Vibe coding is the Polaroid camera of software development Baptiste Fernandez asked me for some short thoughts on vibe coding, which lead to this clip: I wanted to expand on this a lil. I made this statement partly as an experiment to see if it would connect with the particular bias that someone has with vibe coding. It worked. Depending on the perspective, people tended to respond in one of two camps: Camp 1: "Yes! All you can do is create crappy photos!" Camp 2: "Yes! I remember when I got a Polaroid camera. The friction of developing film was gone, which changed habits and resulted in so many more photos. I have so much nostalgia for that time." Ben Thompson and John Gruber discussed the nature of AI slop as we get increasingly good GenAI image creation. Ben spoke about how we have seen changes as we democratized the ability to create and distribute information. We went from newspapers to blogs to social media posts. We dramatically raised the levels of slop, but one persons slop is another one's gem. The Algorithms™ were created to help find the signal that matched personal preferences from the noise. How may this transpire with software? Is vibe coding going to create a sea of crappy software? Are software engineers doomed? Let's look at the evolution of photography again. Each of us have cameras in our pocket, that are with us 24/7, and are dramatically higher quality than Polaroid cameras. The number of images created per day is staggering, and once again you can argue that one persons slop it another's gem. The photo of my kid at the Eiffel Tower is highly personal, and truly not interesting beyond a small family cohort. On the other hand, some amazing photos are captured from someone else's phone, due to skill, luck, and the fact that these cameras are impressive and constantly improving. This doesn't mean there are no longer skilled professional photographers. If you look at the average number of photographers per year in the US, based on Bureau of Labor Statistics data, you see: 2000s: ~126k photographers2010s: ~132k2020s: ~123k 2000s: ~126k photographers 2000s 2010s: ~132k 2010s 2020s: ~123k 2020s Some shrinkage given the population increase but COVID had an obvious effect too in the early 2020s. The total number also hides the fact that the composition of photographers changed, and the job became more polarized between high-earning specialists and lower-earning generalists. Just as there is a large spectrum of "photographers", the same is even more true for "developers". You aren't going to have an average app developer rock up to work on the Linux kernel, or a nuclear missile launcher. There are so many specialties in the world of software, and there will always be lot of room for "professional" software engineers. There is also room for the democratization of software. We can put the "soft" back in "software" and build very flexible, personal solutions, akin to taking our own photos. Just as we can all pull the camera out of our pocket, why shouldn't anyone be able to pull out creation tools for software? We may start with a Polaroid camera, improve to a mobile phone camera, and continue on the evolution so more people can develop higher quality solutions that work for them. We won't have to wait for software that was created by a designer for a particular persona... and if you differ the experience will be sub optimal. This also doesn't mean everyone creates software from scratch. Platforms that embrace this new reality could allow for the easy creation of extensions to allow more customization. What are your thoughts on vibe coding? Let us know if there is room for it as an AI Native by joining the Discord!

Stanford, Laude Institute Unveil Benchmark to Test AI Agents in the Terminal

Too Long; Didn't Read

Make resilience your competitive advantage

The Polaroid Generation of Software Is Here

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

Untitled Story

Stanford, Laude Institute Unveil Benchmark to Test AI Agents in the Terminal

The Noonification: How Often Do NFTs Pass The Howey Test? (1/13/2023)

Darwin's Hybrid Intelligence to Align AI & Human Goals for Startups & VCs

The Noonification: White Man (11/26/2022)

The Noonification: The Metaverse is a Sh*tshow (11/2/2022)

Stanford, Laude Institute Unveil Benchmark to Test AI Agents in the Terminal

The Noonification: How Often Do NFTs Pass The Howey Test? (1/13/2023)

Darwin's Hybrid Intelligence to Align AI & Human Goals for Startups & VCs

The Noonification: White Man (11/26/2022)

The Noonification: The Metaverse is a Sh*tshow (11/2/2022)

Light-Mode

Classic

Newspaper

Dark-Mode

Neon Noir

Minty

HN StartUps