Monitoring 29+ vendors across payments, cloud, and AI.
Every policy change, pricing shift, and compliance update
broken down weekly for indie devs and SaaS teams.

FEATURED

Anthropic Is Closing the 1M Token Context Window Beta for Sonnet 4 and 4.5 on April 30

Apr 9, 2026

•

3 min read

Anthropic Is Closing the 1M Token Context Window Beta for Sonnet 4 and 4.5 on April 30

If you're using the extended context header, you have 21 days before requests over 200k tokens start returning errors.

Anthropic

Trish T.

THE LATEST

Apr 9, 2026

•

3 min read

There's an Active Exploit in Flowise Right Now — Update Before Your Next Agent Run

CVE-2025-59528 is a perfect-10 severity code injection flaw being actively exploited. The patch has been out for six months. Most instances are still exposed.

FlowWise

Apr 9, 2026

•

3 min read

OpenAI Is Retiring DALL-E 2 and 3 on May 12 — The Replacement Isn't a Simple Swap

The new models are cheaper at low quality and more expensive at high quality. Know the difference before you migrate.

Pricing & Billing

Apr 9, 2026

•

4 min read

Anthropic Just Made the Hardest Part of Shipping Agents Someone Else's Problem

Production infrastructure for AI agents, available now — no months of setup required.

Agents

AWS Is Training on Your WorkSpaces Content Unless You Actively Stop It

Your client work, your proprietary code, your internal documents — all of it is fair game by default.

Amazon

The Infrastructure Connecting Your AI Agents to Everything Has a Security Debt Problem

97 million installs. Optional authentication. And the bill is starting to arrive.

MCP

Replicate Now Charges You for Failed API Calls — And Disputing That Just Got Harder

A quiet update to Replicate's API terms adds billing rules that most developers won't notice until something goes wrong.

Pricing & Billing

The Drift Intel TL;DR - Weekly RoundUp

April 3, 2026

Agents

Microsoft Just Launched Three AI Models in One Day — And They're Aimed at OpenAI

Microsoft's superintelligence team shipped MAI-Transcribe-1, MAI-Voice-1, and MAI-Image-2 today. The story isn't the models. It's who they're replacing.

vendor-watch

Anthropic Leaked Its Own Blueprint — And Every Competitor Already Has a Copy

The "safety-first" AI lab had its second major security lapse in a week. Here's what was actually inside the code.

Anthropic

The Company Building the Most Powerful AI Agent Just Leaked Its Own Code

Three stories converged in one week — and they all point to the same uncomfortable gap

Agents

GitHub Is About to Train On Your Code, Unless You Say No First

Every Copilot user just got an email. Most of them won't act on it before the deadline.

GitHub Copilot

SUBSCRIBE TO OUR NEWSLETTER

What's moving in your stack. Curated intel from vendor blogs, regulatory bodies, and tech news — relevant to you and your stack, delivered every Friday

FEATURED

Anthropic Is Closing the 1M Token Context Window Beta for Sonnet 4 and 4.5 on April 30

THE LATEST

There's an Active Exploit in Flowise Right Now — Update Before Your Next Agent Run

OpenAI Is Retiring DALL-E 2 and 3 on May 12 — The Replacement Isn't a Simple Swap

Anthropic Just Made the Hardest Part of Shipping Agents Someone Else's Problem

TRENDING

AWS Is Training on Your WorkSpaces Content Unless You Actively Stop It

The Infrastructure Connecting Your AI Agents to Everything Has a Security Debt Problem

Replicate Now Charges You for Failed API Calls — And Disputing That Just Got Harder

The Drift Intel TL;DR - Weekly RoundUp

Microsoft Just Launched Three AI Models in One Day — And They're Aimed at OpenAI

Anthropic Leaked Its Own Blueprint — And Every Competitor Already Has a Copy

The Company Building the Most Powerful AI Agent Just Leaked Its Own Code

GitHub Is About to Train On Your Code, Unless You Say No First

SUBSCRIBE TO OUR NEWSLETTER

STAY CONNECTED