*This was originally posted on\nX and blew up. To plant\nmy flag, to say that these are things I believed in September 2026, here it is\non the blog, non-ephemeral. Some of these predictions are just observations,\nthey’ve long been true at companies like Amp. Others will take some time to play\nout.*
\nCode review will die. I mean: it’s already dead. But in the future, humans\nwon’t find a bug or an issue with the code produced by a model, at least not in\na reasonable time. Humans will only review the system and its composition, but\nit won’t be in PRs and it won’t be by looking through every line of the code.
\nUnit tests might die too. Why have training wheels if you never fall over?\nI’ve had models write 900 lines of Arduino C, compile it without *a single\nerror*, and send it to the device, where the program ran perfectly. 900 lines\nwill be nothing in the future.
\nThe craft of writing code will disappear. Yes, there are still Italian shoe\nmakers around. But look at your feet.
\nThe craft of building software will be more important than ever. Knowing\nhow to solve business problems with software, how other software did it and why\nand why not, when and how to ship it, how to get feedback on it – that’s the\nnew game.
\nMost bugs won’t be “coding” bugs. They’ll be “you asked for the wrong thing”\nbugs.
\nOpen source in its current form doesn’t make sense anymore. “Given enough\neyeballs, all bugs are shallow” is still true but now we have artificial\neyeballs.
\nPerformance critical contributions by humans will stay what they are: an edge\ncase. In 99% of software it does not matter that you could’ve written a\nfaster algorithm or picked a better data structure. You have no customers, no\nusers, no one’s executing the code. It does not matter. When it matters, the\nmodels can fix it. Do not compare the top of the top 1% of software\n(developers) to 99%.
\nThe terminal is dead. Most developer tooling will be washed away by tokens.\nShells, text editors, CLI tools won’t be used by humans anymore. There’s no\nneed to know command line flags and jq invocations anymore. It’ll be seen as\narcane as knowing how to write everything as a Perl one-liner. (I’m saying this\nas a lover of the terminal & dev tools.)
\nTokens are the new computing paradigm. Everything will be re-made on top of\nit. We’ve had deterministic computers for 80 years, so we confuse “how\ncomputers have worked” with “how computers must work.” We’re entering the\npost-binary era.
\nThe triad of PM/Design/Eng will disappear. It does not make any bit of\nsense anymore. Agile, SCRUM, whatever – dead. “Engineers” who act as “meat\nproxies” and shove tickets into agents and report back to humans will no longer\nbe valuable.
\nThe difference between software engineering at large corporations and small\nones will increase. Start-ups adopting practices of Google will look even\nsillier than before, because they can now build so much faster and with so much\nless constraints.
\nThere’s no proof that “good code” will matter in the future. The notion of\n“good code” itself is mostly based on the idea that it’s easy/cheap/efficient\nfor humans to work with. But humans won’t modify the majority of code. Think of\nhow dumb it is to assume that “only 80 columns wide” or “but newlines here and\nthere” matters for agents – now consider all the other properties you have in\nmind for “good code”. Yes.
\nSome people will be priced out of producing software. Just like only some\npeople could afford a personal computer in the 80s and 90s, for the next few\nyears, if you can’t get enough tokens, you’re playing second league. You need\nto get to the tokens.
\nIt’s questionable whether cheaper models will be used. Tokens will be\neverywhere and we’ll swim in tokens. And what we consider a smart model today\nwill be considered very dumb in the future. But a smarter model makes less\nmistakes, needs fewer turns. When do you really think “I’m okay with it being\nwrong a few times?”
\nModels will become so fast that UI will be generated on the fly. A lot of\nUI exists because software can’t understand what you want. Menus, settings\nscreens, dashboards, filters: much of it is a human-accessible API to a dumb\nmachine. Smart machines need much less UI.
\nIt’ll take a while for this to play out. It’ll take a generation for the\n“new software” to replace the “old software”. Just like there are people\nhappily employed as ASP developers today, there will be people employed to\nwrite code in 10 years. But do you want to have that job?
","headings":[{"level":1,"text":"What I believe about the future of software development","id":"what-i-believe-about-the-future-of-software-development"}]}}