. The camera turned photography from a slow technical craft into a skill picked up in an afternoon. Personal computers moved design and layout work out of professional print shops and into home offices. GarageBand turned music production into something a teenager could do in a bedroom. YouTube turned broadcasting into something anyone with a phone could attempt. Each of these moments brought the same worry, that quality would collapse under the weight of new entrants. Each time, a market eventually separated the good work from the rest, over years rather than overnight. AI reads like the newest chapter of a pattern that keeps repeating.

Here is where education comes in. For a long time, education worked like a gate placed before the work began. You studied for years, earned a degree, and that degree stood as proof to an employer that you were ready. AI is starting to change that order, though the change is more layered than it first appears. It is tempting to say credentials are fading out, but a more accurate way to put it is that credentials have lost their standing as the only path into a field. A resume can now be compared against a live product with real users attached to it, where before the resume stood alone. A cover letter can now sit beside a public thread that shows exactly how something was built, with real numbers included. That comparison was rarely possible before, and now it happens often.

The loop underneath this shift looks different too. Traditional education followed a fairly fixed order, learn first, earn a credential, then work. What is emerging alongside it looks more like this, try something, build it, watch it fail in places, learn from exactly where it broke, build again, and then show the result. Education, in this loop, becomes a companion to building rather than a locked gate standing in front of it. A person can study a subject deeply after shipping a rough version of an idea, once real feedback shows precisely where the gaps sit. That is a genuine change in order, and it is a more useful way to frame this shift than simply saying credentials matter less.

When building something turns cheap and fast, a few things become the real differentiators. Taste, meaning a sharp sense of what actually matters to build and what only looks impressive on the surface, becomes harder to fake and easier to spot. Choosing a sector matters more too, since raw execution ability is spreading across every field at once, so where a person points that ability starts to carry more weight than how well they can execute in general. Reaching an audience turns into its own skill, separate from building the thing itself, since a great product with zero visibility still goes unnoticed. And trust, built slowly through a public track record over time, starts to carry weight that a single credential earned years ago used to carry alone.

A few fields hold their ground against all of this, and naming them keeps the argument honest. Medicine, law, aviation, and structural engineering keep their credential requirements firmly in place, regardless of how good the surrounding tools get. A license in these fields carries legal weight tied to insurance and public safety, and that weight has little to do with skill signaling alone. AI already changes daily practice inside these fields too, helping doctors read scans faster and helping lawyers draft contracts quicker, yet the credential itself remains a firm legal requirement, and that is unlikely to change soon.

So where does this leave a person deciding what to do with their time today. Pick a sector and get specific about it. Build things in public, even small ones, and let people react to them honestly. Treat whatever AI produces as a first draft that still needs your judgment and your editing. Spend real time forming opinions about what good work actually looks like in the area you care about, since that opinion is quickly becoming one of the most valuable things a person can carry into any room. The gate is still there. It has only moved, from the entrance to the end, where the work itself gets judged.

Hacker News

View all
  • Proposal for universal AI ethics standard against country censorship

    1 point · 0 comments · Jul 17

  • AI Censorship Tested Across India, China, US, and Europe

    6 points · 0 comments · Jul 17

  • If AI Can Produce Everything, Who Buys Anything?

    3 points · 0 comments · Jul 16