[ Blog ]
AI, automation, and the tools we use
Articles on AI agents, automation, and developer tools.
Jev AI Limitations: What the 193x Benchmark Doesn't Measure
TypeSafe's Jev returns typed decisions instead of text, answers in 70ms, and charges nothing for output tokens. The 193.6x speed and 444.6x cost claims come from an eval where the correct answer is defined as what two other models said. Here is what Jev actually does, and what the benchmark measures.
Laya vs Jev: Is the Open-Source Alternative Actually as Good?
Convai Innovations released Laya under Apache 2.0 days after Jev launched. It answers in 33ms locally, is free to self-host, and beats Jev on several classification tasks. It also scores below the majority-class baseline out of the box, and Convai says so on its own model card. A fair comparison.
AI Sales Replies That Sound Like Your Team, Not Like Everyone Else
Your sequencer already handles volume. What fails is sameness. How a shared persona lets every rep on a small sales team draft in the same voice.
One Prompt, One Draft, Eight Social Posts
A week of social posts usually means a patchwork of design tools and one picture re-cropped eight times. Latent replaces that with one prompt and one draft.
The Agency Operating System: Why We Stopped Patching Workflows and Built One Portal
Why Silverthread Labs stopped patching workflows and built an agency operating system: one portal for the work, the leads, the spend, and the team, with the data on what six tabs actually cost.
Claude Fable 5 Banned 3 Days After Launch — Over a Request to Fix Code
The US government suspended Claude Fable 5 and Mythos 5 three days after launch, citing national security. The "jailbreak" behind it was asking the model to read a codebase and fix bugs — something GPT-5.5 does without any workaround. Here's what happened and why it matters.
Claude Fable 5 or Mythos 5: The Difference Comes Down to Who You Are
Anthropic shipped the same frontier model in two versions: Fable 5 for everyone, Mythos 5 for vetted cyber defenders. Here's exactly what separates them — and which one you'll actually be using.
Claude Agent SDK vs OpenAI Agents SDK vs Google ADK (2025)
A technical comparison of the three leading AI agent SDKs: Claude Agent SDK, OpenAI Agents SDK, and Google ADK. Architecture differences, feature matrix, and when to use each one.
AI for Dental Practices: The Complete Guide
Voice AI, workflow automation, and patient management — how dental practices are using AI to answer every call, cut admin time, and fill their schedule. Practical guide.