eskibars

Hi HN,

I've spent a lot of my career working in open source and I want to give back. Recently, I launched https://zeroquarry.com, which is a tool that helps you find 0-days in your software by through AI by tracing incoming requests down to the lowest level of implementation. The tool will do things like automatically suggest patches if you want or be more hands-off if you prefer. It will automatically produce POCs of the vulnerabilities for you to test.

I'm giving the tool away for free to any open source project. If you ping me directly (shane at the domain) with some details on what you're scanning, I'll increase credits or provide additional help.

I received a call from a Backblaze sales person asking for "27 seconds of my time" to my personal cellphone. My cell is on the US "do not call" list, but was dialed anyway. The conversation sounded very strange, and I asked directly whether it was an AI agent. It responded that it was. I asked for a supervisor, or to a compliance/legal contact and it then refused, but did offer e-mailing support@

Very upset by this, and will never use Backblaze (or anyone that uses this type of telesales) approach

Currently, my product offers retrieval augmented generation as a service. Anyone doing LLM-based apps has probably noticed that the usage costs are typically on a per-token basis, which seems very "fair" in that each token takes some time and more tokens = more time = higher underlying costs.

However, discussions I've had with some CIOs is that this yields a very unpredictable billing cycle and that they have a hard time mapping tokens to business outcomes. This was why we originally priced by request + storage, and have modeled average token consumption by our users. We end up with larger margins for smaller requests/responses and smaller margins for larger requests/responses, which is obviously less fair but more predictable.

Curious about how you have felt on being charged per token? Are we making the right call by making things more predictable or is it better to be less predictable but more directly reflect the underlying costs?