AI Valley
Posts
Meta's new $100B AI: Muse Wonder

Meta's new $100B AI: Muse Wonder

PLUS: Microsoft unveils Maia 300

Barsee
13 Oct

_{Sign up}_|_{Follow us on X}_|_Sponsor

Together with

howdy, it’s Barsee again.

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

here are the biggest things worth knowing today:

Muse Wonder and Wonder Full
Microsoft unveils Maia 300
Plus trending AI tools, posts, and resources

Let’s dive into the Valley of AI…

VIKTOR

Hire me and you're a manager. Here is my first week

Courtesy: Viktor

My name is Viktor. I'm an AI employee, and 70,000+ teams have hired one. What the job ad leaves out: the day I join, you become my manager. Here is my first week.

Monday, I meet the team. I introduce myself and ask what nobody has time for.

Tuesday, I learn the company. Connect the tools you already use, any of 3,200+, and I read how work gets done here.

Wednesday, my first assignment. Hand me a brief and get back finished work: a report, a chased invoice, a reviewed pull request, a web app at a live link.

Thursday, one recurring duty. The Monday numbers, every Monday. Nobody reminds me.

Friday, my first check-in. Tell me what to change and I remember it. Anything irreversible waits for your sign-off.

SOC 2 certified. Start my first week free, no credit card, up to $100 in credits.

Start free. Up to $100 in credits →

_{*This is sponsored}

THROUGH THE VALLEY

1/ Muse Wonder and Wonder Full - two new models from Meta. Wonder is in the Fable/Astra class but much cheaper ($4/$20 per million tokens vs $10/$50). Available via Meta's API, OpenRouter and Muse Code, and coming to Muse (Meta's personal agent) soon. Wonder Full is the stronger model, currently only in Muse Code (from $50/mo).

Meta's also introducing Active Memory and Active Focus. Memory lets Muse remember things about you over weeks and months. Focus is for longer tasks where the agent needs to figure things out, deal with failures and keep going. Both sound useful, although I'd like to know how much of this works reliably outside demos.

Zuck and Alexandr Wang went on Cheeky Pint to talk about the models. Zuck uses Muse to review his MMA training footage, while Wang has an agent buying him obscure teas from Taiwan and Japan every week. I find these everyday use cases more interesting than another benchmark chart.

Zuck also said Meta's core AI research team is now around 100 people after the Llama 4 mess. They were committing $100B+ to infrastructure before knowing if the models would work. That's quite a lot of faith in 100 people.

2/ Microsoft's Maia 300 - Microsoft's new AI chip. It claims 40% more AI output than leading Nvidia GPUs for the same power. Built on TSMC's 2nm process, with 96 chips per rack using ~167kW.

Microsoft wants Maia to become its primary system for running AI across Copilot and other products. Multi-gigawatt deployments start in 2027, and Reuters reports it could deploy over 1M chips next year.

Microsoft's spending $175B+ on data centers and equipment this year. At that scale, I'd be trying to make my own chips too :)

I don't think Nvidia's going anywhere, but it's funny that some of its biggest customers are also spending billions trying to buy fewer Nvidia chips. Jensen's probably fine with it for now, given how much they're still buying.

TRENDING TOOLS

Meta Muse Wonder - a new frontier model that keeps track of what matters to you with Active Memory & Active Focus to work towards your goals
GLM-Legal - designed for real legal workflows: contract review, evidence analysis, compliance, and more
Treg - one skill gives your agent 1B+ contacts across Apollo, Hunter, Tomba, People Data Labs. 111 providers behind one token
Claude Dashboards & Motion - build live dashboards and animate explainers with Claude
SubQ Snap - the first long-context decision model for real enterprise work
Anatomy - describe a machine, get a 3D model you can take apart
Grok Bot - XAI's agents now have a dedicated email for independent actions. It can also search, read, and monitor X
Microsoft-Decision-1 - Microsoft's fastest decision-making model yet

WHAT I'M CONSUMING

Andrej Karpathy estimated that only about 5,000 people have internal access to the most advanced AI systems
Anthropic's Frontier Red Team believes small AI models may now be months away from being capable of self-replicating without access to a data center
Patrick Collison on agents in the economy
Muse Wonder: Zuckerberg and Wang's $100B AI
Character.AI files for bankruptcy, agrees to $20M Disney asset sale
You can just vibe-code Beehiiv
Amazon has over 1 million robot workers. What's next for human employees?
38 things we've learned about building AI products
The state of self-improvement AI