Efficiency needs to become a product feature again

I caught Google Chrome peaking at 7.13 GB of RAM on my machine.

7.13 GB, for a browser, how the hell did this become normal?!

The original PlayStation had 2 MB of main RAM, and Metal Gear Solid still delivered gameplay systems, enemy behavior, scripts, cutscenes, audio, and an entire world that people remember almost 30 years later.

Today, I have seen a single LinkedIn tab use more than 700 MB.

Word 2003 could run comfortably in tens of megabytes. Today, a basic document workflow can consume hundreds.

Yes, modern browsers do more, they have sandboxing, JIT compilation, extensions, video, synchronization, security systems, and dozens of background processes.

But complexity is not a blank check for waste, for years, the software industry has covered inefficient code with more hardware.

More RAM, more cores, more cloud infrastructure, more cost.

Instead of forcing software to become better, we simply allowed it to become bigger, i refuse to accept that as progress.

AI should not be another excuse to add more code, more wrappers, more dependencies, and more layers.

Its greatest opportunity may be the exact opposite, use AI to find waste, delete unnecessary code, remove bad abstractions, rewrite bottlenecks, simplify architectures, make software radically smaller and faster.

At our startup, we stopped accepting “that is just how modern software works”, some parts of our code are now more than 1,000x faster than the traditional implementations they replaced.

One program went from using 96 MB of RAM to just 192 KB, while delivering the same result, same output, a roughly 500x reduction in memory use.

That is what AI should unlock, mot more AI slop, not more bloated applications, not software that needs a supercomputer to open a document or render a feed.

Leaner code, faster products, lower infrastructure costs, more respect for the user’s machine and time.

Efficiency needs to become a product feature again.