Hacker News new | ask | show | jobs
by torginus 124 days ago
Isn't the reason companies are doing this because they're offering tokens at a discount, provided they're spent through their tooling?

Considering the tremendous amount of tokens OpenClaw can burn for something that has nothing to do with sofware development, I think it's reasonable for Google to not allow using tokens reserved for Antigravity. I don't think there's such a restriction if you pay for the API out of pocket.

5 comments

> Isn't the reason companies are doing this because they're offering tokens at a discount, provided they're spent through their tooling?

Then maybe they should charge for that instead of banning accounts?

Google decided on their own business plan without any guns to their backs. If they decide to create a plan that is subsidized that's entirely on them.

So the issue is the same as Anthropic. They do charge for it though their API. The users, however, want to use the discounted "unlimited" flat rate through the first-party app instead, then get mad when they are told they have to use the same API every other third-party app does.
No, the problem is that the discounted rate exists in the first place. Essentially these are unfair business practices, product cross subsidization to ensure market dominance. See also: Microsoft and a whole bunch of other companies.

And once they've got their monopoly position there is inevitably the rug-pull. I wonder if some CPO somewhere actually had the guts to put a 'rug pull' item on the product roadmap.

It's not unfair its how every business works. When your product is new or not yet good enough and you want people to try it you give them discounts, or if you want to drive traffic to your service you also do the same.

Even traditional businesses do this with coupons. Is it unfair that Costco sells chickens for under cost because it drives usage to them?

Companies like Uber did use massive funding and price subsidization to try and kill competition and then take a monopoly, but it is hard to assert that this is what google is doing now. And given that other competitors in the space, Anthropic are doing the exact same thing again its not as though they are alone.

Also they could be subsidizing it because they want that usage type as it helps them train models better.

Chatgpt and gpt4 were all ran at a loss and subsidized people just didn't know that. Almost all of the llm companies have been selling 1 dollar of llm compute for 50 cents as they valued the usage, training data, and users more than making profit now.

This next generation of MOE and other newly trained models. Like opus 4.6, Cursor Composer 1.5, gpt 5.3 codex, and many of the others have been the first models where these companies are actually profitably serving the tokens at the api cost.

This year has been the switch where ai companies are actually thinking of becoming profitable instead of just focusing on research and development.

I'd agree with you if this was some new SaaS just opening its doors.

But Google are banning entire accounts, with years, even decades, of personal history, photos, even phone accounts and app development projects.

They very easily could just negate the anti-gravity access, which would be much, much more reasonable.

>But Google are banning entire accounts, with years, even decades, of personal history, photos, even phone accounts and app development projects.

Source? It seems to me only the anti-gravity access was blocked. The link says

> Our product engineering team has confirmed that your account was suspended from using our Antigravity service.

> there’s no way we can restore our accounts to use Antigravity anymore yeah?

Disclosure: I work at Google, but not on anything related to this.

> It's not unfair its how every business works.

Not. On both counts.

> Essentially these are unfair business practices, product cross subsidization to ensure market dominance.

Offering a different discounted rate for a service, though their first-party platform is not an unfair business practice whatsoever, though. The bar isn't what you disagree with, or what you think their motives are without any substantial proof. They could even make a honest argument that they can aggressively key-value cache default prompts from their own software reducing inference costs.

>See also: Microsoft and a whole bunch of other companies.

What does that have to do with Google?

Offering goods or services below the cost of their production is often illegal, though. It's called "dumping".

Although in this case it's probably impossible to define, given the complexity of calculating the true cost of tokens.

> Offering goods or services below the cost of their production is often illegal, though. It's called "dumping"

No.

Dumping is an international-trade term. It doesn’t even require pricing below cost, just aiming “to increase market share in a foreign market by driving out competition and thereby create a monopoly situation where the exporter will be able to unilaterally dictate price and quality of the product” [1].

Loss leaders are common in commerce and entirely legal, as are free trials. I struggle to think of a competent jurisdiction that bans them.

[1] https://en.wikipedia.org/wiki/Dumping_(pricing_policy)

So every company that is not immediately selling enough to cover its fixed costs and its variable cost should be illegal? Every company and every new initiative must be profitable from day one in your world?
This obviously cannot be true, otherwise Costco would have been sued to oblivion for “dumping” their rotisserie chickens.
And in this case the subsidy is paid for by tied sales from other users that don't actually use the service, which is another illegal business practice.
First of all, I doubt they’re losing money in inference. Even across subscriptions. This is a tired argument that has been repeated so many times on HN.

Second, that’s not what dumping means. It’s a specific term for international trade.

Third, it’s not illegal to sell something for below the cost to make it. That’s another common misunderstanding.

"PAYGO API access" vs "Monthly Tool Subscription" is just a matter of different unit economics; there's nothing particularly unusual or strange about the idea on its own, specific claims against Google notwithstanding.

Of course, Google is still in the wrong here for instantly nuking the account instead of just billing them for API usage instead (largely because an autoban or whatever easier, I'm sure).

The ban hammer is the scary part.

I am afraid of using any Google services in experimental way from the fear that my whole Google existence will be banned.

I think blocking access temporarily with a warning would be much more suitable. Unblocking could be even conditioned on a request to pay for the abused tokens

I doubt they would have the ability to charge them for it. They never signed up for api token usage?
Just because all you can eat buffet exists doesn't mean that the food is free or you can take away the food. The food exists in discounted rate only if you consider it unlimited food. For normal folks they make profit.

Claude code could possibly make profit because the average usage doesn't come close to exhausting the limits.

This exactly. I'm using 10% of my max plan on the weeks that I'm working a lot. Hit a 4-hour limit once over few months and never let it run overnight. And I'm very happy with my subscription
No, it's more like stuffing your pockets in all-you-can-eat buffet
There's nothing wrong or illegal with subsidizing products and that's not what Microsoft or others have gotten in trouble for doing. It's when they tie a strong monopolistic position (Windows) with bundling to prevent competition (Internet explorer). This is how Apple has operated with far tighter bundling and cross collateralization of their ecosystem without facing monopoly allegations. Google does not have a monopoly position in AI.
Its called economies of scale. When they server 200000 ai subscriptions they dont expect everyone to use the max. They expect some will use more and some will use less and at the end of the day it will even out. Thats how every service works that is for the masses. As soon as you want a guaranteed 1000 tokens you should pay for that.
If I pay for a subscription that guarantees 1000 tokens, I am paying for guaranteed 1000 tokens.
So you are saying a company should never reinvest profits in the company to support another money losing business until it’s profitable?

Should Netflix for instance not invested money from renting DVDs to invest in a streaming service?

Apple not use the profits it was making from selling Apple //e’s to create the Mac?

> So you are saying a company should never reinvest profits in the company to support another money losing business until it’s profitable?

If it makes it impossible to set up a competitor? Absolutely, yes.

> Should Netflix for instance not invested money from renting DVDs to invest in a streaming service?

Netflix was not priced below the cost of production from the beginning. You're confusing sustainable pricing and paying off all the capital spending immediately at launch.

A better example is Doordash when it was heavily subsidized by VC money: https://news.ycombinator.com/item?id=23216852 And it now faces several anti-trust lawsuits.

Netflix very much was priced below the cost of production for years and had to borrow money to make Netflix originals.
Subsidizing the cost of developing a product isn't necessarily bad, but predatory pricing that prohibits competition would be.

Not sure that this case is either. This is just idiots breaking the TOS.

So exactly what’s the difference between “predatory pricing” and pricing to gain customers and market share? Should Sony have to sell the first PlayStation off the line at $2000 (making up a number) so it can sell it at a marginal profit from day one or should it sell it below cost knowing that that over its lifetime if it stays at that price, it will both gain customers and sell at a profit in year 4 as the price of technology comes down and it gets economies of scale?
However someone else said this, and I agree, if I have an AI use my claude-code CLI how is not valid first-party app use? It would be different if they would disallow others to use your claude-code account, and I think most including these AI companies would argue AI is supposed to replace and augment humans. So they aren't banning AI's from using the CLI, right- though thats what some of them are seemingly wanting to do.
> Then maybe they should charge for that instead of banning accounts?

They do, though - you're free to buy tokens and use Google's AI/LLM via the API?

What OpenClaw was doing was pretending to be a different product (Anti gravity) in order to use the cheaper tier.

Google wants usage that earns them street cred, not usage from bots who will never evaluate the output. They're all fighting tooth and nail to acquire customers, both free and paid... they didn't want their giveaways to be burned.
They're about to find out that if you aim to wholesale replace your workers with AI you can't really complain if your users replace themselves with AI...
So they ban a group of early adopters who picked their product and who shape opinions.
But banning accounts wholesale is not going to earn them more customers. They could have just disabled Gemini access, or even given a warning first.

I don't use OpenClaw, I do pay hundreds per month for AI subscriptions, and I will not be giving that money to Google while they treat their customers like this.

> They could have just disabled Gemini access

They just disabled Antigravity access.

> But banning accounts wholesale is not going to earn them more customers.

it has the chilling effect - people getting banned by google might imagine their entire google account getting banned (whether that's true or not is irrelevant).

> They could have just disabled Gemini access

Yes, please!

Hah, yeah, I'm seriously considering downgrading my Google Home Premium subscription to avoid the Gemini features on my Nest cameras.
If I say “you can use my car for $250/month if you don’t smoke in it” and then you pay me that money and you drive around until one day you smoke in it, I’m not going to let you smoke in my car. I told you not to smoke in it and you smoked in it. That’s the deal. All seems fine to me tbh.
I think it's a bad analogy. For one - smoking does very high permanent damage to a car interior.

Two - the usage pattern was Shaun's toc but not obviously against the spirit.

More like "you can use my car to drive around as much as you want" And then going: Obviously I didn't mean driving to another coast on a highway

More like "you can use my car to drive around as much as you want so long as you don't drive to another coast on a highway" and then you drive to another coast on a highway and get mad when I won't give you my car next time.
No, it’s more like you have 10k km to drive in this car.

And then, after you actually drive it for 10k km on “unauthorized” highways, they ban you.

That’s why analogies are for fucking idiots. There is a true example — you bought access to use your account for n tokens, and google got pissy that you didn’t use the tokens in their spyware ecosystem.

For you analogy loving fucks, it’s like the arcades selling their proprietary 50 cent worth token coins, that only works on their machines, when they could just accept 50 cent coins.

Well, your premise is false because you do not actually buy access for a stated number of tokens. I used the analogy because you have trouble getting facts straight so indeed you fall into the category that you describe there.
They’re not banning for excessive use. They’re banning for use with unauthorized software. Big difference.
Yep it sounds like Google is charging too little, and taking losses that would be unsustainable for other companies, to try and win the market on AI coding products. Which is a violation of anti trust law, I think. Now that people are using their pricing in an unexpected way where their product isn’t the one winning from their anti competitive practices, they’re punishing the users. Classic monopolistic behavior. And why we need to tax mega corp more and break them up.
I agree. As others have mentioned here, the authenticate with AntiGravity web popup clearly says that this authentication is only to be used with Google products.

How can Claws users miss this?

What Google could have done better: obviously implement rate throttling on API calls authenticated through the Gemini AI Pro $20/month accounts. (I thought they did this, buy apparently not?) Google tries hard to get people to get API keys, which is what I do, and there seems to be a very large free tier on API calls before my credit card gets hit every month.

Given how popular OpenClaw is (and that OpenClaw itself supports antigravity), I think it's shortsighted to not publicly state that it's not allowed and to warn users. Permanently banning people from Antigravity (much like any Google product) feels really harsh.
Can I at least log in one last time and download my gmail messages from 2004?
Then it should be “This is your first and final warning. The next time we catch you, it’s a ban.”. People are building their lives around this stuff and kneejerk bans erode good faith in your platform.
> Then it should be “This is your first and final warning. The next time we catch you, it’s a ban.”. People are building their lives around this stuff and kneejerk bans erode good faith in your platform.

This is actually the soft-touch approach: the users of these vibe-coded products need to understand that they are delegating their authority to the tool to work on their behalf.

In this case, they delegated to a tool that broke the ToS. The result could have been a lot worse, and in return they learned that the tool is acting with their full authority.

-----------------

EDIT:

One of the users got this response from google support:

> Our product engineering team has confirmed that your account was suspended from using our Antigravity service. This suspension affects your access to the Gemini CLI and any other service that uses the Cloud Code Private API.

Their decision? To break ToS on some other provider:

> I guess it is time to move on to Codex or Claude Code.

So, yeah, perhaps the users really are too stupid to understand what's going on, and even this soft-touch approach has done nothing to clue them in.

Except it's expressly NOT against the TOS of codex to use it via oAuth with Openclaw (the jury is currently out re Anthropic)
The difference is ChatGPT Pro/Plus plans have one shared pool of token limits shared across all use cases.

In contrast Google's AI plans give you at least three seperate pools of token usage limits: Gemini App + Antigravity/Other Code Assist tools like Android Studio + AI Studio free usage limits.

Google limit the context of where you can use their tokens but in exchange they give you substantially more.