3 ms·

I don't know why people here can't accept the simple fact that AI companies are offering cheap "unlimited" plans as a loss leader to tie you to their ecosystem,

by paxys 8mo ago

I don't know why people here can't accept the simple fact that AI companies are offering cheap "unlimited" plans as a loss leader to tie you to their ecosystem, and then make up for it via add-ons, upsells, ads etc. If you use those API tokens to access external services it defeats the purpose. The hack may have worked so far, mainly because no one was checking, but they are all going to tighten the access eventually (as Anthropic and Google have already done). Either stick to first party products or pay for API use.

techpression 8mo ago
When reading HN I get the impression that a lot of people are convinced monthly plans are very profitable for the companies, I don’t have any numbers but to me it always seemed like a bait and switch or ”bait and make you pay with your data too”.
- vineyardmike 8mo ago
  I'll bite. I suspect that these plans aren't as intensely subsidized as people assume. I believe that API usage is probably also not subsidized at all. First, yes, subs are probably subsided, but I bet a significant % of users are profitable to serve, especially the "chat" users who don't use dev tools and have short context window conversations. Yes, I think the subs also exist as a driver to get lock-in and market share. Claude Code, for example, is very good and I stopped using their competition when they released their superior product. That said, I assume that (1) their long-term goal is to create cheaper-to-serve models that fit within their pricing targets, and use the (temporarily) subsidized subscriptions to find the features and costs that best serve the market. Maybe even while capturing more margin on the API in comparison (eg keep API prices high while lowering cost to serve a token). I've largely stopped using Opus, and sometimes even chose to use Haiku, because the cheaper models are fast and usually serves my needs. It's very possible to work all-day and barely hit the usage limits with Haiku on the $20/mo option. Long term, that could be profitable outright. And (2) subscriptions with lower SLOs than API calls have the potential to provide "infill" usage for high fixed-cost GPUs as an alternative to idling, similar to their batch APIs. I'd believe that overnight usage limits could/should be higher than during California work-hours. I assume most big providers have pre-paid fixed cost servers, so pumping more tokens through an otherwise idle GPU is "free". They can also do a lot more cost-optimization behind the scenes, such as prompt caching, to reduce the cost of tokens.
  - NitpickLawyer 8mo ago
    > I'll bite. I suspect that these plans aren't as intensely subsidized as people assume. I believe that API usage is probably also not subsidized at all. First, yes, subs are probably subsided, but I bet a significant % of users are profitable to serve, especially the "chat" users who don't use dev tools and have short context window conversations. Yes, I think the subs also exist as a driver to get lock-in and market share. Claude Code, for example, is very good and I stopped using their competition when they released their superior product. I somewhat agree, somewhat disagree with this. I think API based is not subsidised. If you do some basic napkin math they should have enough room there to serve the models below cost if the models aren't insanely large (you can compare with 3rd party openrouter offerings and have an idea of what $/Mtok you can serve per model size. e.g. Haiku level models can be ~700B tokens and still be profitably served) I think 20-200$ all-you-can-prompt are likely subsidised. If you track token usage (there are many 3rd party tools that do this) you can get 4-5x the API usage out of them (it used to be even higher before they added weekly limits. People were seeing 10-20x usage). Now I think that's a bit tough to make the napkin math work out. I've compared sessions served over API with sessions from subscriptions, and you get much more usage out of them, even with 5h / weekly limits. Strictly for coding, I think they're subsidising them. I somewhat disagree that they're doing it for market share / user lock-in. I think signals and usage trends are much more valuable for them. While there might be user retention for "casual" users (i.e. web) I think the power users in coding will move as soon as the competition has a better product. So at the end of the day having data to improve models and have the "best" model in a niche is more productive than retaining users with an inferior product. That is an assumption tho, and there isn't much math you can do to figure that out from the outside.
    - int_19h 8mo ago
      One thing to remember is that not all users are going to max out their plan. This is more likely to occur on $20 plans though, especially since those are often necessary to unlock the more useful features (e.g. deep research) so people might be paying for that even if they don't actually use the tokens. OTOH someone who's paying $200 will likely want to squeeze the most out of their subscription for that amount of money. So I wouldn't be surprised if it turns out that $20 users are subsidizing the $200 ones.
CuriouslyC 8mo ago
OpenAI and the Chinese companies let you all you can eat openly. Anthropic's lead vs OAI is slight and these things are going to homogenize quickly. The market is going open and the people trying to keep it closed are just generating ill will pointlessly.
- javascriptfan69 8mo ago
  So what are they supposed to do? Race to burn as much cash as possible in hopes that the other goes bankrupt first? These models aren't profitable at the fixed subscription tiers.
  - andersmurphy 8mo ago
    That is the plan yes.
  - beAbU 8mo ago
    > Race to burn as much cash as possible in hopes that the other goes bankrupt first? This has been standard in the VC playbook for the last decade or so. The only moat these companies have is the size of their war chests.
- NewsaHackO 8mo ago
  >OpenAI and the Chinese companies let you all you can eat openly. You say this, but I guarantee that when they do offer a plan similar to Google/Anthropic's dedicated coding "unlimited" subscription, they will do the exact same thing. Maybe they will let OpenClaw in as a first party because of their partnership with the creator.
  - LinXitoW 8mo ago
    But none of these are unlimited, that was never the expectation. It's a flat rate for a flat (but hidden) amount of usage. What's disgusting is that they want the good parts of subs (low usage subs), but then just ban the bad parts (high usage people). I don't care whether that's technically possible, it's incredibly scummy.
stevage 8mo ago
No one is shocked that they don't allow this. Everyone is shocked that they silently, permanently banned the user with no recourse and it took significant effort even to find out that much.
- DavidPiper 8mo ago
  Sorry to be that guy, but given how often Google has done this for lesser infringements (some reported here on HN), is anyone really "shocked" by the permabans? The apparent shock around this sort of thing always feels like cope for the fact that we (myself included) understand the power imbalance between Google and its customers but don't want to admit it. There's plenty of evidence at this point, and I feel like we should be using that emotional energy to actually do something about it (like switching providers for critical personal services, for example).
  - s3p 8mo ago
    >Sorry to be that guy, but given how often Google has done this for lesser infringements (some reported here on HN), is anyone really "shocked" by the permabans? Yes, we are surprised. Google understands our surprise, too. To quote the Google employee another commenter mentioned: "We understand that a subset of these users were not aware that this was against our ToS and will get a path for them to come back on" https://x.com/_mohansolo/status/2025766889205739899 https://x.com/_mohansolo/status/2025766889205739899
    - s3p 8mo ago
      Update: Version 1.19.6, released Feb 26, 2026! >Account Remediation Pathway >Introduced a formal remediation process for accounts suspended due to Terms of Service violations. See our official announcement.
danny_codes 8mo ago
Seems like a hassle when open source models are just as good. Can go with any hosting provider. Might have to wait 3-4 weeks for them to duplicate whatever Anthropic is doing with token caching. But then you get 10x cheaper inference. I feel like this game is just a hot potato, can you get retail to hold the bag game
- disiplus 8mo ago
  I have them all. They're not just as good. Whoever tells you that looked only at the benchmarks, not real use. They all fall short at some point. Kimi K2.5 is the best one, but it's still not at the level of what Anthropic released with opus 4.5.
  - danny_codes 8mo ago
    We’ll have to give it 3 weeks.
- paxys 8mo ago
  Open models are very far in performance from the top models of Anthropic, OpenAI and Google. And that's skipping over the fact you need somewhere to host them.
neop1x 8mo ago
Recent open-weights models(MiniMax, Kimi K2, GLM, Mistral) are also quite good, can be self-hosted or accessed through 3rd-party hosters or OpenRouter and they are sufficient for most of the tasks. Just stop paying overpriced "unlimited" subscription bul**hit.

by paxys 8mo ago

techpression 8mo ago

When reading HN I get the impression that a lot of people are convinced monthly plans are very profitable for the companies, I don’t have any numbers but to me it always seemed like a bait and switch or ”bait and make you pay with your data too”.

vineyardmike 8mo ago
I'll bite. I suspect that these plans aren't as intensely subsidized as people assume. I believe that API usage is probably also not subsidized at all. First, yes, subs are probably subsided, but I bet a significant % of users are profitable to serve, especially the "chat" users who don't use dev tools and have short context window conversations. Yes, I think the subs also exist as a driver to get lock-in and market share. Claude Code, for example, is very good and I stopped using their competition when they released their superior product. That said, I assume that (1) their long-term goal is to create cheaper-to-serve models that fit within their pricing targets, and use the (temporarily) subsidized subscriptions to find the features and costs that best serve the market. Maybe even while capturing more margin on the API in comparison (eg keep API prices high while lowering cost to serve a token). I've largely stopped using Opus, and sometimes even chose to use Haiku, because the cheaper models are fast and usually serves my needs. It's very possible to work all-day and barely hit the usage limits with Haiku on the $20/mo option. Long term, that could be profitable outright. And (2) subscriptions with lower SLOs than API calls have the potential to provide "infill" usage for high fixed-cost GPUs as an alternative to idling, similar to their batch APIs. I'd believe that overnight usage limits could/should be higher than during California work-hours. I assume most big providers have pre-paid fixed cost servers, so pumping more tokens through an otherwise idle GPU is "free". They can also do a lot more cost-optimization behind the scenes, such as prompt caching, to reduce the cost of tokens.
- NitpickLawyer 8mo ago
  > I'll bite. I suspect that these plans aren't as intensely subsidized as people assume. I believe that API usage is probably also not subsidized at all. First, yes, subs are probably subsided, but I bet a significant % of users are profitable to serve, especially the "chat" users who don't use dev tools and have short context window conversations. Yes, I think the subs also exist as a driver to get lock-in and market share. Claude Code, for example, is very good and I stopped using their competition when they released their superior product. I somewhat agree, somewhat disagree with this. I think API based is not subsidised. If you do some basic napkin math they should have enough room there to serve the models below cost if the models aren't insanely large (you can compare with 3rd party openrouter offerings and have an idea of what $/Mtok you can serve per model size. e.g. Haiku level models can be ~700B tokens and still be profitably served) I think 20-200$ all-you-can-prompt are likely subsidised. If you track token usage (there are many 3rd party tools that do this) you can get 4-5x the API usage out of them (it used to be even higher before they added weekly limits. People were seeing 10-20x usage). Now I think that's a bit tough to make the napkin math work out. I've compared sessions served over API with sessions from subscriptions, and you get much more usage out of them, even with 5h / weekly limits. Strictly for coding, I think they're subsidising them. I somewhat disagree that they're doing it for market share / user lock-in. I think signals and usage trends are much more valuable for them. While there might be user retention for "casual" users (i.e. web) I think the power users in coding will move as soon as the competition has a better product. So at the end of the day having data to improve models and have the "best" model in a niche is more productive than retaining users with an inferior product. That is an assumption tho, and there isn't much math you can do to figure that out from the outside.
  - int_19h 8mo ago
    One thing to remember is that not all users are going to max out their plan. This is more likely to occur on $20 plans though, especially since those are often necessary to unlock the more useful features (e.g. deep research) so people might be paying for that even if they don't actually use the tokens. OTOH someone who's paying $200 will likely want to squeeze the most out of their subscription for that amount of money. So I wouldn't be surprised if it turns out that $20 users are subsidizing the $200 ones.

CuriouslyC 8mo ago

OpenAI and the Chinese companies let you all you can eat openly. Anthropic's lead vs OAI is slight and these things are going to homogenize quickly. The market is going open and the people trying to keep it closed are just generating ill will pointlessly.

javascriptfan69 8mo ago
So what are they supposed to do? Race to burn as much cash as possible in hopes that the other goes bankrupt first? These models aren't profitable at the fixed subscription tiers.
- andersmurphy 8mo ago
  That is the plan yes.
- beAbU 8mo ago
  > Race to burn as much cash as possible in hopes that the other goes bankrupt first? This has been standard in the VC playbook for the last decade or so. The only moat these companies have is the size of their war chests.
NewsaHackO 8mo ago
>OpenAI and the Chinese companies let you all you can eat openly. You say this, but I guarantee that when they do offer a plan similar to Google/Anthropic's dedicated coding "unlimited" subscription, they will do the exact same thing. Maybe they will let OpenClaw in as a first party because of their partnership with the creator.
- LinXitoW 8mo ago
  But none of these are unlimited, that was never the expectation. It's a flat rate for a flat (but hidden) amount of usage. What's disgusting is that they want the good parts of subs (low usage subs), but then just ban the bad parts (high usage people). I don't care whether that's technically possible, it's incredibly scummy.

stevage 8mo ago

No one is shocked that they don't allow this. Everyone is shocked that they silently, permanently banned the user with no recourse and it took significant effort even to find out that much.

DavidPiper 8mo ago
Sorry to be that guy, but given how often Google has done this for lesser infringements (some reported here on HN), is anyone really "shocked" by the permabans? The apparent shock around this sort of thing always feels like cope for the fact that we (myself included) understand the power imbalance between Google and its customers but don't want to admit it. There's plenty of evidence at this point, and I feel like we should be using that emotional energy to actually do something about it (like switching providers for critical personal services, for example).
- s3p 8mo ago
  >Sorry to be that guy, but given how often Google has done this for lesser infringements (some reported here on HN), is anyone really "shocked" by the permabans? Yes, we are surprised. Google understands our surprise, too. To quote the Google employee another commenter mentioned: "We understand that a subset of these users were not aware that this was against our ToS and will get a path for them to come back on" https://x.com/_mohansolo/status/2025766889205739899 https://x.com/_mohansolo/status/2025766889205739899
  - s3p 8mo ago
    Update: Version 1.19.6, released Feb 26, 2026! >Account Remediation Pathway >Introduced a formal remediation process for accounts suspended due to Terms of Service violations. See our official announcement.

danny_codes 8mo ago

Seems like a hassle when open source models are just as good. Can go with any hosting provider. Might have to wait 3-4 weeks for them to duplicate whatever Anthropic is doing with token caching. But then you get 10x cheaper inference. I feel like this game is just a hot potato, can you get retail to hold the bag game

disiplus 8mo ago
I have them all. They're not just as good. Whoever tells you that looked only at the benchmarks, not real use. They all fall short at some point. Kimi K2.5 is the best one, but it's still not at the level of what Anthropic released with opus 4.5.
- danny_codes 8mo ago
  We’ll have to give it 3 weeks.
paxys 8mo ago
Open models are very far in performance from the top models of Anthropic, OpenAI and Google. And that's skipping over the fact you need somewhere to host them.

neop1x 8mo ago

Recent open-weights models(MiniMax, Kimi K2, GLM, Mistral) are also quite good, can be self-hosted or accessed through 3rd-party hosters or OpenRouter and they are sufficient for most of the tasks. Just stop paying overpriced "unlimited" subscription bul**hit.