| Sponsored by |  |
|
|
| Welcome Automaters, 👋 | If you thought Anthropic’s new Claude Mythos model was locked away behind an impenetrable velvet rope, think again. According to a Bloomberg report, a group of people with absolutely zero clearance has reportedly gotten their hands on Anthropic’s brand-new, invite-only cybersecurity tool. | And honestly? The way they got in is almost impressive. | The Backstory: Mythos is a monstrously powerful AI built for enterprise security, but Anthropic has been very vocal about the fact that this thing is a double-edged sword. In the wrong hands, it could flip from a digital shield into a terrifying weapon. That is exactly why they locked it behind a super-exclusive rollout called Project Glasswing, shared only with a handful of elite, pre-approved partners. | So, how did they get in? | They didn't break down the front door; they picked the lock. The group reportedly made a clever, educated guess about the model's online location by analyzing patterns from Anthropic's other models. From there? They leveraged access belonging to someone employed at a third-party contractor working for Anthropic. | The group, whose members are still anonymous, are part of a Discord channel dedicated to sniffing out unreleased AI models. They’ve reportedly been using Mythos regularly since the very day it was announced, even showing Bloomberg screenshots and a live demo as proof. | Should we be panicking? | Before you start spiraling: a source described the group as people "interested in playing around with new models, not wreaking havoc." Anthropic also told TechCrunch that it has found no evidence its own internal systems were compromised. | Still, an investigation is underway. Anthropic is definitely going to be auditing every single connection and contractor credential after this. | But hey, it is a harsh reminder that even the most advanced, "unhackable" AI in the world is only as secure as the weakest link in the chain—which, in this case, was just a very smart person with way too much curiosity. | So yeah, is it just me, or does this feel like the beginning of an AI heist movie? or is this just the reality of building software in 2026? | I’d love to know your take. You can also learn more here | Here's what we have for you today | | | | | | | 🥳 OpenAI Launches ChatGPT Images 2.0 With Advance Text Rendering and Reasoning That Actually Works | | Remember those dark, primitive days when AI image tools were spectacularly bad at spelling? We're talking restaurant menus full of gibberish, street signs with scrambled letters, and enough chaotic typos to make your high school English teacher weep. | Well, good news! Those dark ages are officially over. | OpenAI dropped ChatGPT Images 2.0 on Tuesday, April 21, and the headline feature is genuinely wild: the model can finally render clean, legible, and accurate text inside its images. | Ask it for a professional menu, a polished poster, or a product ad, and you’ll get something you could actually use without a second glance. |  | ChatGPT Images 2.0 |
| But readable text is just the beginning. Images 2.0 comes packed with what OpenAI calls "thinking capabilities." When you select the Thinking models, this thing doesn't just blurt out an image. It browses the web, plans out the layout, generates multiple distinct images from a single prompt, and—wait for it—proofreads its own work before handing you the final result. | But that’s not all: OpenAI is promising "unprecedented specificity and fidelity." We’re talking: | Small text and UI elements that you can actually read. Dense layouts that aren't a jumbled mess. 2K resolution that looks sharp on any screen.
| The model is also making massive strides in multilingual support. It is finally rendering non-Latin scripts, including Japanese, Korean, Chinese, Hindi, and Bengali, with actual accuracy. | Who gets it: All ChatGPT and Codex users have access starting now. The "Pro" Perks: Paid subscribers unlock the more advanced, "thinking" outputs. For the Builders: Developers can tap into the gpt-image-2 API, with pricing that scales based on resolution and quality.
| One last piece of gossip? OpenAI is staying tight-lipped about exactly what model architecture is powering Images 2.0 under the hood. But hey, a little mystery never hurts anybody, right? | So are you ready to retire your design software, or do you still have a soft spot for doing things the old-fashioned way? I’m dying to know if this replaces your current creative workflow! | Learn more here. |
| |
| | |
| | The 8-week course that teaches you how to actually evaluate deals. | | Learn the same investing frameworks used at BlackRock, KKR, Ares, and other top firms. | Over 8 weeks: | Underwrite live deal scenarios end to end Build and analyze institutional-level real estate models Make go/no-go decisions based on risk-adjusted returns
| Join the June 8 cohort of the Wharton Online Real Estate Investing program. Use code SAVE300 to get $300 off tuition. | Explore the Program |
|
| | | | | | | | | | 🤖 AI Workout Of The Day: How to Master Interviews with AI | | Ever wish you could step into the hot seat and face a real interviewer without the stakes of an actual interview? | This prompt turns AI into your personal interview simulator, giving you realistic questions, instant feedback, and actionable tips to refine every answer. | Think of it as a virtual rehearsal room where you can practice, experiment, and perfect your delivery until you’re confident, sharp, and interview-ready. | Here’s How: | Set the Scene: Begin by providing the context for the interview—name the broadcast, show, or company, and describe the role or topic. Example: “I’m being interviewed on a tech podcast about AI in healthcare.” Answer Naturally: Respond to each question as if it’s real. Speak your thoughts aloud or type your answers verbatim. Treat this as a safe space to experiment with phrasing, tone, and storytelling. Iterate & Improve: Revise your answers based on the feedback and try again. Ask for advanced tips: Request additional coaching for body language cues, phrasing improvements, or storytelling techniques for maximum impact in real-life interviews. Build Confidence Over Time: Use this prompt repeatedly to track your progress, sharpen your responses, and gain the confidence you need to shine in the actual interview.
| 💡 Prompts to try: | Simulate an interview with [broadcast name or type]. You will act as the interviewer, asking relevant and realistic questions. I will respond as the interviewee. After each of my responses, provide detailed feedback covering: clarity, tone, confidence, conciseness, and impact. Suggest actionable improvements and tips for making my answers more effective. Keep the simulation engaging, realistic, and structured like an actual interview. Ask follow-up questions as needed to challenge me and refine my responses.
|
| |
| | |
| | Is this your AI Workout of the Week (WoW)? Cast your vote! | | That's all we've got for you today. | Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇 | What'd you think of today's email? | | Your feedback means a lot to us and helps improve the quality of our newsletter. |
|
