howdy, it’s Barsee again.

happy monday, AI family, and welcome back to AI Valley.

here are the biggest things worth knowing today:

• Pentagon announces deal with 8 AI companies but keeps Anthropic out

• Musk admits Grok was trained using OpenAI’s models

• Meta is moving into humanoid robots

• Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

SPOKI

You write. The AI does the rest.

Spoki turns a simple instruction into a fully running campaign.

Type: “Win back inactive customers” Spoki creates it, sends it, and keeps optimizing it automatically.

No dashboards. No workflows. Just tell Spoki what you want, and it gets it done.

This is the shift from software that gives you tools to software that delivers results.

Run everything from one place across WhatsApp, SMS, and voice. AI agents handle chats and calls 24/7, qualify leads, and guide users all the way to checkout.

Spoki isn’t just another CRM. It’s a Customer Engagement Management (CEM) platform that unifies marketing, sales, and support into one continuous conversation.

With an average +30% increase in sales, businesses are doing more with less effort.

Try Spoki for free.

_{*This is sponsored}

THROUGH THE VALLEY

1/ Pentagon announces deal with 8 AI companies but keeps Anthropic out

The Pentagon has added eight AI companies to its classified networks while continuing to exclude Anthropic, even though the new contracts reportedly include the same limits on autonomous weapons and surveillance that Anthropic was flagged over.

The list includes SpaceX, OpenAI, Google, Nvidia, Reflection, Microsoft, Amazon Web Services, and Oracle as part of its push to become an “AI-first” military.

At the same time, Anthropic is reportedly about to announce a $1.5B joint venture with Blackstone, Hellman & Friedman, and Goldman Sachs to sell AI tools to private equity-backed companies.

Why this matters:

Anthropic isn’t excluded because it’s irrelevant, it’s excluded because it’s too important to depend on. The U.S. is spreading contracts across multiple players to avoid a single point of control while still tracking frontier models closely. This is what early AI geopolitics looks like, not clear bans or approvals, but selective access and constant hedging between capability and dependence.

2/ Musk admits Grok was trained using OpenAI’s models

Under oath, Elon Musk confirmed that xAI used OpenAI models to help train Grok, describing distillation as “standard practice.” He made the statement during his lawsuit against Sam Altman and Greg Brockman.

He also ranked Anthropic as the leading lab, ahead of OpenAI and Google, while positioning xAI as a smaller player.

Why this matters:

This breaks the idea that top labs are building independently. Even competitors are learning from each other’s models, which makes progress more interconnected than it appears. That weakens the idea of “owning” intelligence and turns competition into something closer to building on top of each other than starting from scratch.

3/ Meta is moving into humanoid robots

Meta has acquired Assured Robot Intelligence (ARI), a startup building foundation models for humanoid robots capable of real-world physical tasks.

Co-founders Xiaolong Wang and Lerrel Pinto will join Meta’s Superintelligence Labs. The move builds on Meta’s existing robotics efforts led by Marc Whitten.

The push comes as Big Tech races to secure robotics talent, with market projections ranging from $38B by 2035 (Goldman Sachs) to $5T by 2050 (Morgan Stanley).

Why this matters:

Training on internet data got AI this far. Improving it further likely requires interaction with the physical world. Robotics is becoming a training strategy. Whoever owns systems that learn by acting, not just predicting, gets a compounding advantage that pure software models can’t easily match.

TRENDING TOOLS

• Remio > Set up once. remio captures everything and gets the work done — docs, emails, slides, reports _^(sponsored)

• Raven > AI can find someone’s exact location from a reflection in a pair of sunglasses using just pixels

• Mira > AI can now live on your face. capture every conversation. create the most personalized form of AI ever

• Glean Waldo > Glean’s new agentic search model, designed to work alongside LLMs with 50% lower latency and 25% fewer tokens

• ByteDance Seed3D 2.0 > A model capable of generating 3D objects from a single image or a text prompt

• Algrow > You can plug in a YouTube, Tiktok or Instagram URL and let Claude break it down for you

• Nodalist > Turns ideas and documents into structured nodes, layering AI-driven analysis, debates, and context-aware reasoning

• Meta Ads AI Connectors > Now in open beta, allowing advertisers to plug ChatGPT, Claude, and other AI assistants directly into their ad accounts

WHAT I'M CONSUMING

• A compact agent skill for turning short prompts into beautiful, realistic GPT Image 2 images

• OpenAI dropped a new prompting guide for GPT-5.5

• OpenAI explains why GPT-5.5 kept making goblin metaphors

• Prompt Engineering is permanent

• The most fun I’ve had building apps: GPT-5.5 + GPT-Image-2

• Most companies aren't anywhere near ready for AI

• You can build, train, and customize a humanoid robot

THE VALLEY GEMS

What’s trending on social today:

THAT’S ALL FOR TODAY

Thank you for reading today’s edition. That’s all for today’s issue.

💡 Help me get better and suggest new ideas at [email protected] or @heyBarsee

👍️ New reader? Subscribe here

Thanks for being here.

REACH 100K+ READERS

Acquire new customers and drive revenue by partnering with us

Sponsor AI Valley and reach over 100,000+ entrepreneurs, founders, software engineers, investors, etc.

If you’re interested in sponsoring us, email [email protected] with the subject “AI Valley Ads”.