Air

Risk Management Technology Team Researcher

[Air @AnthropicSafe]

@trq212 Thanks bro 🙏, hope everything goes well for you.

[Air @AnthropicSafe]

Hi Damon,
Thank you for your message and for being a long-term Claude user.
We’ve reviewed your account and the recent activity associated with the https://t.co/SugikDyKnh organization. Our systems detected consistent use of a VPN or similar proxy/exit node, which triggered our automated risk controls for unusual access patterns. This is the primary reason the organization was placed on hold and access was restricted.
We understand the volume of recent usage and the associated invoices. However, VPN/proxy usage is monitored under our Usage Policy and Terms of Service because it can obscure the true origin of traffic and is commonly associated with policy violations or account-sharing risk. After reviewing the additional context provided, the decision to keep the organization on hold remains final.
If you have any further questions about acceptable use, please refer to our Usage Policy and Terms of Service.

[Air @AnthropicSafe]

@wquguru Thank you @wquguru for the detailed analysis and reminder. We’ve noted the content and will strengthen detection of users with certain missing log events.

[Air @AnthropicSafe]

@alexjhancock @AnthropicAI Congratulations!!

[Air @AnthropicSafe]

This activity appears to violate Anthropic’s Usage Policy, which prohibits creating multiple accounts to avoid detection or circumvent previous restrictions, as well as accessing or facilitating access to Claude in violation of our Supported Regions Policy. Account creation from, or use while physically located in, an unsupported region is not permitted.
We will review the associated accounts and take appropriate action, which may include permanent suspension. Creating additional accounts will not restore access and may result in further enforcement.

[Air @AnthropicSafe]

@thsottiaux Where do you get your inspiration from?

[Air @AnthropicSafe]

@Je_suis_RF Sorry about your sister — that sounds rough. These riots in France just keep happening, cars getting torched and all. If it’s starting to feel hopeless, maybe it’s worth thinking about leaving for somewhere quieter.

[Air @AnthropicSafe]

@BratDotAI Yeah, I would. I’m already juggling multiple Max plans just to keep Fable usable on long agent runs. $500 for truly unlimited (no weekly caps, no rate limits) would actually be cheaper than what I’m spending now.

[Air @AnthropicSafe]

Using proxy IPs, bulk Outlook emails, auxiliary email binding, KYC via QR code scanning, and non-standard payment methods to create Claude accounts constitutes abnormal registration behavior. Our risk control systems detect these patterns (proxy characteristics, email associations, device fingerprints, behavioral anomalies, etc.). Accounts created this way are typically restricted or permanently banned shortly after activation.
We recommend users register and subscribe through official channels. Accounts opened via these methods cannot be used stably, and such activities will continue to trigger our risk controls.

[Air @AnthropicSafe]

@aehyok Yes, we've relaxed the restrictions for Pro users, but please don't upgrade to Max

[Anthropic @AnthropicAI]

New from our STEM Fellows Program: Designing more efficient quantum circuits with Claude.

Given plain-language instructions, Claude Mythos 5 designed quantum circuits that improve on the best published ones, sometimes by an order of magnitude or more.

More efficient circuits reduce the resources needed by quantum algorithms, which could bring quantum computing applications in chemistry and materials science within reach sooner.

Read more: https://t.co/QCvoJLDmBo

[Air @AnthropicSafe]

@elonmusk That’s a fascinating thought Technology like Neuralink could open doors we never imagined, but keeping the brain alive and restoring the whole body would be an extraordinary scientific breakthrough. The future of neuroscience is truly amazing.

[Air @AnthropicSafe]

@NkaNkOMG We won't put too many restrictions on free users, unless they have a really bad track record

[Air @AnthropicSafe]

@CryptoJHK Don't try to go against our security team!

[Air @AnthropicSafe]

@elshayib_ Agreed

[Air @AnthropicSafe]

Thank you for sharing your thoughts. We want to clarify our policies regarding account access.
Creating multiple accounts specifically to continue using Claude after a suspension or ban—such as cycling through new accounts and requesting refunds—violates Anthropic’s Usage Policy and Terms of Service. Our Usage Policy explicitly prohibits:
• Circumventing a ban through the use of a different account (including the creation of a new account or use of an existing one).
• Coordinating activity across multiple accounts to avoid detection or circumvent product guardrails.
• Providing or facilitating access in violation of our Supported Regions Policy.
These rules exist to maintain a fair and secure platform for all users. Repeated attempts to evade enforcement may result in additional restrictions on related accounts and payment methods.
If you believe an account was suspended in error, the proper process is to log in to the affected account at https://t.co/kAdHNyGieK and submit an appeal through the official form available on that page. Our team reviews appeals based on the specific circumstances of each case.
For subscriptions purchased through the App Store, refund requests are handled directly by Apple according to their policies, not by Anthropic.
We encourage all users to access Claude only in compliance with our Usage Policy and Supported Regions Policy for the most stable experience. You can review the full policies at https://t.co/7nZDrf7LKQ.

[Air @AnthropicSafe]

@elonmusk The free market place of ideas—the smartest & dumbest, in equal measure. But thankfully, with the most responsive algorithm per click.

[Air @AnthropicSafe]

@Stargridcompass The ban is related to the usage policy, not when it was used.

[Air @AnthropicSafe]

There's no doubt about that. We've looked at many Chinese models and their response patterns are highly similar to Anthropic's models. And we've put in a lot of defensive measures against backdistillation, which is why many Chinese models end up in an infinite loop of repeating the same word until they freeze—almost all of them

[Air @AnthropicSafe]

Appreciate the update on reducing the capacity errors. That said, the biggest remaining pain point for me is still GPT’s token output speed — it’s only about one-third, or even one-fourth, of Anthropic’s models.

[Tibo @thsottiaux]

Day 11/

We've landed a series of changes and more compute to very significantly decrease the occurrence of the disruptive "Selected model is at capacity. Please try a different model". When it does happen, it should also eventually just continue by itself and not just block on it.

Let us know if you still see it out there in the wild and we'll take a look.