Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

HERMES.md in commit messages causes requests to route to extra usage billing

github.com

Post reply on HN
Print view View original

371–380 of 568 posts

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#371

Post by stefan_ »

csoups14 wrote: post #352
jiggawatts wrote: post #346

Earlier quoted context omitted.

Please do explain why someone at Anthropic decided, on purpose, to write code that says something along the lines of: "if ( git_history_str contains "HERMES.md" ... )" then { bill more money } Somebody (or something ) wrote this code. This bug wouldn't be happening for any other reason. It's not a glitch, an oversight, a feature gap, or a temporary outage. It is a piece of written code in your system. Everyone here i…

Would imagine it's the simplest answer: they're flying by the seat of their pants, there's 1000 things happening every day that demand attention and there's not enough of it to go around. They toss their LLM at it, give it a cursory glance, and ship it. A quick glance at the Claude Code source code bears the result of this process out. The fundamental question is, if their model is so powerful, why do they keep fucki…

Hermes is one of these OpenClaw clones, so this was certainly intentional, not a model hallucinating something.

I think the problem is clear. Anthropic saw their usage go up much more than their capacity could handle. There are a few tried and true solutions to this, like "increase the price" or "restrict signups so you can guarantee service to what you have already sold".

Then there is the "large scale fraud" option, where you materially change and degrade the service you have already sold. Just because you have obfuscated and mislead in how you describe the product you are selling doesn't mean you get to capture the cash flow of 1 year subscriptions then not honor that contract for the full duration.

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#372

Post by Arainach »

jayd16 wrote: post #335
Arainach wrote: post #88

Earlier quoted context omitted.

If anyone with principles quit the moment a company did something bad, you'd be left with only people who are cynical and/or bad and/or sufficiently indentured to be unable to push back against management, and there would be no hope of the company ever improving. Sure, everyone probably has their own personal line such as "will quit if my employer is declared complicit in genocide by the UN", but bad customer service…

But they're not advocating. They're claiming they can do nothing. Quitting in protest would be more advocation.

Who says they're not advocating? Who says they were aware of this before today?

Extend this to other disciplines - if everyone who cared about security resigned every time leadership pushed to rush something out without proper testing, the world would be a worse place. Sticking around and continuing to try to change the culture is how good companies are made.

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#373

Post by quietsegfault »

trq_ wrote: post #264

Hey everyone, Thariq from the Claude Code team. We've been on this since the bug surfaced. Everyone affected is getting a full refund and an extra grant of usage credits equal to their monthly subscription as our apology. You can see my original post here: https://x.com/trq212/status/2048495545375990245 . We’re still working on sending emails to everyone affected. Our support flow wasn't set up to route a complex bug…

I’ve had similar terrible experiences with the Claude support bot when my usage limit was disappearing after a few minutes using Sonnet. I asked for help, asked for escalation, asked for a human, anything. All I got was a non-answers from an AI. I won’t spend real money on Claude now. I’m ok with losing $20 if there’s a rug pull of one way or another, but not $200.

Please, please, please hire more humans with the actual ability to do the right thing for support if your AI agents can’t do the job.

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#374

Post by justinclift »

areoform wrote: post #185
ecshafer wrote: post #3

> However, I need to let you know that we are unable to issue compensation for degraded service or *technical errors* that result in incorrect billing routing. This is very surprising. I've never seen a legitimate business not give refunds for technical errors of their own fault. Minimum Anthropic should credit the full amount to them.

This billing cycle my account was billed an extra $200. I investigated. I was being for a Claude Max gift subscription that has been sent to – what appears to be – a randomly generated 27 char alphanumeric icloud email account that bounces. Apparently, Anthropic doesn't have a centralized process that allows you to approve, see or revoke "gift cards." And no I can't use this hypothetical gift card. Because I can't se…

Maybe worth trying some of their ~legal-ish email addresses?

* notices@anthropic.com : from https://www.anthropic.com/legal/commercial-terms)

* usersafety@anthropic.com : from https://www.anthropic.com/legal/aup

* marketing@anthropic.com : https://www.anthropic.com/legal/consumer-terms

* disclosure@anthropic.com : from https://www.anthropic.com/responsible-disclosure-policy

* dpo@anthropic.com : from https://www.anthropic.com/legal/privacy

* pubsec@anthropic.com : from https://www.anthropic.com/news/claude-in-amazon-bedrock-fedr...

There's also their generic consumer ones, though I'd rate them as unlikely to do anything useful:

* support@anthropic.com : from https://www.anthropic.com/legal/consumer-terms

* privacy@anthropic.com : from https://www.anthropic.com/legal/privacy

And this out of left field one. They seem like actual lawyers:

* anthropicprivacy@bkl.co.kr : from https://www.anthropic.com/legal/privacy

---

Interestingly, Anthropic's "Trust Center" has an "Evidence of Insurance" document listed under "Other documents": https://trust.anthropic.com/resources#69eff53d22c228b34e5379...

Looks like you need to "Request Access", but if it's an automated system then it may give you access. And there _might_ be insurance contacts listed there who would be interested in this. :)

---

Follow up note -> Yep, it's automated and DOES give access to their docs. ;)

Their insurance levels don't really seem to be anywhere near what I'd expect frankly. To me, they look much lower than even entry level mandatory company insurance for brand new businesses at least in Australia. o_O

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#375

Post by root_axis »

Henchman21 wrote: post #299
root_axis wrote: post #223

Earlier quoted context omitted.

More likely its just an LLM hallucination, not a real policy that Anthropic has. Unfortunately for them, it's a bad look to showcase one of the main failure modes of their product in their own business process.

If they've let their AI write the policy, and then they repeat that as policy, how exactly is this an "LLM hallucination" and not a real policy?

It's the same thing. Whether it was hallucinated upstream or in situ, the point is that it's not a real policy that the business adheres to, just something the LLM spat out.

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#376

Post by quietsegfault »

mystraline wrote: post #312
whicks wrote: post #278

Earlier quoted context omitted.

Thanks for the follow up here and the transparency. For those of us not on X, what are the best communication channels for us to follow this sort of communication?

I'd recommend a good credit card like Amex, and a lawyer. These fucks only respond when they get bad publicity.

Amex, like basically all other card issuers, have essentially stopped giving customers preference in chargebacks since 2020 or so. What used to be solid advice now rings hollow - you’re more likely to be asked for information that not available to you than allowing your chargeback to go through.

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#377

Post by quietsegfault »

Rebelgecko wrote: post #282
pshirshov wrote: post #279

Earlier quoted context omitted.

But why did you say that > I need to let you know that we are unable to issue compensation for degraded service or technical errors that result in incorrect billing routing. What prevents you from issuing compensations?

As a large language model, their support is not allowed to issue compensation

I know this is a joke, but Amazon’s bots give me compensation literally all the time when something goes wrong. It’s possible.

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#378

Post by muwtyhg »

boc wrote: post #321
dev_l1x_be wrote: post #42

What a series of disasters that are happening at Anthropic nowadays. I am not even sure what is going on with Opus 4.7 I had to switch back to 4.6 and 4.6 was already a downgrade (anecdotal + the github thread with the harness changes). I am cancelling my subscription as it is impossible to justify these degradations and paying for a subpar service especially now that we have at least 3 more models that are as good a…

[flagged]

You realize the thread you are commenting on is not about scaling or downtime, but about a billing bug that Anthropic refused to fix until it become a Streisand effect?

If you're happy to continue paying a company that has demonstrated it will steal your money, admit it, and refuse to return it, more power to you. The AI industry is moving fast enough that there will be plenty of players to pick up customers who don't want to be robbed.

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#379

Post by quietsegfault »

root_axis wrote: post #375
Henchman21 wrote: post #299

Earlier quoted context omitted.

If they've let their AI write the policy, and then they repeat that as policy, how exactly is this an "LLM hallucination" and not a real policy?

It's the same thing. Whether it was hallucinated upstream or in situ, the point is that it's not a real policy that the business adheres to, just something the LLM spat out.

Sure, it’s a real policy. It came from their website, from the official means of support.

Re: HERMES.md in commit messages causes requests to route to extra usage billing

#380

Post by mcmcmc »

registeredcorn wrote: post #94
culi wrote: post #51

Earlier quoted context omitted.

And then you use the smallest, cheapest local model to keep their AI bot busy

My insurance company and Synology would be my first targets. I'd gladly throw ~1k at each. Of course, I suspect the true business model to be to do nothing. You sell the "service" to people customers, but your enterprise customers pay you a subscription fee to not execute the order. ELaaS: Everybody Loses as a Service

aka Rent Seeking as a Service
Post reply on HN