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 8 of 11
AI News
Opus 5 on SlopCodeBench: Why Standard Coding Benchmarks Lie to You
Claude Opus 5 scores 85% on SWE-bench but fails hard on real-world slop. We break down SlopCodeBench, why it matters for Forward Deployed Engineers, and how to actually evaluate AI coding agents today.
10 min read
AI News
Kimi K3’s Attention Explained: A DeltaNet Walkthrough for Engineers
Kimi K3 didn't invent magic; it rediscovered DeltaNet. Walk through the linear attention math, see why it matters for long-context inference, and learn how to implement it today.
9 min read
AI News
Debian's LLM Policy: What the Vote Means for Engineers Who Ship Code
Debian's 2026 vote on LLM use in package maintenance isn't just governance theater—it's a blueprint for every engineer deploying AI-assisted code. Here's what the three proposals actually say and how to apply the rules today.
9 min read
AI News
MAI-Cyber-1-Flash: Microsoft’s Security-Tuned Model Inside MDASH
Microsoft drops a domain-specific small model for cybersecurity. We cut through the announcement to explain what MAI-Cyber-1-Flash is, how it fits into the MDASH agent, and why it signals a shift toward specialized, deployable AI for security engineers.
9 min read
AI News
Cloudflare's AI Crawler Block: What Engineers & FDEs Need to Know
Cloudflare now lets customers block AI crawlers with one click. We break down the mechanics, the real-world deployment impact for FDEs, and a practical guide to enable it today.
9 min read
AI News
LLM on an $8 ESP32: How a 29M-Param Model Runs on a Microcontroller
A 28.9M parameter LLM now runs on an $8 ESP32-S3. We break down the quantization, memory tricks, and inference engine that make it possible, and why this matters for embedded engineers and Forward Deployed Engineers building on the edge.
10 min read
AI News
Scriptc: Compiling TypeScript Straight to Native Binaries With Zero Runtime
Vercel's Scriptc compiles TypeScript to standalone native binaries without a JavaScript runtime. We unpack the mechanics, why this matters for CLI tooling and edge deployments, and how to try it today.
8 min read
AI News
UK AISI's Kimi K3 Cyber Audit: An Engineer's Breakdown
We reverse-engineer the UK AISI's evaluation of Kimi K3's cyber capabilities. Learn the CTF-based methodology, why it matters for forward-deployed engineering, and how to apply these adversarial testing patterns today.
10 min read
AI News
Corporate AI Spending Pullback: Budget Cuts or Efficiency Gains?
Corporate AI budgets are shrinking, but it's not a retreat. It's a shift from speculative moonshots to engineered efficiency. Here's what the pullback really means for engineers and FDEs.
8 min read
AI News
Claude Opus 5 Elevated Errors: What the Incident Reveals About API Architecture
Anthropic's Opus 5 outage wasn't just a blip—it exposed the fragile layers between model inference and API surface. Here's what engineers and FDEs should take away about retry logic, circuit breakers, and the new shape of AI reliability.
10 min read
AI News
DeepSeek Halts Fundraising After Leaked Comments on US Compute Gap
Leaked investor transcripts reveal DeepSeek’s leadership grappling with a massive compute gap behind US labs. Here’s the unvarnished engineering reality and what it means for AI builders.
9 min read
AI News
Why Open-Weight AI Is Repeating the Kubernetes Operational Playbook
Open-weight models are having their Kubernetes moment: powerful but operationally raw. Engineers and FDEs must now solve the same scheduling, serving, and security problems that cloud-native infrastructure solved a decade ago.
10 min read
AI News
Framework Desktop Packs 192 GB RAM & AMD Ryzen AI Max+ Pro 495 for Local LLMs
Framework’s new desktop board crams 192 GB of unified memory and an AMD Ryzen AI Max+ Pro 495 into a single tile. For engineers running large local models, this changes the calculus on quantized vs. full-precision inference.
12 min read
AI News
Claude-Thermos: Keep Your Claude Session Alive and Avoid Cold Starts
A deep dive into Claude-Thermos, a utility that prevents Anthropic Claude Code sessions from dying. Learn why cold starts kill engineer flow, how the tool works, and how to deploy it.
9 min read
AI News
Palmier Pro: An Open-Source macOS Video Editor Built from the Ground Up for AI Pipelines
Palmier Pro isn't a Premiere clone. It's a native macOS video editor designed as an API-first node in an AI pipeline. Here's why forward-deployed engineers should care.
11 min read
AI News
Hetzner LLM Inference: What Bare-Metal GPU Pricing Means for Self-Hosters
Hetzner is building an LLM inference service on bare-metal GPU instances. We break down the leaked specs, pricing implications for self-hosters, and how to benchmark it against your current stack.
10 min read
AI News
Context Engineering for Claude 5: Prompting When the Model Actually Reads the Docs
Claude 5 doesn't need hand-holding—it reads documentation natively. Here's how to engineer context for a model that parses your entire codebase, not just your prompt.
11 min read
AI News
How Terence Tao Used ChatGPT to Stress-Test a Jacobian Conjecture Counterexample
Terence Tao's live ChatGPT session wasn't just math—it was a masterclass in AI-assisted exploration. Here's what engineers and FDEs can steal from his method of probing a complex counterexample.
10 min read
AI News
The Open-Weight AI Fight: What the Nvidia/Meta/Microsoft Letter Means for Your Stack
Nvidia, Meta, and Microsoft just warned against overregulating open-weight AI models. Here's what 'open-weight' actually means for engineers, how it changes your deployment stack, and why Forward Deployed Engineers should care.
9 min read
AI News
Why AI Coding Agents Stall: The Context Engineering Gap No One Talks About
AI coding agents don't fail because of weak models. They fail because of weak context. Learn why software factories stall and how engineers can inject the missing signal to ship faster.
11 min read
AI News
Flux 3 Mimic: Why Video-Action Models Kill Frame Prediction
Black Forest Labs' Flux 3 Mimic shifts video generation from pixel-prediction to action-modeling. We break down the architecture, why it matters for robotics and UI automation, and how to start prototyping today.
9 min read
AI News
Claude Opus 5: What Extended Thinking Means for Complex Engineering Work
Anthropic's Claude Opus 5 introduces visible extended thinking. We break down what this means for Forward Deployed Engineers and how to leverage it for high-stakes debugging, schema migration, and code review.
11 min read
AI News
Petals: Running 100B+ LLMs at Home with a BitTorrent-Style Network
Petals lets you run massive language models like Llama 3 70B on consumer hardware by splitting inference across a decentralized peer-to-peer network. Here's how it works under the hood and how to use it today.
10 min read
AI News
Can a MUD Evaluate LLMs? A $99 Game-Based Benchmark POC
Crucible ran LLMs through a text adventure game for $99. Here's why MUD-based evals expose reasoning gaps standard benchmarks miss, and how you can build your own.
9 min read
259 articles and counting