Good morning, AI enthusiasts. French startup Mistral just turned AI document processing on its head — with a brand new model that makes complex data extraction so simple as an API call.
With speeds of as much as 2000 pages per minute and the flexibility to handle multilingual texts, images, charts, and more, is that this the tech that finally converts static archives into the AI-powered gold mines of tomorrow?
P.S. Our next workshop is today at 3:30 PM EST! Join to learn the best way to use the most recent AI tools to take your vibe coding to the subsequent level. RSVP here.
In today’s AI rundown:
-
Mistral OCR’s AI-ready document processing
-
China’s ‘fully autonomous’ Manus AI agent
-
Design landing page mockups with AI
-
AI avatars getting emotional intelligence
-
4 latest AI tools & 4 job opportunities
LATEST DEVELOPMENTS
MISTRAL
📜 Mistral OCR’s AI-ready document processing

Image source: Mistral
The Rundown: Mistral AI just launched Mistral OCR, a robust latest API designed to extract and comprehend detailed information from complex documents with exceptional speed and accuracy.
The small print:
-
The API can accurately analyze docs with images, equations, tables, and advanced formatting, converting them to markdown outputs for AI processing.
-
OCR can process as much as 2000 pages per minute and supports multilingual evaluation across hundreds of languages, including Hindi and Arabic.
-
Benchmark tests place Mistral OCR well ahead of rivals like Google’s Document AI, Azure OCR, and GPT-4o across different document evaluation categories.
-
Users may also deploy the OCR technology on-premises, which is right for organizations handling classified or sensitive datasets.
Why it matters: With a lot of the world’s data still trapped in complex documents, unlocking it efficiently is crucial. Mistral OCR capabilities could supercharge archive-heavy industries like financial analytics, legal discovery, historical preservation, and more — transforming static information into dynamic, AI-ready knowledge bases.
TOGETHER WITH TURING
The Rundown: When LLMs fail, it isn’t in regards to the model’s raw power. Turing’s “Maximizing Your LLM ROI” whitepaper shows how the true culprits — misaligned training, poor evaluation, and optimization gaps — can quietly sabotage ROI.
Inside, you’ll learn the best way to:
-
Reduce hidden inefficiencies that drive up LLM costs
-
Strengthen evaluation techniques for higher accuracy and reliability
-
Construct an AI strategy that maximizes performance and business value
MANUS AI
🤖 China’s ‘fully autonomous’ Manus AI agent

Image source: Manus
The Rundown: A Chinese startup just introduced Manus, calling it the world’s first fully autonomous AI agent — able to handling real-world tasks independently and achieving latest SOTA performance on agentic benchmarks.
The small print:
-
Within the demo, Manus will be seen handling tasks like resume screening and property research, accessing its own independent computer instance.
-
The agent also shows skills like web browsing, coding, and creating visuals while reportedly with the ability to handle tasks on sites like Upwork and Fiverr.
-
It outperformed leading general-purpose assistants like ChatGPT and Gemini on the GAIA benchmark, a comprehensive evaluation of AI performance.
-
Manus currently operates on an invite-only basis — with the team committing to open-source the models behind the agent later this yr.
Why it matters: We’re at the purpose of acceleration where relatively unknown labs are dropping (reportedly) state-of-the-art level agentic tools. While the early iterations of agents handled more easy tasks that needed human handholding, we’re quickly approaching the subsequent step of more autonomous complex workflows.
AI TRAINING
🖥️ Design landing page mockups with AI

The Rundown: On this tutorial, you’ll learn the best way to use Ideogram’s latest 2a image generation model to create professional-looking landing page mockups for what you are promoting using just text — no design skills required.
Step-by-step:
-
Enroll for a free Ideogram account and navigate to the creation interface.
-
Write an in depth prompt describing your landing page layout, including text elements, color scheme, and visual style.
-
Customize settings: select 16:9 aspect ratio for desktop views and select “Design” style for skilled results.
-
Click “Generate” to create multiple variations after which download or further refine your favorite design.
Pro tip: You’ll be able to take your mockup to the subsequent level by uploading it to AI coding assistants like Windsurf or Cursor and asking them to “code this landing page,” immediately turning your design into functional code.
PRESENTED BY INNOVATING WITH AI
💼 Start your profession as an AI Consultant

The Rundown: Innovating with AI’s latest program, AI Consultancy Project, equips AI enthusiasts with all of the resources they should capitalize on the booming AI consulting market – which is about to grow 8x to $54.7B by 2032.
This system offers:
-
Tools and framework to search out clients and deliver top-notch services
-
A 6-month roadmap to construct a 6-figure AI consulting business
-
Student landing their first AI client in as little as 3 days
Click here to request early access to The AI Consultancy Project.
TAVUS
🎭 AI avatars getting emotional intelligence

Image source: Tavus
The Rundown: Digital twin developer Tavus just unveiled a significant upgrade to its Conversational Video Interface (CVI) platform, launching three latest AI models that work together to make video interactions with AI feel more humanlike and personalized.
The small print:
-
Phoenix-3 handles full-face animation, creating natural facial expressions for avatars, including eye movements, eyebrows, and subtle micro-expressions.
-
Raven-0 acts because the AI avatar’s eyes, analyzing cues like body language and facial expressions in real time to reply more naturally to human emotions.
-
Sparrow-0 handles conversation timing, eliminating awkward pauses and interruptions by understanding when to talk and when to listen.
-
The corporate showcased the tech through “Charlie,” a demo AI avatar that may hold conversations while searching the net, analyzing screens, and more.
Why it matters: While many scoffed at Sam Altman’s proof-of-personhood startup, tech like that is showing how hard it’s about to be to discover AI from humans online. The times of AI customer support reps and digital avatars feeling robotic and scripted of their interactions are coming to an end very soon.
QUICK HITS
🛠️ Trending AI Tools
-
🔎 Google Search AI Mode – Get better-reasoned answers to tough questions
-
🧠 QwQ-32B – Qwen’s low cost, efficient, and open-source reasoning model
-
⚙️ Windsurf Wave 4 – Agentic coding with Previews, tab-to-import, and more
-
🎬 Ray2 – Powerful video AI with features like KeyFrames, Extend, and Loop
💼 AI Job Opportunities
-
🧮 Faculty – Financial Accountant/Finance Manager
-
💰 OpenAI – Compensation Partner, Go To Market & Sales
-
🗂️ Harvey – Executive Assistant
-
🔬 Hippocratic AI – Senior Applied Scientist
📰 The whole lot else in AI today
Google co-founder Larry Page is starting a brand new AI company called Dynatomics, which is able to leverage LLMs to create factory-ready designs for quite a lot of products.
Tencent open-sourced HunyuanVideo-l2V, a brand new high-quality image-to-video model with custom computer graphics, audio, and lip-syncing capabilities.
Anthropic submitted latest AI Motion Plan recommendations to the White House, calling for enhanced national security testing, stricter export, and infra expansion.
OpenAI released an update bringing IDE integration to ChatGPT for macOS, allowing Plus, Pro, and Team users to edit code directly inside development environments.
Privacy browser DuckDuckGo rolled out latest AI features, including expanded anonymized access to leading chatbots and AI-assisted search answers.
Former OpenAI policy head Miles Brundage criticized the corporate’s latest safety document, saying it causes a “dangerous mentality for advanced AI systems.”
Convergence AI unveiled Template Hub, a community-driven marketplace allowing users to create, share, and deploy task-specific AI agents in a single click.
COMMUNITY
🎥 Join our next live workshop

Join our next workshop this Friday at 3:30 PM EST with Dr. Alvaro Cintas, The Rundown’s AI professor. You may learn the best way to use the most recent AI tools to take your coding to the subsequent level, and the way we’re leveraging AI-powered coding at The Rundown.
RSVP here. Not a member? Join The Rundown University on a 14-day free trial.
🤝 Share The Rundown, get rewards
We’ll at all times keep this article 100% free. To support our work, consider sharing The Rundown with your pals, and we’ll send you more free goodies.

That is it for today!Before you go we’d like to know what you considered today’s newsletter to assist us improve The Rundown experience for you.
|
|
Login or Subscribe to take part in polls. |
See you soon,
Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team