I'm trying to cancel my Claude plan right now, because mid-research I got hit with a one hour time out for the first time, and I realized just how little I need the "expensive" research stuff and just want basic language refinement and some basic augmented knowledge search to get more terms to search the web. And all that basic stuff gets cutoff when you hit some invisible barrier that you have no way of controlling or monitoring.
But I literally can not cancel. Trying the app says "you signed up on a different platform, go there" but it doesn't tell me which platform that might be.
Trying to cancel on mobile web gives several upgrade options but no cancel options.
So, do I need to call my credit card? This is the worst dark pattern on subscription I have seen of any service I have ever paid for!
Anthropic had a fairly positive image in my head until they cut off my access and are not giving me a way to cancel my plan.
Edit: after mucking with the Stripe credit card payment options I found a cancel plan button underneath the list of all invoices. So there is an option, I just had a harder time finding it then I have had with other services. Successfully cancelled!
The huge positive publicity boost that Claude Code gained during its unlimited phase was surely easily worth the slight negative publicity it is getting now. This post is just marketing copy by a small competitior that can't afford to do the same thing.
I think the difference this time is that there is no lock in with a contract like AT&T customers had when they bought iPhones.
Unlimited for startups work better because they have zero idea on load challenges that come in the future. And they don’t have much idea how well their product will be taken in the market.
Anthropic got the experience and decided they needed to maximize on reasonableness over customer trust. And they are a startup so we all get this.
OTOH there is no such thing as unlimited. Atoms in the universe are finite. Your use is finite. Your time is finite. Your abuse is limited and finite. You are a sucker for believing in the unlimited myth just like think others are suckers for believing in divine intervention or conspiracy theorists are suckers to believe in unlimited power.
I am not a heavy user of Claude in any way, sometimes I go days without using it. And I've hit the limit a week ago. Now with the limit changing to per week, I am 100% canceling my subscription. I'll likely spend a max of $5 using an API key.
When you offer "unlimited", you're immediately exposing yourself to the 0.1% of userbase that will absolutely treat it is unlimited. Whale users/hoarders. This has been a known problem since the dawn of hosting...hell, it goes back to times before computers.
One thing I miss for the other users, i.e. the casual users that never use anywhere near of their quota, is rollover. If you haven't used your quota this month, the unused will roll over to the next month.
Did Anthropic ever use the term unlimited? I understand the general frustration with the pattern, but it seems weird to put unlimited in quotes when it wasn’t the way Claude was sold.
I feel like this is where the heavyweights are really going to start throwing their weight around and basically drive the smaller startups out of town.
Can you really ever compete when you are renting someone else's GPUs?
Can you really ever compete when you are going up against custom silicon built and deployed at scale to run inference at scale (i.e. TPUs built to run Gemini and deployed by the tens-of-thousands in data centers around the globe)?
Meta and Google have deep pockets and massive existing world-class infrastructure (at least for Google, Meta probably runs their php Facebook thing on a few VPS dotted around in some random colos /s ) . They've literally written the book on this.
It remains to be seen how much more money OpenAI can burn, but we've started to see how much Anthropic can burn if nothing else.
I would absolutely do this if I ran anthropic. Of course unlimited implicitly means “unlimited without abuse.” There are always these “power users” who run it 25/8 and use all the resources, or sell access to others. To Anthropic: of course the “power users” are also going to be the top 5% of extremely online folks who are going to angrily pen long form tweet storms and this whole thing will die down soon for the other 95% of us. Weather the storm.
There is no free lunch when VC money is involved, and the new weekly rate limits in Claude code now are just the latest example of that. I knew there was going to be a rugpull and it definitely changes the psychology of using it when you know you could be locked out for multiple days. Also saying that "Less than 5% of users are affected" = anyone actually USING this tool for real work...
"I firmly believe that AI will not replace developers, but a developer using AI will replace a developer who does not."
Ugh, anyone who says that and really believes it can no longer see common sense through the hype goggles.
It's just stupid and completely 100% wrong, like saying all musicians will use autotune in the future because it makes the music better.
It's the same as betting that there will be no new inventions, no new art, no works of genius unless the creator is taking vitamin C pills.
It's one of the most un-serious claims I can imagine making. It automatically marks the speaker as a clown divorced from basic facts about human ability
I didn't look at the URL at first and was surprised when this turned in to an ad. Oh well!
> Stop selling "unlimited", when you mean "until we change our minds"
The limits don't go in to affect until August 28th, one month from yesterday. Is there an option to buy the Max plan yearly up front? I honestly don't know; I'm on the monthly plan. If there isn't a yearly purchase option, no one is buying unlimited and then getting bait-and-switched without enough time for them to cancel their sub if they don't like the new limits.
> A Different Approach: More AI for Less Money
I think it's really funny that the "different approach" is a limited time offer for credits that expire.
I don't like that the Claude Max limits are opaque, but if I really need pay-per-use, I can always switch to the API. And I'd bet I still get >$200 in API-equivalents from Claude Code once the limits are in place. If not? I'll happily switch somewhere else.
And on the "happily switch somewhere else", I find the "build user dependency" point pretty funny. Yes, I have a few hooks and subagents defined for Claude Code, but I have zero hard dependency on anything Anthropic produces. If another model/tool comes out tomorrow that's better than Claude Code for what I do, I'm jumping ship without a second thought.
84 comments
[ 3.9 ms ] story [ 60.8 ms ] threadBut I literally can not cancel. Trying the app says "you signed up on a different platform, go there" but it doesn't tell me which platform that might be.
Trying to cancel on mobile web gives several upgrade options but no cancel options.
So, do I need to call my credit card? This is the worst dark pattern on subscription I have seen of any service I have ever paid for!
Anthropic had a fairly positive image in my head until they cut off my access and are not giving me a way to cancel my plan.
Edit: after mucking with the Stripe credit card payment options I found a cancel plan button underneath the list of all invoices. So there is an option, I just had a harder time finding it then I have had with other services. Successfully cancelled!
"The real damage: You're not frustrating 5% of users—you're breaking trust with the exact people who drive growth and adoption."
"It's not rocket science. It’s our way of attracting users. Not bait and switch, but credits to try."
- note: "unlimited" does not mean free.
quote source: "Apple Just Found a Way to Sell You Nothing" https://www.youtube.com/watch?v=ytkk5NFZGjs
Unlimited for startups work better because they have zero idea on load challenges that come in the future. And they don’t have much idea how well their product will be taken in the market.
Anthropic got the experience and decided they needed to maximize on reasonableness over customer trust. And they are a startup so we all get this.
OTOH there is no such thing as unlimited. Atoms in the universe are finite. Your use is finite. Your time is finite. Your abuse is limited and finite. You are a sucker for believing in the unlimited myth just like think others are suckers for believing in divine intervention or conspiracy theorists are suckers to believe in unlimited power.
One thing I miss for the other users, i.e. the casual users that never use anywhere near of their quota, is rollover. If you haven't used your quota this month, the unused will roll over to the next month.
There is a case to be made that they sold a multiple and are changing x or rate limiting x differently, but the tone seems different from that.
Can you really ever compete when you are renting someone else's GPUs?
Can you really ever compete when you are going up against custom silicon built and deployed at scale to run inference at scale (i.e. TPUs built to run Gemini and deployed by the tens-of-thousands in data centers around the globe)?
Meta and Google have deep pockets and massive existing world-class infrastructure (at least for Google, Meta probably runs their php Facebook thing on a few VPS dotted around in some random colos /s ) . They've literally written the book on this.
It remains to be seen how much more money OpenAI can burn, but we've started to see how much Anthropic can burn if nothing else.
Ugh, anyone who says that and really believes it can no longer see common sense through the hype goggles.
It's just stupid and completely 100% wrong, like saying all musicians will use autotune in the future because it makes the music better.
It's the same as betting that there will be no new inventions, no new art, no works of genius unless the creator is taking vitamin C pills.
It's one of the most un-serious claims I can imagine making. It automatically marks the speaker as a clown divorced from basic facts about human ability
> Stop selling "unlimited", when you mean "until we change our minds"
The limits don't go in to affect until August 28th, one month from yesterday. Is there an option to buy the Max plan yearly up front? I honestly don't know; I'm on the monthly plan. If there isn't a yearly purchase option, no one is buying unlimited and then getting bait-and-switched without enough time for them to cancel their sub if they don't like the new limits.
> A Different Approach: More AI for Less Money
I think it's really funny that the "different approach" is a limited time offer for credits that expire.
I don't like that the Claude Max limits are opaque, but if I really need pay-per-use, I can always switch to the API. And I'd bet I still get >$200 in API-equivalents from Claude Code once the limits are in place. If not? I'll happily switch somewhere else.
And on the "happily switch somewhere else", I find the "build user dependency" point pretty funny. Yes, I have a few hooks and subagents defined for Claude Code, but I have zero hard dependency on anything Anthropic produces. If another model/tool comes out tomorrow that's better than Claude Code for what I do, I'm jumping ship without a second thought.
wait until their investors get fed up with pouring money down the drain and demand they make a profit from the median user
that model training and capex to build the giant DCs and fill them with absurdly priced nvidia chips isn't free
as an end user: you will be the one paying for it
“Claude is promising unlimited and it isn’t sustainable.”
“For a limited time only pay us $20 and get $80 worth of credits.”
Look at what these people on HN said!
Come on.