Every generation gets at least one storage format it treats like bedrock.
For my parents, it was tape. For mine, discs. For younger teams, cloud dashboards and collaborative docs that seem impossible to imagine replacing.
And every generation eventually learns the same uncomfortable truth: what felt permanent was often just a bridge.
Not a failure. A bridge.
The nostalgia trap: we remember impact, not duration
Some of the most emotionally vivid technologies of the last 30 years were surprisingly short-lived in their dominant phase.
- VHS felt culturally invincible, then got outcompeted by DVD speed and convenience.
- DVD felt like the final answer for home video quality, then lost relevance to streaming and subscription libraries.
- Blu-ray won a format war and still arrived at the same destination: niche status in a streaming-first world.
- Floppy disks once represented modern portability itself, then became an icon of obsolescence.
We tend to misremember these as “eras that lasted forever,” when in practice many had a brief period of mass dominance before the stack moved upward.
Why “revolutionary” technologies disappear quickly
When we study adoption patterns, the same sequence repeats:
- A technology removes a painful bottleneck.
- It scales quickly because the pain was universal.
- New infrastructure appears above it.
- The old breakthrough becomes friction.
VHS solved recording and distribution economics for home video. DVD solved random access, quality, and durability pain. Streaming solved distribution latency and inventory constraints.
Each format was excellent for the problem landscape of its moment.
Then the problem landscape changed.
Floppy disks: from symbol of the future to symbol of “save”
IBM introduced the floppy disk in the early 1970s. By the 1980s and 1990s, removable magnetic media became the default everyday transfer mechanism for consumers and offices.
Then capacity demands rose, networks improved, and software grew beyond floppy-era assumptions. By 2011, Sony ended floppy disk production, and the format moved from utility to iconography.
That icon still survives in software as the “Save” button — a tiny museum exhibit embedded in modern UX.
VHS, DVD, Blu-ray: three wins, one pattern
The home-video sequence is especially instructive because each step was legitimately better than the previous one.
- VHS delivered recording and rental economics that changed household media behavior.
- DVD made media smaller, cleaner, easier to navigate, and better to store.
- Blu-ray raised quality and storage again, and even won a major standards battle against HD DVD.
Yet none remained the default center of gravity.
The center moved from format ownership to service access.
Netflix ending its DVD-by-mail service in 2023 was less a funeral for one product and more a marker of that larger shift.
The real lifecycle is layer migration
We usually describe these transitions as “one device replacing another.”
That misses the deeper shift.
What changes is not only the device or disc; what changes is the layer where value concentrates:
- from hardware mechanism,
- to media format,
- to software interface,
- to network service,
- to recommendation and personalization systems.
People did not stop wanting movies, music, files, and memory.
They stopped needing to manage the same physical layer in order to get them.
Why this matters now: AI feels permanent for the same reasons discs did
Today, AI products feel like the new inevitability. That may be true at a broad level.
But specific AI interaction patterns we think are “the final form” are likely intermediary too.
Current habits — giant chat panes, prompt gymnastics, copy-paste pipelines, model roulette — may look as quaint in ten years as CD binders do now.
If history is any guide, the durable winners will be systems that:
- reduce operational friction,
- fit naturally into existing workflows,
- and move complexity to a layer people do not have to manually maintain.
The breakthrough is rarely the thing we initially fetishize.
A more useful question than “Will AI stay?”
AI is likely to stay. That is not the interesting question.
The interesting question is: which part of today’s AI stack is just a bridge?
- Model providers?
- Chat interfaces?
- Agent frameworks?
- App-level copilots?
- Enterprise orchestration layers?
Some of what feels foundational now will become implementation detail.
That is not a tragedy. It is how technological maturity works.
The paradox of progress
We call a technology “revolutionary” when it changes behavior quickly.
But fast behavior change is often exactly what shortens a technology’s visible dominance, because it accelerates the conditions for the next layer.
In that sense, being temporary is not evidence of failure.
It can be evidence that the technology succeeded so thoroughly it made itself replaceable.
The formats we miss did their job.
And if we are wise, we will treat today’s AI tools with the same respect: as powerful, consequential, and possibly transitional.
References
- IBM — Floppy disk history: https://www.ibm.com/history/floppy-disk
- Toshiba — Notice regarding termination of HD DVD business (2008): https://www.toshiba.co.jp/about/press/2008_02/pr1901.htm
- Netflix — DVD-by-mail closure announcement (2023): https://about.netflix.com/en/news/netflixs-dvd-by-mail-the-next-episode
- Wikipedia — Floppy disk: https://en.wikipedia.org/wiki/Floppy_disk
- Wikipedia — VHS: https://en.wikipedia.org/wiki/VHS
- Wikipedia — DVD: https://en.wikipedia.org/wiki/DVD
- Wikipedia — Blu-ray: https://en.wikipedia.org/wiki/Blu-ray
- Wikipedia — Compact disc: https://en.wikipedia.org/wiki/Compact_disc