There's something to this. I recently shipped a music curation site and deliberately avoided React/Next/etc - just HTML, CSS, vanilla JS. The cognitive load difference is stark. The 'peak' might be less about capability and more about us rediscovering that simpler tools often suffice.

Same. I build stuff for local businesses in my area with nothing but boring old HTML, PHP, CSS and JS. I guess my shit isn't "web scale" but it works, and it works consistently, with minimal downtime, and it worked during both Amazon and Cloudflare's latest outages.

I don't need my software to eat the world, I'm perfectly content with it just solving someone's problems.

It's all about picking the right tools for the job. The "cognitive load" might be larger in a vanilla project compared to React when your interface is more complex and interactive.

Speaking of JavaScript's evolution - I've been building a music player (muz11.com) and it's remarkable how far we've come. The Web Audio API, MediaSession for lock screen controls, smooth animations via requestAnimationFrame... all running client-side with no framework, just vanilla JS. Thirty years ago this would have required a desktop app and probably a record label deal.

The irony is that 'freeing' JavaScript from Oracle's trademark might matter less than freeing ourselves from the framework churn. The platform itself is incredibly capable now.

> Thirty years ago this would have required a desktop app and probably a record label deal.

And that would have been just fine.

Until you have to pay back your advance to the record label.

This reminds me why simple single-purpose web toys used to be so satisfying. No account, no onboarding, no "upgrade to pro" - just a thing that does one thing well. The world counter is a nice touch without being gamified into oblivion..

Tried this on iPhone - the category tabs (Sports, World News, Business) get cut off on the right and there's no horizontal scroll indicator, so I didn't realise there were more options at first. The story cards also aren't using the full screen width, leaving wasted space on both sides.

Cool concept though - the source count and "+N" spread metrics give a quick sense of which stories have legs.

One thing missing from this framing: the feedback loop speed. Animal evolution operates on generational timescales, but LLM "commercial evolution" happens in months. The optimisation pressure might be weaker per-iteration but the iteration rate is orders of magnitude faster.

Curious whether this means LLMs will converge toward something more general (as the A/B testing covers more edge cases) or stay jagged forever because no single failure mode means "death".

> Animal evolution operates on generational timescales, but LLM "commercial evolution" happens in months.

But LLMs have all been variations on transformer neural networks. And that is simply not true with animals. A nematode brain has 150 neurons, a bee has 600,000. But the bee's individual neurons are much more sophisticated than the nematode. Likewise between insects and fish, between fish and birds, between rodents and primates....

Animal evolution also includes "architectural breakthroughs" but that's not happening with LLMs right now. "Attention Is All You Need" was from 2017. We've been fine-tuning that paper ever since. What we need is a new historically important paper.