Hunter's Blog

AI and B2B Slop

Or, why that finance bro's AI startup will fail in the next 2 years

2026-07-22

Across the Internet, whether it's a meme on Reddit or an actual post on LinkedIn, I think we've all seen this type of pitch:

Listen, bro, I've got an awesome idea for a product. You know how employees are always complaining about their email inboxes? What if we made an AI vibe-coded app that sorts everything in a spreadsheet, stores it all in a database, and displays them to the user as text? We can call it, "em_AI_l". We're gonna be trillionaires!

And for some reason, this idea gets funded with an angel investment round of $20 million.

Contrary to what this post might initially sound like, I wouldn't say I'm an AI-doomer. Specifically, the technology itself is a useful tool. I would be outright lying if I said that, as a software engineer, using Claude Code didn't speed up my workflow. It 100% does. You know what else has made me more efficient? Autocomplete for my function and variable names.

Treated as the tool that it is, as a cog in the machine of computer software, AI can be used for good purposes. Too often on the Internet, there forms this strict dichotomy, "you're with me, or you're my enemy"-type arguments.

I think the truth of AI is slightly more nuanced. There's a lot of potential for good uses here. For instance, people with sight impairments are able to have AI describe environments, texts on their phone screens, website layouts, etc. far better than any current screen readers. The AI can still hallucinate, but it's still a huge leap in the right direction, compared with the technology we had before.

Conversely, there's also been an enormous surge in vibe-coded AI slop, charging businesses absurdly high subscription prices for garbage. And it makes me ask...

Why would I buy your shitty app?

At its core, AI slop is defined by maximum extraction for minimum effort, often with heavy monthly subscriptions for a wrapper that took five minutes to vibe-code.

The title of this post is perhaps a little misleading. I don't think this applies just to B2B sales slop. I think this applies to a lot of other applications. But I'm getting ahead of myself.

Recently, I saw an article discussing how Starbucks has put all of their software vendors "on notice". They're ramping up an internal software department that will heavily use AI to, essentially, re-build these software applications as Starbucks-specific, internal-only applications.

Do I support everything Starbucks does? No, not at all. I have a strong suspicion that this is only there to get Starbucks evaluated as an "AI-first company", so their stock price can go up.

But the core idea of what they're doing, that stuck with me. I think this might be one of the core uses of AI that will make it truly useful: building custom applications for yourself/your business.

Think about it: why would you ever pay for, say, an online dashboard tracking your incoming and completed orders, when you can instead build it yourself? Every single B2B application has such a glut of unused features; their color themes suck; and they're so, sooooo slow, specifically because they need to appeal to as many businesses as possible.

Instead, a given business can create their own internal application that does exactly what they want, and no more.

Now, it would be irresponsible of me, as a professional software engineer, to not point out some of the obvious pitfalls here. Anyone can vibe-code an application. But maintaining the application over time requires some level of programming literacy, and it might not be fully viable for every business to hire multiple software engineers to maintain all their code.

There's also lots of services that simply cannot be replaced by vibe-coding. Cloud storage will always be an expense for businesses, simply due to the unreliable nature of self-hosting databases and systems. Same for things like financial systems, payroll, health insurance, etc; but even with all these caveats, there's a lot of space for businesses of all sizes to stop throwing money out the door on B2B subscription service.

AI Slop in Everyday Life

I don't own a business, and I don't work in B2B sales. So why am I writing about this? Because the exact same economic trap facing software vendors is coming for consumer media. Just as companies are realizing they don't need to pay a premium for bloated, generic software when they can vibe-code their own, everyday consumers are going to realize something fatal about AI-generated art, videos, and text: Why pay attention to a machine's generic output when you can just prompt your own?

When I see an AI-generated video, image, or large swaths of text, not a single part of my brain wants to engage with it. And for a little while, I couldn't quite grasp why. Sure, there's an uncanny valley effect sometimes, and AI writing in particular has a distinct flavor of "it's not just X, it's Y" that makes it frustrating to read. And then it hit me.

Why would I pay attention to this thing, when I could instead generate my own version that is more specific to me and my preferences?

When I read a well-written book, I'm not engaged with the text purely because "I hadn't thought of it yet". I'm engaged with the phrasing of the author, with the imagery conjured up in my head. And I have some vague comprehension of the amount of effort put into learning this craft, all the rounds of editing and polish. Having written several pages of several very unfinished novels myself, yeah, it's really, really hard to write a whole damn book.

When I go see a movie in the theaters, I'm not engaged with it purely because "huh, I'd never thought of Matt Damon playing Odysseus". I'm engaged with the cinematography, with the thousands and thousands of man-hours poured into bringing a story to life.

When I see art, I'm not engaging with something "new to me". I'm engaging with the craft, the humanity. I see something I can't do, and I'm in awe.

If art is replicable, is it less impressive? Not at all. I can play the guitar decently well, mostly just chords and some light fingerstyle. I can strum along to most songs on the radio. But the composition, the specific details of how melodies and chords fit together, is a little more obfuscated to me. I can compose my own songs, but the magic of listening to other people's music, for me, lies in the mysticism of how they came up with their lyrics and melody and chord structures.

When I see an AI video of a blonde catburglar, breaking into a rich man's penthouse, only to find an array of armed guards that she has to fight? I'm not at all impressed. I just think to myself, "I would've made her a redhead, and the guards should be robots". This isn't a real example, it's just something random I thought up right now.

Why would I ever want to pay to see an AI movie? Why would I ever spend money on an AI-generated video game? Why would I waste my time reading documentation generated by an AI? All of that is influenced by the people doing the generating. Without the craft of making the art, what is the point of seeing this AI "art"?

Conclusion

The rant's almost over, I promise. I just want to make some things clear.

I'm not advocating for everyone to generate their own artwork. I love human-made art, and I love the dedication, sweat, and skill put into the act of creation.

What I’m saying is that AI slop creates its own consumer repellent. If I want something low-effort and personalized, screw it, I’ll generate it myself. But if I want to actually engage with something, whether it's music I wrote by hand, campaign notes typed out late at night, or a movie that took hundreds of man-hours to make, I want the friction. I want the humanity.

While the finance bros keep trying to shove automated, generic slop down our throats, the market is inevitably going to hit a wall. Once everyone realizes that mass-produced AI content is entirely valueless because anyone can make it, the real, human-made craft can stop competing with the noise and rise back up.