Possible Futures in Software
[I don’t like predictions and I don’t think much of soothsayers. But here’s an attempt to make some and be one]
From my early twenties, I’ve identified as an engineer. Even as my career evolved into leadership, building organizational capabilities, and using PowerPoint more than an IDE, coding has always been my happy place.
After spending the past six weeks immersed in AI tools to build software, it’s clear that the discipline of software engineering has radically changed.
Things happening now:
AI’s knowledge of popular programming languages is better than the vast majority of human engineers
AI’s ability to write code is very good
AI’s ability to create software is ok
AI Agents’ ability to replace human experts in real-world tasks isn’t great
When assigning a task to an AI, the more important it is, the greater the need for human review
It’s nearly impossible to keep up with the speed at which new models and tools are being released
Things I would bet $1 on happening soon:
The number of apps is going to explode
AI-generated code will make up the majority of new software
A new job, Software Producer, will emerge
Virtual content houses will pop up where groups of Software Producers pump things out at the speed of TikTok videos
Vibe coding bulls will be on parade for some time
Bears will wait for an inevitable, high-profile vibe coding fail before saying, told ya
Impressive security snafus will emerge
How these changes may impact software creation:
From the software producer side, AI will shift software creation from being engineering-focused to user experience-focused.
From the user side, the “thing to be done” will become the focal point. Budgets, not spreadsheets. Plans, not project tools.
Analogous to how there are great graphic designers who can’t draw, and punk rockers who can’t read music, we’ll see some amazing products from non-engineers who vibe code
Software production will start to resemble the music industry - there will be individual stars, popular groups, short-running hits, trends, etc
New software teams will be smaller than existing ones
Existing software teams will crank out more stuff faster, or at least be expected to
Fast food apps — quick, useful, low-effort — will be everywhere
Meme apps will spike and vanish like viral tweets
A build-first culture will begin in startups and move to enterprises
Enterprise IT departments will become the CoE for corporate process design as demand grows for custom AI Agents handling back-office tasks
Software companies will have to work harder to retain users
Engineers who rebrand as AI Engineers or Software Producers will command higher salaries — and higher expectations
Policy-heavy organizations will need to shift to being more principle-focused in order to deal with the tsunami of custom software being built inside their walls
The number of custom apps will rival the number of PowerPoints at big companies
Software production strategies will need to be agile and adaptive, focusing on speed without ignoring outcomes, quality, or cost
Big, garbage enterprise software with awful user experiences will improve or die
Speed will be the primary consideration for managers, with crazy assumptions about high quality and low cost
Niche software shoppes, like website design studios, will quickly adopt AI tools or have trouble staying in business
We will all expect super-simple interfaces, chatting-with-everything-for-everything, and just-do-it-for-me experiences with software
There will be more, not less, demand for builders over the next few years. Including vibe code-native junior engineers
Software will become a medium for artistic and creative expression
Software products are about to get much more fun