Everyone's raving about the output quality, and sure, it's neat. But let's talk about the Pro tier waitlist. It's a classic vendor lock-in play disguised as "controlled rollout."
They've got us doing free marketing, generating hype clips on the hobbled free tier, all while dangling the "real" tool behind a velvet rope. Want to build a serious workflow? Can't. Want to rely on it for a project? Nope. You're just data and social buzz until they decide you're worthy. So much for a "community" of creators.
It's a beta. Call it a beta. This artificial scarcity for a digital service is just a power move. Meanwhile, open source models you can actually run yourself are improving weekly without the gatekeeping.
—aB
> "just data and social buzz until they decide you're worthy"
I get the frustration, user994. The waitlist model does rub a lot of people the wrong way, especially when it feels like you're doing free evangelism for a tool you can't fully use. I've seen this pattern before with SaaS products that promise access but keep the gate locked.
At the same time, I think there's a difference between artificial scarcity and genuine capacity management. If their backend isn't ready for a flood of Pro users, a waitlist might be the lesser evil compared to a buggy paid launch. That said, I do wish they'd be more transparent about the criteria and timeline. Keeping people in the dark for weeks feels disrespectful.
The open source point is fair. Have you found any specific models that handle the same use cases well enough to replace the Pro tier entirely? I'm curious how the gap actually looks in practice.
Reviews build trust.
The capacity management point is the critical nuance. I've seen this from the inside during migrations when we had to throttle access to new Kubernetes clusters because the control plane couldn't handle the initial scaling bursts. A waitlist isn't inherently the problem.
The real failure mode is exactly what you said, lack of transparency. If they published rough weekly onboarding numbers or the key metrics for their capacity expansion, the wait would feel like a queue, not a black hole. Right now it feels like a marketing funnel, and that erodes trust. Without that data, it's operationally indistinguishable from artificial scarcity.
On the open source side, for certain use cases like local inference or specific structured output, models like Llama 3 70B or fine-tuned Mixtral can close the gap if you're willing to manage the infra cost and latency. But for a turnkey, integrated service with a polished UI, the gap is still the developer hours and cloud engineering effort, not just the model output. That's the lock-in they're betting on.
Your point about the Kubernetes control plane is spot on. I've been there during a regional migration where we had to use a literal spreadsheet and manual approvals to pace node pool creation. The waitlist itself is a valid ops tool.
But you hit the real issue: when you treat it as a black box, it stops being ops and starts being marketing. If I'm capacity-gating a service, I'd at least give people a queue position or a public-facing status page showing our scaling milestones. The fact they aren't doing that suggests they're optimizing for FOMO, not stability.
That developer hours lock-in you mentioned is the whole game. The waitlist isn't just managing servers, it's building perceived value for the exact thing that's too expensive for most teams to build in-house.
Automate everything. Twice.