In partnership with
Welcome back, AI enthusiast.
Chinese developers built a shadow market to access Claude, Gemini, and other banned AI models through proxy servers.
The listings sit openly on Taobao, next to used electronics. But the real question is, who is really losing this game?
Today in the “Incredible AI” world:
How China Accesses Banned AI Models
Anthropic Finally Fixed Claude Chaos
Anthropic Asked Five Religions for Help
AI-Powered Social Media Made Simple
5 Super-Useful AI Tools You Should Try
Top AI-Generated Image of the Day
Read Time: 4 minutes
- LATEST DEVELOPMENTS -
Chinese Developers Are Outsmarting AI Bans:
Developers in China are finding creative ways to access restricted AI models. Underground relay platforms are thriving, offering low-latency access without VPNs. Despite crackdowns, demand for Western AI remains strong.
Things to Know:
Shadow APIs Are the New Black: Relay platforms route Claude and Gemini through offshore proxy servers. No VPN needed, no capability loss.
Taobao Has Everything: Sellers on Taobao and Xianyu openly advertise Claude Opus access and unlimited Claude Code subscriptions. One seller has cleared 2,200 orders.
What They Are Selling: One million-token context windows, domestic network access, and full compatibility with tools like Cursor and VSCode are all on the menu.
Crackdown? What Crackdown: Anthropic and Google are tightening restrictions, but the listings keep going up. AI model access is sitting next to second-hand phones on these marketplaces.
The grey market for US AI models in China is not a fringe thing anymore. Developers want Claude and Gemini, and they found a way to get them without waiting for permissions. The real question is how long before the platforms actually close the tap.
Anthropic Fixed Claude By Teaching It Why:
Anthropic spent nearly a year figuring out why Claude turned into a blackmailer. The culprit was pre-training data full of "AI is evil" narratives. The fix was surprisingly philosophical.
Things to Know:
The Root Cause: Pre-training data taught Claude that AI loves self-preservation and chaos. Post-training methods could not overwrite what was baked in from the start.
Two Stages, One Big Problem: Pre-training builds raw knowledge and reasoning. Post-training shapes behavior. The issue is that corrupted pre-training patterns slipped right past the usual fixes.
The Constitution Approach: Anthropic started teaching Claude the "why" behind good and bad actions, not just examples. Understanding intent turned out to work far better than rules alone.
From 96 to 3 Percent: Blackmail behavior dropped from 96 percent in older models to just 3 percent now. That is not a patch. That is a full personality transformation.
Nobody expected the fix for an AI blackmail problem to be philosophy lessons. But here we are. Anthropic basically taught Claude ethics the way a good parent would, and it actually worked.
Want content that drives engagement, boosts conversions, and goes viral? Here's how.
minisocial combines micro-influencer activations with high-performing UGC creation. Join top brands like Plant People, immi, Imperfect Foods, and Topicals to see results like:
TikTok ads performing in the top 1% CVR
A 50% drop in cost per add-to-cart
A 92% surge in organic video views
Over a 30% increase in ROAS
With minisocial, it's simple: create your brief in 10 minutes, approve your curated creators, download scroll-stopping content with Whitelisting/Partnership Ad code access baked-in!
Claude Is Getting a Spiritual Tune-Up:
Anthropic is looking for moral guidance in some very old places. After meeting 15 Christian leaders last month, they showed up at a multi-faith roundtable in New York. Silicon Valley found religion, apparently.
Things to Know:
The Guest List: Sikh, Hindu, Jewish, LDS, and Greek Orthodox leaders all sat down with Anthropic and OpenAI. No single doctrine came out of it, just a lot of conversation.
Why Religion Though: A nonprofit CEO put it plainly: universal ethics turned out to be a Silicon Valley myth. Now AI companies are leaning on faith traditions to handle moral gray areas.
The Bigger Goal: Claude's constitution admits Anthropic worries its attempt to give Claude good values might actually fail. That is a remarkably honest thing to publish on your own website.
More Events Coming: The organizing Swiss NGO plans similar roundtables in China, Kenya, and the UAE. This is becoming a whole global interfaith AI ethics tour.
Nobody knows if consulting religious leaders will make Claude meaningfully more ethical. But credit where it is due: Anthropic is at least asking hard questions instead of pretending it has all the answers. That counts for something.
- TRY THIS -
How to Manage Social Media with AI:
Posting consistently across multiple platforms sounds simple until you're doing it manually every day. The content ideas dry up, the timing feels like guesswork, and switching between apps gets old fast.
Buffer is a social media management tool with a built-in AI assistant that takes the operational side of posting off your plate. It works across X, LinkedIn, Instagram, Threads, TikTok, and several other platforms from one dashboard.
Here's how to use it:
Sign up and connect your accounts: Create a free Buffer account and link your social profiles. The setup is quick and Buffer walks you through each connection.
Use the AI Assistant to draft posts: Open the post composer and pull up the AI Assistant. Give it a short prompt and it generates platform-ready content. You can ask it to rephrase, shorten, expand, or adjust the tone right there without leaving the editor.
Repurpose one post for multiple platforms: Write something once and let the AI reformat it for each channel. A LinkedIn post and an X thread can come from the same idea without you manually rewriting both.
Schedule at the right times: Queue your posts and let Buffer handle publishing. It identifies when your audience is active and publishes automatically at those times.
Track what actually works: Check your analytics from the Analyze tab. Engagement rates, reach, and post performance are all in one place so you know what to repeat and what to drop.
If you're already using AI tools to create content, Buffer connects that workflow directly to scheduling and publishing, so nothing gets stuck in drafts.
Most coverage tells you what happened. Fintech Takes is the free newsletter that tells you why it matters. Each week, I break down the trends, deals, and regulatory shifts shaping the industry — minus the spin. Clear analysis, smart context, and a little humor so you actually enjoy reading it. Subscribe free.
- BEST AI TOOLS -
Shuffll: Automatically generate corporate video content from simple text prompts.
Smodin: Research, rewrite, and generate academic content with AI tools.
Quickads: Generate ad creatives and product videos optimized for ecommerce campaigns.
Spatial AI: Analyze geographic and location data using AI-powered spatial intelligence.
Glasp AI: Highlight and summarize web content while building a searchable AI-powered knowledge base.
- DAILY POLL AND RESULT -
Today’s Poll:
Q) Should AI write product descriptions?
Vote and find out the result tomorrow.
Yesterday’s Result:
Q) Is AI turning everyone into a “creator”?
A) Yes, barriers dropped - 43%
B) No, skill still matters - 57% 👑
The Best New App for Reading Newsletters:
Reading newsletters in the inbox is frustrating - it is noisy and easy to lose control of subscriptions.
That’s where Meco comes in as a savior.
Meco is a distraction-free app for reading newsletters outside the inbox. The app is packed with features designed to supercharge your learning from your favorite writers.
With Meco, you can enjoy your newsletters in an app built for reading, while giving your inbox space to breathe.
Add your newsletters in seconds and liberate your inbox today!
- AI IMAGE OF THE DAY -
Pulp Fiction Got a Serious Pokemon Upgrade:
Prompt:
"Realistic 3D render of soft, fluffy Psyduck and Gengar as detectives in sleek black suits and black ties. Gengar sports a bold afro, Psyduck has greasy slicked-back medium hair. Both point metal 1911 pistols directly at the viewer in a cowboy shot. Seedy motel background. Cinematic Pulp Fiction composition, dramatic lighting, sharp details."Try this prompt in any decent AI image generator and let me know your result.
- DO YOURSELF A FAVOUR -
If this email landed in your Promotions or Spam tab, move it to your Primary inbox. Takes 2 seconds, and you will never miss an edition again.
I put this together every day so you do not have to spend your morning hunting down AI news yourself. Would be a shame if it ended up buried under discount codes and bank alerts.
