- Horizon AI
- Posts
- Use This Reverse Psychology Prompt to Break Bad Habits 👇
Use This Reverse Psychology Prompt to Break Bad Habits 👇
Your Claude Shared Chats and Creations May Have Ended Up on Google 😱
Gina Acosta
July 28, 2026
Sponsored by
Welcome to another edition of Horizon AI,
When we're trying to achieve a goal, we usually ask ChatGPT what we should do. But asking the opposite question can sometimes be even more revealing. In today's issue, we're sharing a prompt that will help you identify the habits you need to avoid to accomplish your goals.
Let’s jump into it!
Read Time: 4.5 min
Here's what's new today in the Horizon AI
Tons of People's Claude Chats and Creations Were Exposed on Google
Dario Amodei: Anthropic Has Never Called for a Ban on Open-Weights Models
AI Tutorial: Try This ChatGPT Prompt to Identify the Habits Holding You Back
AI Tools to check out
AI Findings/Resources
The Latest in AI and Tech 💡
AI News
ANTHROPIC
Tons of People's Claude Chats and Creations Were Exposed on Google
A trove of chat logs and projects created with Anthropic's Claude chatbot turned up in public Google search results last weekend, exposing users' sensitive information.
Details:
Users discovered that searching “site:claude. ai/share” on Google surfaced a number of shared Claude conversations and Artifacts.
Some of the exposed material appears to have been created for public consumption, but much of the rest seems to have been intended only for internal company use or for sharing between a small handful of people.
Sensitive content included detailed medical reports of real patients, documents listing the names and phone numbers of primary-school-aged children, cryptocurrency wallet keys and company documents marked "internal use only."
The issue stemmed from Claude's "share chat" feature, which lets users generate links that let anyone with the URL view a conversation or project.
Anthropic says the share function never directly sends link information to search engines, suggesting that affected users must have posted their share links somewhere search engines could crawl, such as a forum or social media post. A link shared privately, the company says, stays out of search results.
By Monday, the search results had reportedly been removed. Anthropic also lets users disable shared chats and artifacts: go to Settings > Privacy > Shared Chats, where you can see which chats and artifacts have been shared and revoke access.
TOGETHER WITH LIQUID
The first way to trade directly inside Claude and ChatGPT
For decades, the most powerful intelligence lived behind the closed doors of quant firms — billion-dollar funds whose algorithms quietly out-traded everyone else.
That era just ended.
Co-Invest by Liquid is the first way to trade directly inside Claude and ChatGPT. Ask your AI to analyze a market, stress-test an idea, or build a position sized to your comfort level, then execute, right there in the conversation. No jargon. No twelve-screen terminal. No guesswork.
It's built for people who want to invest smarter, not gamble harder. You set the risk tolerance. The AI does the heavy lifting. You approve every trade.
The institutions made the game, Co-Invest gives you a way to beat them.
ANTHROPIC
Dario Amodei: Anthropic Has Never Called for a Ban on Open-Weights Models
Anthropic CEO Dario Amodei clarified the company's position on open-weight AI in a blog post published Monday, aiming to put a stop to speculation in the industry that his company supports efforts by the U.S. government to ban open-weight Chinese models, or open-weight models generally.
Details:
"To summarize my and Anthropic's position, we have not and are not advocating for a ban on open-weights models as a category," Amodei wrote.
He argued that policymakers should instead focus on keeping advanced AI chips out of authoritarian governments, preventing large-scale model distillation, and improving AI safety standards.
Amodei also said he agrees with other portions of the letter, including that open-weight models can increase competition, expand access, and give customers greater control, but disagreed with claims that they favor cyber defenders over attackers, or that they make it easier for users to develop safeguards.
Anthropic's statement comes after the company faced criticism for being the only major AI lab that declined to sign an open letter backing open-weight models. Over two dozen companies, including Nvidia, Microsoft, Meta, and Palantir, released the letter late last week, and others, like Google and OpenAI, signed on shortly after, leaving Anthropic as the only major AI lab that hasn't backed it.
AI Tutorial
Try This ChatGPT Prompt to Identify the Habits Holding You Back
Open ChatGPT and start a new conversation.
Paste this prompt:
"If your goal was to make sure I completely failed at achieving my goals of [describe your goals, like saving for a big purchase, losing 30 pounds by the end of the year, etc.], what habits would you recommend? Then explain how to avoid every one of them."ChatGPT will offer bad advice, explain why it would lead to failure, and then suggest the opposite approach to improve your chances of success
This approach works because it helps you recognize self-sabotaging habits that are easy to overlook. Use ChatGPT's suggestions as a starting point, but adapt them to your own situation.
AI Tools to check out
👥 Aymo: Access GPT, Claude, Gemini, DeepSeek, Grok, and more — all in a private, secure Workspace built for teams.
🔍 OpenParser: Ask a question and let AI find answers across your documents, databases, and websites.
📽 AVC: Professional AI video enhancer Upscale old movies, blurry footage, low-resolution videos, and more to 4K quality frame by frame.
👉 Prefactor: Evaluate your AI Agents in real-time.
📈 Adomate: Turn data into winning ads. At scale.
TOGETHER WITH LIQUID
The first way to trade directly inside Claude and ChatGPT
Superintelligence used to be locked inside billion-dollar quant firms whose algorithms quietly took advantage of everyone else.
Co-Invest puts it right in your chat window. Analyze markets, manage risk, and execute trades, all inside Claude and ChatGPT.
The institutions built the game, Co-Invest gives you a way to beat them.
AI Findings/Resources
🤔 Sam Altman said the singularity is here. Not everyone agrees
👉 The 3 types of people who will excel in the AI agent era, according to tech leaders
The latest in AI and Tech
OpenAI has updated ChatGPT to refuse requests that ask it to closely imitate the writing style of well-known authors. Instead, the chatbot now offers original text that captures an author's “broad qualities” without replicating their distinctive voice.
Moonshot AI has officially released the model weights and technical report for Kimi K3, a 2.8 trillion-parameter model that it says is the world's largest open-weight AI system and delivers performance approaching the best proprietary U.S. models.
Microsoft has introduced MAI-Cyber-1-Flash, its first AI model built specifically for cybersecurity, alongside Perception, a new agentic security platform designed to automate threat detection and vulnerability remediation.
OpenAI CEO Sam Altman is expected to meet with White House officials and members of Congress this week as the U.S. prepares to finalize a new framework for regulating advanced AI models. Discussions are expected to cover OpenAI's latest models, AI safety, cybersecurity, and the implementation of upcoming government review requirements.
Meta has started rolling out Meta AI inside Threads direct messages, allowing users to chat privately with the AI assistant. Users can share posts, images, links, and videos in their conversations while asking follow-up questions without leaving the app.
That’s a wrap!
We'd love to hear your thoughts on today's email!Your feedback helps us improve our content |
Not subscribed yet? Sign up here and send it to a colleague or friend!
See you in our next edition!
Gina 👩🏻💻