You should absolutely judge a book by its cover, that's the entire point of covers. Sure, hold your full judgement until you understand more, but probably everyone's already heard this whole clichéd spiel before — so what is my point?

Well, it used to be the case that when you looked at a cover, you could immediately tell to some significant level of accuracy how much work likely went into it. Is this a beautifully illustrated, embossed art piece specific to the third novel in a series, or is it Papyrus font with a drop shadow for someone's first Amazon self-publish?

We make these judgments instantly, subconsciously, and we correlate the effort that went into creating the quality of the cover to the effort that must have gone into making the subject itself high quality. You know this, it's obvious, but I promise I'm going somewhere.

Everything Has a Cover

Some of them look how we're accustomed to book covers looking e.g. website landing pages, or a game's Steam store page. Some of them are a little less obvious. If someone puts a report on your metaphorical desk that has 1 page of research, vs a report that has 100 pages of research, you could rightly infer a difference in effort and (likely) therefore a difference in quality, before you even begin reading. That's not a cover, exactly, but the sheer chonk of the report is filling the exact same role as a flowery front page.

For essentially all of human history, this has served us pretty well, but AI completely breaks this dynamic. Generating good-looking artwork is now trivial. Generating pages upon pages of coherent text is equally effortless. It has made the 'covers' for many mediums completely disconnected from their quality, and it doesn't seem like it's about to stop any time soon.

The fundamental tools we have always used to tell at a glance how much human effort went into the making of something are now broken forever.

Sure, the pitfall of style over substance has been open since time immemorial, but at least that style was something that a person actively had to work to produce. No longer.

Perhaps we are approaching a time where rather than delivering a 300 page report as evidence of work, we must instead take personal accountability for a single page providing all of the information needed to effect change, leaving no AI-generated room in which to hide if it isn't worth its weight in solid gold (and I hear that's quite pricey these days). Perhaps the lack of quantity will serve as a 'cover' in the next era… For now though, aesthetics and bulk of work are simply no longer a reliable indication of the actual human effort input.

Hold On...

But that's not quite true, is it? You can tell (at least for the moment) when the aesthetics of something are clearly human-made. It's not about how well rendered it is, or its resolution, or how shiny it appears. It's something else. It's intent.

It's a tiny detail of a character's pocket watch design which obviously fits their personality. It's an imperfection in a guitar solo that is left on the record because it sells the emotion of the song far better than any of the other takes could hope to. It's a lingering tracking shot that remains on a character in silence for long enough to make the audience uncomfortable, and then deliberately continues past that point.

These are things that scream out to us that the human on the other side of the creation had something specific to say, and this is how they wanted to say it.

Software's Cover is Broken Too

Now, there is a lot of debate about the role AI can possibly play within this dynamic — it's a very complicated topic — but rather than opening that can of worms I'd rather take a second to draw a parallel with software. Just like these other mediums of creation, software has historically been extremely difficult to produce. A decent UI has been a very reliable cover flying the flag of human work and intent for the entire time that software has even existed. Once again, that's broken now.

I've found this to be especially true when building with Rool. So many of the old barriers to app creation just don't exist any more. I'm not worrying about data safety because there's infinite undo. I expect stuff to just work multiplayer by default. I don't even really consider data types any more because if I throw it into a space it'll just smoosh itself into whatever I need it to be next time I want to work with it. These days I'm so lazy I don't even provide a link to the coding agents, I just ask them to integrate the Rool SDK. That is crazy, but that's where we are in 2026.

The Speling Errors

In a world where making functional software is that simple, we will have to turn to other signals to ascertain which creations were labours of passion. I expect the best apps to get more and more opinionated, as strong experience design becomes a differentiator in a market where pure function is trivial to deliver.

There is a muscle to be built for all of us that is somewhat the software creation equivalent of leaving the speling errors in so that people looking at it can tell it's not just the AI's default way of solving a problem — because we'll all be able to generate that for ourselves trivially.

There's a very particular impression you get from masterwork media, when you can tell that the creator is reaching through the medium with purpose.

I think the oncoming saturation of basic functionality could push software design to compete on a similar level, and I look forward to seeing what gets created as a result.