Hacker Newsnew | past | comments | ask | show | jobs | submit | jamescontrol's commentslogin

That is a huge red-flag. While I understand that they will do some policing/censoring, this is way beyond what I would consider acceptable.

They can have a different price plan for agentic stuff, but these things where they “accidentally” whoops match on specific keywords and trigger extra usage charges is giving a evil-microsoft-vibe


What I don't quite understand is why would one of the most advanced AI labs use rudimentary broken text match heuristics to track and detect abuse. Why not run simple inference on actual turns out of band, and if abuse is detected, adjust the quotas semi-retroactively.

> What I don't quite understand is why would one of the most advanced AI labs use rudimentary broken text match heuristics to track and detect abuse.

It's vibe-coded. What's hard about understanding that?


They’re idiots who hacked together a shockingly useful tool by leveraging the billions of dollars they received from shamelessly hyping up chatbots. The Claude Code leak makes this very clear.

Pretty wild to say that the company with one of the best models (arguably the best) is a bunch of idiots.

You seem to be implying that the company that employs the best chemists should therefore also make the best cakes. I don't see an obvious reason why this should hold true. I think it's fair to ridicule a bunch of chemists acting as master patissiers.

> Pretty wild to say that the company with one of the best models (arguably the best) is a bunch of idiots.

It would be pretty wild if they didn't considering all the money thrown at them!

You're looking at one of the largest investments business (as a collective) has ever made. They had better be one of the forerunners in the space :-/


And you think with all of this money they are employing idiots?

They're completely vibe-coding one of their flagship products. It's not unreasonable to consider that the people who took that decision are, indeed, idiots.

The people working on the models almost certainly aren't the same people writing the code for their harness.

Even idiots can succeed if you uncritically funnel them hundreds of billions of dollars.

You can't just burn money in a pit to get the best AI model out. Undoubtedly some of the smartest people in the world are working on frontier AI.

> most advanced AI labs use rudimentary broken text match

> It's vibe-coded

I called this out when I saw Claude Code CLI source code reach for regex on a certain task a while back and got told it was very unlikely that nobody reviewed the diff. Looks like the bar was lower than imagined.


Maybe running additional inference on all sessions to detect OpenClaw usage would require spending more money than they would save with that detection in the first place (which is the original goal). I also suspect the Claude Code team is just a regular software team without immediate access to ML pipelines (or competence to run them) to quickly develop proper abuse detection systems with extensive testing (to avoid false positives, which people would also complain about), and they're under pressure by the management to do something right now, so a regex is all they can do within those constraints.

Fairly certain it went like this:

Somebody at the top freaked out.

Somebody had to do something fast.

A prompt was given to Claude Code to fix Claude Code to stop Claude Code from being used for non-Claude Code stuff.

Commit made. Emergency release.

OpenClaw number went down. Everybody's pre-IPO stock options continued to go up.


> Why not run simple inference on actual turns out of band, and if abuse is detected, adjust the quotas semi-retroactively.

I suppose because running inference of any kind is a helluva lot more demanding than running a regex and less deterministic.


This is fascinating because it makes me think OpenClaw is something of a trojan horse aimed at draining Anthropic's resources. For them to go to this length to stop OpenClaw usage raises some interesting questions and a precedent for closed model vendors.

Why do they treat is as a trojan horse? More OpenClaw usage means more Claude usage. Isn't more Claude usage what Anthropic wants?

Not when their customers are paying a flat rate subscription.

Calling current AI subscription services (especially Claude) "flat rate" (implying infinite access for a flat fee) is misleading. There are pretty strict hourly, daily, weekly, and monthly limits. So there is a pretty easy-to-reach limit for all these subscriptions. They're hardly unlimited, and given how easy it is to run into limits, it's likely not super complicated/low-stddev for an accounting department to figure out avg cost per customer.

Is flat rate the best way to describe it when there's actually a few different tiers and each one has hard coded rate limits?

Within each tier, each marginal token is an expense with no marginal revenue to offset. So yes. The platonic ideal for any subscription business model is zero usage.

I run an AI subscription business and we have our pricing set in a way that we make an acceptable profit even if all users were to max out their given usage

Of course. My point is that your profit still decreases as you approach max usage, ceteris parabis. It may be acceptable but it is less. Your costs are variable and your revenue is fixed (at least on a unit basis).

No, not for us because we have a lot of different tiers and so as their usage increases they buy bigger and bigger plans.

Why is this a red flag? OpenClaw is basically automated abuse of their subscription plans. This is entirely reasonable.

Silently changing the billing mode based on keyword presence in free text is a garbage implementation

This is why I’m always wait as long as possible to update major versions, seems like there is fuckups big and small in every single major macOS update.


I looked at it for company chat and data, but those weird limits in functionality making in unusable was just too much, so them doing this too is not really surprising. Are they low on money?


What a terrible name and the only reason I know it’s not an aprils fool is because it’s November.


How will this not be used by trump/republicans?


Did anyone say it won't? Welcome to America the land of trump


Nope. I like the idea, but that is just still way too bulky, and just the fact that this is meta makes me want to steer clear.


This reminds me of the cars being produced in Russia at the moment…


I think he means, what causes apple to trigger those notifications. I don’t remember ever seeing that prompt, at least not without myself doing some action to trigger it.


>I think he means, what causes apple to trigger those notifications

Yeah, that's what I tried to guess too. Like, maybe those are sent periodically?

Could be there's some heuristics like "logged in from a different city" or such, too.


Can you explain? Having a tool to detect changes and create a migration doesn’t sound bad? In a nutshell thats how django migrations work as well, which works really well.


This is one of two reasons why I decided to not try anything, ever.

1. What if it’s the most amazing thing ever, like ADHD medication, what if you get insanely productive, and you then have to go back to being mediocre?

2. What if it triggers something bad. If I get 10% worse at focusing on a solving a problem?

Nope


> 1. What if it’s the most amazing thing ever, like ADHD medication, what if you get insanely productive, and you then have to go back to being mediocre?

Probably a good time to remind everyone that ADHD medication will not actually make you insanely productive forever. Good for treating genuine ADHD, but trying to use it like a limitless pill just leads to tolerance, dependence, and burnout.

The more involved doctors will actually start people on low doses and titrate up so patients don’t get the wrong idea about what the medicine is supposed to be doing for them. It’s usually the people who took too high of a dose they borrowed from a friend in college or something who have an idea that stimulants are a magic limitless pill without consequences.


That's probably not true. You'll try (and probably already have) all sorts of drugs which are deemed, by some government authority, to be "safe" which have all sorts of nasty side effects. Many "totally safe" drugs have much worse effects than psychedelics.

You will get 10% worse at focusing on solving a problem and then 20% just as a result of aging and there's nothing you can do about it. Your eyes will get old, your vitreous will start to detach and you'll start to see what looks like worms floating around in your eyes. Again, nothing you can do about it. Your hearing too will slowly start to degrade. High frequencies will fade to dust and those below them will become more apparent and, in some cases, become quite annoying.


I can speak to #1. If you take a normal amount, you will not be productive at all in terms what we all feel society demands of us. All ideas of productivity fall away when you trip. Social pressures fall away, you become a kid again in a way. Fascinated by looking at grass breathing in a meadow. You still think familiar thoughts, but in a vastly different way.

Now, if you microdose that’s a different story. I tried microdosing for a month years back at a remote job. I found it made me more able to connect with others. Like I didn’t want to hop off a meeting right away. It gave me some extra creativity when coding, but didn’t increase the output of my code.

I completely respect anyone who doesn’t want to try though, there are even some psychedelics that scare me and I’ve done quite a bit!


>All ideas of productivity fall away when you trip.

Counter anecdote: I tend to compulsively clean my apartment on mushrooms.


That's why I never smoked a single nicotine cigarette. If I like it, I'll be an addict. If I don't, it was wasted.


I’ve done similar. I’ve actually smoked many cigarettes but I’ve never bought nicotine myself. I’ve seen both of my parents piss away thousands of dollars and years of their lives to cigarettes, but damn a hit of a vape can be fun with friends


Or #3: You didn’t personally meet the chemist to know what you really took and now you don’t even know what’s causing your life’s problems.


Without getting into opinions this made me think of a Richard Feynman quote about abstaining from substances, along the lines of "I like how I think and didn't want to do anything to possibly change that." To each their own and much respect for doing what works best for you.


and on the other side you may get insanely creative and change your life for the better and build something valuable to the humanity.

Tradeoff is clear for me.


This will never happen, and is an absurd cope for people that just want their fix and to be addicts without feeling bad.


I’m sorry, but are you saying that drugs categorically can not lead to inspiration that can help creativity?

Aldous Huxley would like a word…


how would you argue the fact that you don't develop addiction to this stuff?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: