🤯 Gemini cracks 500 year old mystery

Good Morning! Google’s Gemini 3.0 Pro just helped crack a mystery that historians couldn’t solve for over 500 years. Plus, you’ll learn how to turn rough ideas into perfect AI prompts with just one click.

In today’s AI newsletter:

• Gemini Decodes a 500-Year-Old Manuscript Mystery

• Yann LeCun Publicly Slams Meta’s AI Leadership

• Plaud launches a new AI pin + desktop meeting notetaker

• How to Turn Rough Ideas Into Perfect AI Prompts Automatically

• 4 new AI tools worth trying

AI MODELS

📜 Gemini Decodes a 500-Year-Old Manuscript Mystery

Google’s Gemini 3.0 Pro was used to interpret handwritten annotations in a rare 1493 copy of the Nuremberg Chronicle, revealing their meaning for the first time.

• Gemini analyzed high-resolution images combining vision, language, and historical reasoning

• It identified the notes as biblical chronology calculations, reconciling Septuagint vs Hebrew timelines

• The model parsed abbreviated Latin, Roman numerals, and theological context without human guidance

• Findings align with known medieval methods, despite minor numerical errors

This goes beyond OCR or transcription. Gemini showed it can reason across images, text, and historical context, unlocking archives that have resisted human consensus for centuries. Multimodal AI is quietly becoming a powerful tool for historians, not just engineers.

AI NEWS

🚨 Yann LeCun Publicly Slams Meta’s AI Leadership

Meta’s former chief AI scientist Yann LeCun has openly criticized Alexandr Wang and predicted further talent exits, signaling deeper fractures inside Meta’s AI org.

• LeCun called Wang “inexperienced” in research, saying he doesn’t understand what attracts or repels top AI researchers

• He claims Meta’s GenAI org was sidelined after Llama setbacks, eroding internal trust

• LeCun warns Meta is now “completely LLM-pilled,” calling LLMs a dead end for superintelligence

• He predicts more AI researchers will leave, following his own exit to start a new company

Meta is doubling down on LLM scale, while one of AI’s most influential thinkers says that path won’t reach superintelligence. If top researchers agree, Meta’s biggest risk isn’t competition, it’s internal brain drain.

AI HARDWARE

🎙️ Plaud launches a new AI pin + desktop meeting notetaker

Plaud unveiled the NotePin S, a wearable AI notetaker, alongside a new desktop meeting notetaker designed to compete with tools like Granola and Fireflies.

• NotePin S ($179) adds a physical button for start/stop + highlights, with clip, lanyard, magnetic pin, and wristband included

• 20 hours battery, 64GB storage, dual MEMS mics (≈10 ft range), and Apple Find My support

• Desktop app auto-detects live meetings, records system audio, and turns calls into structured AI notes

• Multimodal notes let users combine audio, text, and images in one place

Plaud is building a full-stack notetaking workflow across physical and digital meetings. As work shifts between IRL conversations and Zoom calls, always-on, AI-native memory tools are quietly becoming essential infrastructure.

HOW TO AI

🗂️ How to Turn Rough Ideas Into Perfect AI Prompts Automatically

In this tutorial, you’ll learn how to use a lightweight Chrome extension that upgrades your messy, half-baked ideas into clear, structured, high-performing AI prompts in a single click. It works directly inside tools like ChatGPT and Claude..

🧰 Who is This For

• Creators who want better AI outputs

• Developers who want production-ready prompts with minimal effort

• Beginners who struggle with prompt wording and want instant improvements

• Anyone who wants consistently better results from ChatGPT or Claude

STEP 1: Install the MetaPrompt Chrome Extension

To get started, you first need to add the extension to your browser. Open the Chrome Web Store and search for MetaPrompt. Once you find it, click the blue Add to Chrome button. If Chrome asks for confirmation, approve the installation.

The extension installs instantly and doesn’t require any setup or account creation. It’s designed to live quietly inside your existing AI chat tools, so you don’t need to open a new app or website.

STEP 2: Find the New “Optimize Prompt” Button

After installation, open ChatGPT or refresh the tab if it was already open. Look closely at the message input area where you normally type your prompts. You’ll now see a small button labeled ✨ Optimize Prompt hovering just above the text box.

This button is the core of the tool. If you can see it, the extension is active and ready to use. From this point on, you can optimize any prompt without leaving your chat window.

STEP 3: Write a Rough Prompt

Instead of overthinking your prompt, just type your idea in the most basic way possible. For example, you might write something like, “Create a photo book style magazine cover.” Don’t worry about structure, detail, or wording.

The key step here is to not press Send. Before submitting anything to the AI, click the ✨ Optimize Prompt button instead. You’ll briefly see a small animation or confirmation indicating the tool is rewriting your input.

STEP 4: Review and Send the Optimized Prompt

Within seconds, your original sentence will be replaced by a much more detailed and structured prompt. You’ll notice that it adds useful elements automatically, such as role definitions, context, constraints, formatting instructions, and visual or technical details you probably wouldn’t think to include.

Take a moment to read through the optimized version. You can tweak it if you want, but most of the time it’s already good enough to use immediately. When you’re satisfied, simply hit the Send button and let the AI generate a much higher-quality result than before.

Starlink says it is providing free broadband service to new and existing customers in Venezuela.

Tech workers in Silicon Valley are increasingly experimenting with gray-market peptide drugs from China, legal for personal use, for improving focus.

A profile of June Paik, CEO of Seoul-based chip startup FuriosaAI, valued at ~$700M, whose AI chip dubbed “RNGD” is slated to enter mass production this month.

Startups have started to offer free nicotine pouches as a perk, as some employees claim the products help them focus during the workday, despite health hazards.

⚙️ IQuest-Coder-V1: China's new 40B Model crack 80% on SWE-Bench

🎨GPT Image 1.5: OpenAI’s latest, faster and more controllable image model

⚙️ GLM-4.7V: Zhipu AI’s open-source multimodal model

🧑‍💻 MiniMax 2.1: Strong model for coding and building apps

THAT’S IT FOR TODAY

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas