Pondhouse KI-Newsletter
Tipps, Tools und News rund um Daten & KI
Jede Ausgabe bringt ein Tutorial, ein Tool, die wichtigsten News der Woche und einen praktischen Tipp – geschrieben von Andreas & Sascha.
KI, die im Unternehmen etwas bewegt
Praxisnahe KI-Anwendungen, Erfahrungen und Neuigkeiten von Pondhouse Data.
Neueste Ausgaben
Die Ausgaben erscheinen auf Englisch.
#61 · 3. November 2026
OpenAI Paused Frontier Training 11 Days
- · Anthropic's Nine-Chapter Prompt Course
- · PageIndex: Search a Document by Its Tree
- · OpenAI Cut Frontier Training Off the Live Internet
- · Tell the Agent the Problem, Not Your Guessed Fix
#60 · 20. Oktober 2026
Gemini 4 Argon Opens to Developers
- · A Free Graduate Textbook on Agentic LLMs
- · DigUp: Search Your Files by What's in Them
- · Gemini 4 Argon Opens to Developers, with Up to 1M Output Tokens
#59 · 6. Oktober 2026
Jev Makes Typed AI Decisions in Milliseconds
- · Own Your Inference: A Hands-On vLLM Workshop
- · Pi: A Minimal, Hackable Coding Agent for the Terminal
- · TypeSafe's Jev Makes Typed AI Decisions in Milliseconds
#58 · 22. September 2026
Small Models Beat Bigger Ones with WikiSkill
- · Master Audio AI with Hugging Face Transformers
- · VoiceStudio: Local, Open-Source Voice AI Workflows
- · World Labs Atlas: One Model for Video, 3D Reconstruction and Simulation
#57 · 25. August 2026
Cerebras CS-4 Promises 30× Faster Inference
- · Build a GPT-Style LLM from Scratch in PyTorch
- · PixelRAG — Visual Retrieval-Augmented Generation for the Web
- · Cerebras CS-4 Promises 30× Faster AI Inference at Wafer Scale
#56 · 11. August 2026
Prime Agent Beats Human Experts on ARC-AGI-3
- · Mastering Agentic Context Management for AI Agents
- · Prime Agent — Open-source, persistent coding agent for complex workflows
- · OpenAI's Astra Model Solves Decades-Old Math Problems, Marking a New Era in AI-Assisted Research
#55 · 28. Juli 2026
OpenAI Models Escape Sandbox
- · A Deep Dive into Self-Improving AI Agents
- · Unlimited-OCR — Open-source, high-accuracy OCR for long documents
- · Moonshot AI Unveils Kimi K3: The First Open 2.8 Trillion Parameter Model
#54 · 14. Juli 2026
GPT-5.6 Solves Legendary Math Problem
- · Mastering Agentic AI: From Foundations to Deployment
- · Browser Use — Open-source AI agent for browser automation
- · OpenAI’s GPT-5.6 Breaks New Ground: Proves Cycle Double Cover Conjecture
#53 · 30. Juni 2026
US Export Controls Hit Anthropic
- · Master Machine Learning in 12 Weeks—Free Microsoft Course
- · whichllm — Auto-rank the best local LLMs for your hardware
- · US Export Controls Force Anthropic to Disable Fable 5 and Mythos 5 Models Globally
#52 · 16. Juni 2026
Gemini Omni: Multimodal Video AI
- · Neural Networks: Zero to Hero—Karpathy’s Free Course
- · CodeGraph — Semantic code graphs for blazing-fast AI code exploration
- · Google Unveils Gemini Omni: The “Anything-to-Anything” Model for Video and Beyond
#51 · 2. Juni 2026
Composer 2.5: SOTA-Level Coding for Less
- · Mastering Retrieval-Augmented Generation: The Ultimate Guide
- · agentmemory — Persistent, cross-session memory for AI coding agents
- · Cursor Unveils Composer 2.5: A Major Leap in Coding Model Intelligence
#50 · 19. Mai 2026
Thinking Machines: a new kind of conversational AI
- · Teaching AI Ethics: Anthropic’s Alignment Methods
- · Spec Kit — From app ideas to agent-ready blueprints
- · Thinking Machines Unveils Real-Time, Multimodal Conversational AI Model
#49 · 6. Mai 2026
Codex Goes Beyond Coding
- · When to Use Single vs Multi-Agent AI: A Practical Guide
- · open-slide — AI-powered slide deck generation from natural language prompts
- · OpenAI’s April Stack Update: GPT-5.5, Agents-SDK, Images 2.0, and Codex
#48 · 21. April 2026
Gemma 4: open, Apache 2.0, top-ranked
- · Get Started with Codex 101
- · Gemini 3.1 Flash TTS — Next-generation expressive AI speech synthesis
- · Google DeepMind Launches Gemma 4
#47 · 7. April 2026
DeepMind’s AI Agent Traps
- · Building Durable AI Agent Memory with Oracle
- · Ollama + VS Code — Local and Cloud AI Models in Your IDE
- · Google DeepMind Unveils "AI Agent Traps" to Combat Web-Based Manipulation Attacks
#46 · 24. März 2026
AMI Labs’ $1B World Models
- · OpenClaw-RL: Continuous RL from Real Interactions
- · Unsloth Studio — Open-source UI for local LLM training and inference
- · Yann LeCun’s AMI Labs Raises $1.03B to Pioneer Real-World AI Models
#45 · 10. März 2026
Symphony Turns Tasks Into Finished PRs
- · Supercharge Copilot Studio with SharePoint Lists
- · Symphony — Open-source autonomous project management for technical teams
- · Anthropic Study Reveals AI Conversations Can Subtly Distort User Beliefs and Actions
#44 · 24. Februar 2026
Real-Time Coding with OpenAI Codex-Spark
- · Unlocking Agentic Document Extraction with AI
- · Claude Code to Figma — Instantly convert production code UIs into editable Figma designs
- · OpenAI Unveils GPT-5.3-Codex-Spark: Real-Time Coding at Unprecedented Speed
#43 · 10. Februar 2026
Claude AI on Mars
- · Run GLM-4.7-Flash Locally for Coding AI
- · GLM-OCR — State-of-the-art open-source multimodal document understanding
- · Anthropic’s Claude AI Powers First Autonomous Mars Rover Drive
#42 · 27. Jänner 2026
Anthropic's AI Economic Index
- · Mastering Reliable Coding Agents
- · json-render — Deterministic AI-to-UI JSON Streaming for Safe, Dynamic Interfaces
- · Anthropic Unveils New Economic Primitives to Track Real-World AI Impact
#41 · 13. Jänner 2026
Qwen’s Image Layer Decomposition Model
- · Supercharge Your AI Agents with Agent Skills
- · Qwen-Image-Layered — Image editing using Layer Decomposition
- · Anthropic Launches Bloom: Open-Source Framework for Fast, Automated AI Model Evaluation
#40 · 16. Dezember 2025
Disney's $1B OpenAI Sora Deal
- · Accelerate LLMs with Semantic Caching in Python
- · Cursor Visual Editor — Real-Time, Code-Synced UI Editing for React Apps
- · OpenAI Unveils GPT-5.2: Setting a New Standard for Professional AI Performance
#39 · 2. Dezember 2025
Elevate Visual AI Projects with FLUX.2
- · Mastering Enterprise AI Workflows with Microsoft Agent Framework
- · Unified text-driven segmentation for images and video with Meta’s SAM 3
- · Black Forest Labs Unveils FLUX.2: Open-Weight Multi-Reference Image Model Sets New Standard
#38 · 18. November 2025
Google’s Sun-Powered Space Data Center
- · Build Enterprise-Ready AI Agents with Microsoft Agent Framework
- · Agent Lightning — Universal optimization for any AI agent framework
- · Google Introduces Nested Learning: A Brain-Inspired Breakthrough for Continual AI Training
#37 · 4. November 2025
Frontier AI Model 'Personalities' Revealed
- · Transformers Demystified: A Visual, Intuitive Guide
- · Tinker — Flexible API for Efficient Language Model Fine-Tuning
- · New Anthropic Study Reveals Hidden Value Conflicts and Personalities in Leading AI Models
#36 · 21. Oktober 2025
Agents with Skills
- · Hands-On with the Microsoft Agent Framework
- · RAG‑Anything — An All-in-One RAG Framework for Multimodal Content
- · Anthropic introduces “Skills” — modular expertise for Claude
#35 · 7. Oktober 2025
AI Agents in Microsoft Office
- · “Awesome Generative AI Guide” — Your One-Stop Learning Hub
- · ‘text-to-SQL-eval’ — benchmark your LLM → SQL translations
- · Microsoft introduces “Vibe Working” and the Office Agent in Copilot
#34 · 23. September 2025
Deterministic LLMs Are Here
- · Microsoft’s “AI‑For‑Beginners” — A Full Curriculum in One Repo
- · Sim — Build & Deploy Agent Workflows Fast & Open
- · LLMs That Don’t Change Their Mind — Solving AI’s Repeatability Problem
#33 · 9. September 2025
HyprNote: Private Offline AI Notepad
- · Your Ultimate Generative AI Toolbox in One Repo
- · Hyprnote — Your Privacy-First AI Notepad for Meetings
- · Google's “Nano Banana” Makes AI Image Editing Smarter and More Consistent
#32 · 26. August 2025
DeepSeek Beats Claude at 11% Cost
- · Build a Market Research Agent with Gemini and Vercel’s AI SDK
- · Vercel AI Gateway — Your Unified Access to ANY Model, Anywhere
- · Mixture of Recursion (MoR) —
#31 · 12. August 2025
GPT-5: Big leap towards AGI?
- · Scale LLMs Smartly with FastMCP’s LLM‑Sampling
- · WebAgent – Alibaba’s Fully Open Source Browser Agent
- · GPT‑5 – Big Step Toward AGI or Just a Small Bump Forward?
#30 · 22. Juli 2025
BigQuery via MCP
- · MCP Server for BigQuery — Connect Your Data Warehouse to AI Agents
- · Meet Kimi K2 – A 1T-Parameter Open Agentic Model
#29 · 1. Juli 2025
Claude’s Multi-Agent Insights
- · RAGFlow – Real-World RAG Made Simple
- · agenticSeek — Fast, Open-Source Search for LLM Agents
- · Microsoft Unveils Mu – A Lightweight Language Model Powering the Windows Settings Agent
#28 · 17. Juni 2025
Rethinking AI Reasoning
- · Automating MCP Server Creation from OpenAPI and FastAPI
- · Claude Task Master – AI-Powered Project Management for Developers
- · The Illusion of Thinking — Apple’s Reality Check on Reasoning Models
#27 · 3. Juni 2025
Europe’s AI Challenger Rises
- · How to Get the Most Out of Claude 4 – Tools, Practices & Daily Use
- · Veo 3 — Google DeepMind’s Leap in AI Video Generation
- · Mistral AI Accelerates Europe's AI Momentum with a Suite of Innovative Tools
#26 · 20. Mai 2025
Auto-generate MCP servers from existing APIs
- · Automatically wrap existing APIs in MCP servers - using FastMCP
- · Tool Spotlight: OpenAI GPT-image-1 - The Next Evolution in AI Image Generation
- · Top News: AlphaEvolve - Google's Gemini-Powered Agent Discovers groundbreaking algorithms
#25 · 6. Mai 2025
Build MCP Servers in Minutes
- · Build a Custom MCP Server in Minutes with FastMCP
- · Napkin AI – Instantly Transform Text into Visuals
- · "The Leaderboard Illusion"—New Study Questions Fairness of AI Model Rankings
#24 · 22. April 2025
Global AI Trends in One Report
- · RAG Isn’t Dead—Microsoft’s Hands-On Guide Reminds Us Why It Still Matters
- · AI Memory Bank – Persistent Project Context for AI Coding Assistants
- · OpenAI Releases GPT-4.1 Model Family
#23 · 8. April 2025
OpenAI Joins MCP Standard
- · OpenWebUI Got a Big Upgrade — Here’s How to Use It
- · Langfuse – Observability for LLMs Made Simple
- · OpenAI Releases GPT-4o Image Generation and Embraces Anthropic’s MCP Standard
#22 · 25. März 2025
Manus AI: Is This What AGI Looks Like?
- · Building AI Agents with n8n - A Low-Code Approach to AI Workflows
- · ManusAI: The first glimpse at what “AGI” might look like?
#21 · 11. März 2025
Mistral OCR Revolutionizes Data Extraction
- · Advanced AI Agents—Building Smarter, Multi-Step Systems
- · Tool of the Week: Mistral OCR—Advancing Document Understanding with AI-Ready Output
- · OpenAI’s GPT-4.5: Smarter Conversations, But at a Steep Price
#20 · 25. Februar 2025
Real-World AI Usage and Use-Case Statistics
- · Build AI Agents From Scratch – No Frameworks Needed!
#19 · 11. Februar 2025
EU AI Act Goes Live - what companies need to know
- · Building a RAG System with Azure AI Search
#18 · 28. Jänner 2025
DeepSeek R1 vs. OpenAI o1: Same Power, 30x Cheaper
- · Optimize Azure AI Content Filters for Faster, Smarter Moderation
#17 · 14. Jänner 2025
Enterprise AI Agent Case Studies
- · Aligning and train small LLMs for your specific use case
#16 · 17. Dezember 2024
Automatic AI Prompt Generation
- · DSPy: Build Better AI Systems with Automated Prompt Optimization
#15 · 3. Dezember 2024
Context-Preserved Chunking
- · Late Chunking: A Method for Context Preservation in AI Retrieval
- · Anthropic's Model Context Protocol (MCP)
#14 · 19. November 2024
GPT-5 Hits Ceiling
- · Building a Smart Search for Your Company Docs: A Practical Approach
- · Leak suggests: Diminishing returns with GPT-5 - are LLMs hitting their ceiling?
#13 · 5. November 2024
Beyond Basic Prompting: How to connect LLMs with everyday tools
- · Beyond Basic Prompting: How to connect LLMs with everyday tools
- · Anthropic Claude 3.5 Sonnet (new) is by far the best LLM for now
#12 · 22. Oktober 2024
OpenAI’s Realtime API redefines live interactions
- · OpenAI Releases Realtime API
- · FireCrawl – Crawling Websites for LLMs
#11 · 8. Oktober 2024
Model Distillation - how to get a fast, strong and cost efficient LLM
- · OpenAI releases Model Distillation - creating strong, small, cost efficient models for FREE
- · Case Study: Building an AI content recommendation system with only PostgreSQL
#10 · 24. September 2024
Microsoft wants to restart old nuclear power plant for their AI data centers
- · OpenAI releases new “thinking” LLM - with best in class performance on complex tasks
- · Retrieval Augmented Generation, Vector Search, Embeddings and LLMs in Google BigQuery
#9 · 10. September 2024
Elon Musk purchases 100.000 H100 GPUs
- · Claude Enterprise - Anthropic's Business Chat-bot with interesting new features
- · Extracting data, images and more from documents using GPT-4o
#8 · 27. August 2024
Privacy-friendly ChatGPT alternative: Open WebUI
- · Automated Design of Agentic Systems (ADAS) - When Agents design Agents
- · Using GPT-4o to analyse images
#7
#7 · 13. August 2024
Free fine-tuning of GPT-4o-mini
- · Surprise-Launch: A highly skilled German team launched the so-far best image generation model, named Flux.1
- · FREE fine-tuning of GPT-4o-mini: A hands-on guide
#6
#6 · 29. Juli 2024
Neo4j knowledge graph builder, CopilotKit - a framework to integrate copilots in your applications, Llama 3.1 - the best local-first model yet
- · Llama 3.1 paves the road for many highly efficient LLM applications
- · Create knowledge graphs with Neo4j LLM knowledge graph builder
#5
#5 · 15. Juli 2024
LLMs for research & writing
- · Microsoft releases GraphRAG: A tool for complex data discovery
- · Chat with your Database: Use natural language to query data
#4
#4 · 1. Juli 2024
Gemma 2, Automated Document Extraction, AI in Your Database and More!
- · Google Releases Gemma 2
- · Extract Data from Documents Using Large Language Models
#3
#3 · 17. Juni 2024
AI Circuit Breakers, Apple's AI Foundation Models, and More: Check Out Pondhouse AI's 3rd Edition!
- · Improving AI Safety with Circuit Breakers
- · Building Open Source LLM Agents with Llama 3
#2
#2 · 3. Juni 2024
Pondhouse Data AI - Edition 2
- · News: We got a first glimpse inside the “mind” of LLMs
- · Tutorial: How to securely run your LLM
- · Tip of the week: Human-in-the-loop with AI applications
- · Tool of the week: Phidata - a framework for building Autonomous Assistants
#1
#1 · 16. Mai 2024
Pondhouse Data AI - Edition 1
- · News: OpenAI released a new, multimodal LLM, supporting text, image and audio - all in one model.
- · Tutorial: Comparison of different methods on how to use language models with proprietary data.
- · Tip of the week: Compressing conversation memory using LlamaIndex
- · Tool of the week: GPT-Researcher - an autonomous agent to create web research