OpenAI and Anthropic's New Alliance

Plus, 🗣️ Using Gemini Pro 3 to prompt the best landing pages, Mistral Launches Devstral 2 and Its First Coding Agent

Hey There😎

Ready to dive into the world of AI? We’ve got some fascinating reads, a tip to up your game, and a tool you won’t want to miss. Let’s get started!

Oh and before we go ahead! If your friend sent this to you, be sure to subscribe below! So you don’t miss out on any.

---

🔧 The Agentic AI Foundation: Big Tech’s New Alliance for Open-Standard AI Agents

OpenAI, Anthropic, and Block just teamed up under the Linux Foundation to make sure the AI agent future doesn’t turn into a proprietary mess.

The new Agentic AI Foundation (AAIF) brings their core tools, MCP, Goose, and AGENTS.md, into one open, neutral home. It’s timed perfectly with enterprise AI spend exploding and companies needing tech that actually works together.

1. Three major projects now form the backbone of AAIF

Anthropic’s Model Context Protocol (MCP) is already the go-to way to connect AI models to tools, used across Claude, Copilot, Gemini, ChatGPT, Cursor and thousands of MCP servers.

Block’s Goose framework gives developers a simple, local-first way to build reliable agents. OpenAI’s AGENTS.md standardizes how coding agents understand instructions in any repo, now adopted by 60,000+ projects.

2. The goal: keep the agent ecosystem open, not locked down

The Linux Foundation says AAIF ensures these fast-growing tools evolve transparently and with community input, the same model that helped Kubernetes win. Major members include AWS, Google, Microsoft, Cloudflare, Bloomberg, Block, Anthropic, and OpenAI.

3. The timing couldn’t be better

Enterprises spent $37B on genAI in 2025 (up 3.2x from last year), and 76% are buying solutions rather than building them. Coding AI alone hit $4B in spend, with Anthropic holding 54% of that market. Tools like Cursor even hit $200M revenue with zero enterprise sales reps.

Why it matters

Most “agents” today are still basic automations, but real ones are coming fast, and they need shared standards to avoid fragmentation. If MCP, Goose, and AGENTS.md become the defaults, developers can build once and run anywhere, making agents far more powerful and portable. AAIF is basically the industry trying to set the rules now, before things get messy later.

---

Together with Lindy

You are not behind on ideas. You are behind on speed.

You know the drill. Competitor research scattered across tabs, unfinished messaging docs, late briefs, and creative that misses the angle. That chaos is not annoying; it is expensive.

Lindy AI CMO is a suite of powerful agents that ship beautiful marketing campaigns quickly.

Drop in your website, and agents study your competitors, extract angles, draft messaging, create briefs, and generate launch-ready assets, then organize everything in Airtable for a clean approval flow.

• Cut campaign prep from 10 to 14 days down to 1 to 2 days.

• Generate 30 to 80 on-brief variants per angle to accelerate testing.

• Reduce rework and increase weekly experiment volume without extra headcount.

Teams running agent-led workflows consistently ship 2 to 3 times more weekly tests with the same headcount.

---

🚀 Mistral Launches Devstral 2 and Its First Coding Agent

Mistral just dropped Devstral 2, its new generation of coding-focused models, along with Vibe CLI, its first real step into autonomous coding agents. The release pushes both open-weight performance and local deployability forward.

1. Devstral 2 brings frontier-level coding performance

The 123B Devstral 2 model hits 72.2% on SWE-bench Verified, putting it just behind DeepSeek V3.2 despite being 5x smaller. The Small 2 variant is only 24B parameters, runs on a single GPU or even a laptop CPU, and still competes with the top open-weight models.

Both arrive under a modified MIT license that restricts use for companies making $20M+ in monthly revenue.

2. Vibe CLI marks Mistral’s move into coding agents

Vibe CLI is a terminal-native agent that scans codebases, handles multi-file edits, and is free under Apache 2.0. It’s designed for developers who want agentic workflows without cloud dependencies.

This positions Mistral directly against agent-focused tools emerging across the ecosystem.

Why it matters

Mistral is shipping fast, delivering Devstral 2 just days after Mistral 3, and is now competing at the frontier while still supporting local-first development. The flagship model pushes open-weight performance upward, and the small model could become the go-to for offline, consumer-grade coding workloads.

---

🧩 How to Audit a Landing Page and Find Conversion Leaks Using One Prompt

Turn a vague “something feels off” landing page into a clear, prioritized checklist of fixes that improve UX and boost conversions.

Prompt to paste: Conduct a comprehensive landing page audit of [Insert landing page URL]. Identify friction points across Headline Clarity: [Insert headline]. CTA Visibility: [Describe placement and wording]. Page Load Speed: [Insert load time]. Visual Hierarchy: [List observations]. Output a prioritized checklist with the top friction points ranked by conversion impact, quick wins under 60 minutes, rewritten headline and CTA options, UX improvements for mobile and trust, and 3 A B tests with success metrics. Keep it concise, specific, and action ready.

---

🤳AI Nugget of the Day

---

🗣️ Using Gemini Pro 3 to prompt the best landing pages
Via Meng To on X

---

That’s it for now, plenty to explore and try out! Keep experimenting, stay curious, and we’ll catch you next time with even more AI goodness.🥳