The FDE Coach blog · new articles every day
Practical guides for engineers who ship AI.
Step-by-step builds with free AI tools, what just shipped in AI and why it matters, and Forward Deployed Engineer playbooks on interviews, pay and client work.
What just shipped in AI, and why it matters for engineers.
Page 6 of 11
AI News
Qwen3.8 Max Tops Agentic Index: The New Benchmark for Agent-First AI
Qwen3.8 Max now leads the agentic index rankings. We break down what this shift means for engineers building tool-using AI, how to test it today, and why agentic benchmarks are replacing chatbot evals.
11 min read
AI News
AMD Acquires Taalas to Etch AI Models into Silicon for Inference
AMD's acquisition of Taalas signals a shift from programmable GPUs to hard-wired AI. We break down what 'etching models into silicon' means for inference latency, cost, and the FDE toolkit.
9 min read
AI News
Architecting an Agentic Harness: Tool Use, Memory & Routing Patterns
Strip away the hype. We dissect a concrete agentic harness architecture—covering tool execution, memory management, and intelligent routing—so you can build reliable AI agents that ship.
9 min read
AI News
Humans Miss 1 in 3 AI Threats: The Agent Approval Blind Spot
A 40k game-run study reveals humans miss 33% of threats when approving AI agent commands at scale. We break down the failure mode, why it matters for engineers shipping agentic systems, and practical guardrails you can implement today.
9 min read
AI News
Prime Agent: Self-Improving RL Creates AI Agents That Build Better AI Agents
Prime Intellect's Prime Agent uses self-improving reinforcement learning to automate AI agent creation. Here's how it works, why it matters for engineers, and how to start experimenting today.
12 min read
AI News
Warp Agent CLI: The Native Terminal Coding Agent Engineers Need
Warp drops a coding agent directly into your terminal, not a separate IDE pane. Here's how the Warp Agent CLI works, why it matters for engineers and FDEs, and how to try it today.
12 min read
AI News
Cloudflare OS: The Open Edge Platform for Agents & Apps
Cloudflare OS isn't a new operating system—it's a rebundling of their edge primitives (Workers, Durable Objects, AI) into an open platform for building autonomous agents and distributed applications. Here's what changed, why it matters for engineers, and how to use it today.
12 min read
AI News
Beating GPT-4o on Retrieval with a 100x Cheaper Open Stack
Neon’s Castform stack uses open-source models to beat GPT-4o on document retrieval while being 100x cheaper. Here’s the engineering breakdown and how to build it yourself.
10 min read
AI News
Rust’s Formal LLM Policy: What It Means for Working Engineers
The Rust project just adopted a formal LLM policy. We break down the rules, the rationale, and what this signals for engineers building with AI—plus how to apply similar guardrails in your own stack.
9 min read
AI News
TIME Serves AI Bots an Ad-Loaded Site: The Engineering Playbook
TIME.com is now serving a separate, ad-injected version of its content to AI crawlers. We break down the mechanics, why it matters for the data supply chain, and how engineers can replicate or defend against this tactic.
11 min read
AI News
Inside the Shai-Hulud Supply Chain Attack: How Keyv and Friends Were Compromised on npm
A deep dive into the Shai-Hulud npm supply chain attack that compromised Keyv, cacheable, and other packages. Learn the technical play-by-play, why it matters for FDEs, and how to harden your pipelines.
9 min read
AI News
Shieldstral: Mistral’s 3B Open-Weights Model for Multimodal Moderation
Mistral drops Shieldstral, a 3B open-weights model for text and image moderation. Here’s why it matters for engineers building safe AI pipelines and how to run it today.
9 min read
AI News
80B Qwen on 4.3 GB RAM: How Swiftlet Quantization Works
Swiftlet quantizes an 80B Qwen model to run on a Mac with just 4.3 GB of RAM. We unpack the extreme compression techniques, why this matters for FDEs building local-first AI, and how to try it today.
9 min read
AI News
DeepSeek V4 Flash on a Single AMD MI300X: Inference Engineering Deep Dive
A technical walkthrough of running the massive DeepSeek V4 Flash model on a single AMD MI300X GPU. We break down the quantization and KV-cache wizardry that makes this possible, and what it means for forward-deployed inference engineering.
8 min read
AI News
The Engineering Cost of AI-Generated Blog Images on Reader Trust and UX
AI-generated images are eroding reader trust and tanking content UX. We break down the engineering psychology behind the backlash and provide a high-signal workflow for authentic visual assets.
10 min read
AI News
How AirLLM Runs 70B Models on a 4GB GPU with Layer-Wise Loading
AirLLM inference explained: how layer-wise loading lets you run Llama 70B on a single 4GB GPU. Covers the engineering mechanics, memory math, and a balanced take on when it's actually useful for FDEs.
9 min read
AI News
Why LLMs Amplify the Output Gap Between Senior and Junior Engineers
LLMs aren't a leveling force—they're a force multiplier for expertise. Here's why senior engineers extract 10x more value from AI tools, and how to close the gap deliberately.
10 min read
AI News
Retyping LLM Code to Prevent Cognitive Debt: An Engineer's Deliberate Practice
Blindly copy-pasting AI-generated code creates dangerous cognitive debt. Discover why manually retyping LLM output forces comprehension, builds intuition, and makes you a more resilient engineer.
8 min read
AI News
When LLMs Hallucinate CVEs: The SQLite Vulnerability Report That Wasn't
A researcher used an LLM to generate a SQLite CVE report. The result was a masterclass in hallucination: plausible, detailed, and dangerously wrong. Here's what happened and why it matters for engineers shipping AI features.
10 min read
AI News
Qwen3.8-Max Coding Performance: Rethinking Local Dev Workflows
Qwen3.8-Max challenges proprietary coding models at a fraction of the cost. We break down the benchmarks, what it means for engineers shipping code, and how to run it locally today.
9 min read
AI News
MI355X vs B300 for MoE Inference: Kimi K3 Throughput per Dollar Deep Dive
AMD's MI355X runs Kimi K3 MoE inference at better performance per dollar than Nvidia's B300. We break down the benchmark, why memory bandwidth dominates cost, and what this means for engineers deploying large-scale models.
9 min read
AI News
Running Kimi K3 Locally: Memory Bandwidth and the True Cost of 0.5 tok/s
We ran Kimi K3 on consumer hardware using 29 GB of RAM. The 0.5 tok/s throughput isn't a software bug—it's a physics lesson in memory bandwidth. Here's what happened, why it matters, and how to try it.
9 min read
AI News
The Prototype-Product Gap: Why LLM Code Needs Systems Thinking to Ship
LLMs can generate a prototype in seconds, but that isn't a product. Explore the invisible systems work—state management, error handling, deployment—that separates a demo from shippable software.
9 min read
AI News
Constraint Injection: The Engineering Secret to Hallucination-Free Financial AI
MIT research confirms AI financial advice is surprisingly good when you ask the right questions. Here's the engineer's playbook for structuring prompts with constraint injection to extract useful, grounded LLM advice without hallucination.
11 min read
259 articles and counting