Comments URL: https://news.ycombinator.com/item?id=47859103

Points: 4

# Comments: 3

Comments URL: https://news.ycombinator.com/item?id=47844912

Points: 2

# Comments: 1

Comments URL: https://news.ycombinator.com/item?id=47777280

Points: 3

# Comments: 0

For years, SaaS products were designed around a predictable assumption.

A human logs in, clicks around, and uses the product a few times a day. Pricing models were built around that behavior too.

But with tools like Claude writing code, executing workflows and calling APIs autonomously, that assumption is breaking.

Software is now being invoked by agents. An agent might run a workflow ten times in a minute. Or once in a week. It might spin up five tools for a single task and never touch them again.

The usage pattern becomes bursty, unpredictable, and completely detached from the idea of “seats”. This changes how infrastructure needs to be built.

Rate limits, entitlement checks, credit systems and billing logic now need to operate in real time. The system has to understand not just who the customer is, but what the agent is doing and how resources are being consumed.

This is why traditional subscription models are starting to feel increasingly misaligned for AI-native products.

When agents become the primary users of software, access-based pricing stops making sense. What matters instead is execution, consumption, and outcomes.

The companies that recognize this shift early will design systems that are built for agents from day one.

And the rest will spend the next few years slowly rewriting their billing infrastructure.

Comments URL: https://news.ycombinator.com/item?id=47335024

Points: 3

# Comments: 1

We're building flexprice, a monetization platform that helps making usage-based, credit based and hybrid pricing easier and We recently added a managed-storage option to the S3 export system in Flexprice.

Originally, S3 exports required the usual setup: your own bucket, IAM policy, access keys, region config, and ongoing credential management.

That works if you already have infra in place, but it adds a lot of overhead for teams that just want recurring exports for invoices, events, or credit reports.

Now there are two options:

- bring your own S3 bucket

- use a Flexprice-managed S3 connection

With the managed option, there’s no AWS account or bucket setup required. You can schedule exports, run them manually, and download files directly from the runs history.

The goal here was simple: make exports useful for finance and ops teams, not just infra teams.

Still supports:

- hourly / daily schedules

- manual export

- invoices, events, credit top-ups, credit usage

- execution history and downloadable files

Happy to answer questions on the design tradeoffs or why we kept both managed and BYO storage modes. Feel free to join our community or try our product at admin.flexprice.io

Comments URL: https://news.ycombinator.com/item?id=47283978

Points: 3

# Comments: 0

We kept running into the same issue while building a monetization layer for AI teams.

Small pricing changes would require engineering. Add a commitment. Adjust overage. Create a subscription tweak, everything.

So we tried something different. We built an MCP server for Flexprice.

Now every API operation is exposed as structured MCP tools. If you use Cursor, Claude Code, VS Code, Gemini, Windsurf, or any MCP-compatible client, you can connect your Flexprice account and operate billing through your AI assistant!

You can:

• Create subscriptions • Configure commitments and overages • Issue invoices • Post usage events • Manage wallets and more through your IDE or assistant now!

What we like most about this is that pricing is no longer locked behind engineering.

A non-technical co-founder or someone from finance can prompt a change, and it executes safely through authenticated tools.

Curious what HN thinks about prompt-operable infra like this. Happy to answer questions! Feel free to check out our docs for more info: https://docs.flexprice.io/docs/connect/mcp-server

Comments URL: https://news.ycombinator.com/item?id=47144263

Points: 3

# Comments: 0

The founder of OpenClaw ( Peter Steinberger) joining OpenAI is more than a hiring update. It reflects a structural pattern in AI.

Open-Source has quietly become the ecosystem’s proving ground. Builders experiment in public, take risks without institutional constraints, and earn credibility at the edge. When their work reaches a certain threshold, the largest labs absorb the talent.

It is efficient and strategic. It is also centralizing.

Every time a strong open-source founder moves to a frontier lab, the center grows stronger. The edge becomes slightly thinner. This does not make the decision wrong. Ambitious engineers will always gravitate toward ambitious problems. In many ways, it validates that building in the open still produces serious technical depth.

But we should be clear about the dynamic.

Open-Source in AI is increasingly becoming a pipeline. Build publicly, prove capability, and get pulled into one of a few centralized institutions. The gravity is strong, and it keeps reinforcing itself.

The real question is not whether talent emerges in the open. It clearly does. The question is whether it stays. Because the future of AI will not just be shaped by model performance. It will be shaped by where power consolidates.

And moves like this make that direction harder to ignore!