GenAI/AI News from Jul 11 25: Demystify AI Agents by Building One! MSFT and Liquid AI Announce Powerful New On Device Models and More...
Our friends at Langchain provide paint by numbers agent building.
There are a few things that are more empowering than taking a concept that everybody's talking about and actually building one. There is no substitute for building in terms of the expertise. It gives you our friends at Lang Chang have provided a fantastic step-by-step guide on how to build your own AI agent.
Before getting into the articles, I want to remind you that in 79 days GAI World 2025 Is Coming – Are You Ready? Working hard to get fantastic speakers for introduction everyone from the Mayo Clinic. The John Hancock insurance are going to tell the benefits and challenges of truly implementing generative AI in real companies doing important work — the date is September 29–30 at the Hynes Convention Center in Boston. GAI World 2025 is where AI leaders, builders, and bold decision-makers collide to share what’s working right now in GenAI. With 800+ attendees, 120 powerhouse speakers, and the most curated AI networking experience available, this is the event where strategies get sharpened, products get launched, and C-suites leave with answers. Want in? Attend, sponsor, or exhibit: www.gaiworld.com
Good Morning, GAI Insights Community!
If you missed us Friday, watch the show on YouTube. Join the fun! Watch and comment on our Live feed on Linkedin or YouTube. We post articles for the daily briefing here so you can follow along and comment.
Today there were 1 Essential, 3 Important, and 2 Optional articles.
Rating: Essential
Rationale: This guide provides a clear, step‑by‑step primer on designing, testing, and deploying AI agents in production—covering data collection, MVP building, safety checks, and orchestration. Our analysts praised it for demystifying agent creation, helping everyone from leaders to developers gain practical confidence in AI‑driven agentic content creation. Check out the handy dandy table below — from the article.
Introducing LFM2: The Fastest On‑Device Foundation Models on the Market
Rating: Important
Rationale: LFM2 delivers record‑setting speed and memory efficiency—2× faster than Qwen3—with hybrid architecture optimized for edge deployment on phones, laptops, and wearables. Our analysts highlighted how this marks critical innovation in on‑device AI, enabling smarter devices in 2026 and beyond. Also, I love Liquid AI's name and logo.
How Rocket streamlines the home buying experience with Amazon Bedrock Agent
Rating: Important
Rationale: Rocket Mortgage leverages Amazon Bedrock Agents to deliver context‑aware assistant—boosting retention and scheduling efficiency with real‑time engagement. Our analysts emphasize this as a strong AI case study showing measurable ROI and operational improvements.
Reasoning reimagined: Introducing Phi‑4‑mini‑flash‑reasoning
Rating: Important
Rationale: Microsoft’s Phi‑4‑mini‑flash‑reasoning delivers a Samba hybrid architecture—achieving up to 10× faster throughput and 2–3× lower latency on edge and mobile reasoning tasks. Our analysts noted its practical value for IoT and robotics, making efficient on‑device reasoning increasingly achievable.
Upgrading agentic coding capabilities with the new Devstral models
Rating: Optional
Rationale: Mistral’s DevStral (Small and Medium) offers high‑performance, Apache‑licensed agentic coding models that hit strong SWE‑Bench scores at low cost—beating Gemini 2.5 Pro and GPT‑4.1. Our analysts saw promise in its open‑source, tool‑aware design, though it landed in a crowded space with marginal incremental novelty.
Ramp agents: Let finance teams do finance
Rating: Optional
Rationale: Ramp's agent automate finance workflows, policy enforcement, and conversational interactions within finance teams. Our analysts regarded it as a product announcement in an emerging automation trend, noting significant potential but limited differentiation from other solutions.
Like this email? Refer a friend! They can sign up here.
GAI Insights is an industry analyst firm helping AI leaders and their teams achieve business results with GenAl.
Enjoy! John Sviokla, Paul Baier and our analysts Luda Kopeikina and Adam Rappaport.