- AI Valley
- Posts
- Anthropic just gave Claude your mouse and keyboard
Anthropic just gave Claude your mouse and keyboard
PLUS: “We’ve achieved AGI”
Together with
howdy, it’s Barsee again.
happy tuesday, AI family, and welcome back to AI Valley.
we’re keeping things tighter and clearer today.
here are the biggest things worth knowing today:
Anthropic just gave Claude your mouse and keyboard
“We’ve achieved AGI”
Elon Musk’s AI chip factory could rival a city
Plus trending AI tools, posts, and resources
Let’s dive into the Valley of AI…
VIKTOR
Viktor is an AI coworker that lives in your Slack. Not a chatbot. Not another tab. A colleague that connects to 3,000+ tools and ships real work.
One message: "Pull last week's revenue from Stripe and ad spend from Google Ads, compare them, and post a summary to #growth." Done. A formatted report lands in your channel.
It doesn't stop at data pulls. Viktor builds and deploys web apps, generates board-ready PDFs, manages Google Ads campaigns end-to-end, and runs scheduled tasks 24/7. Your Monday morning briefing? Automated. Weekly ad performance reports? Running before you open your laptop.
3,000+ integrations. SOC 2 certified. Your data never trains models. 84% of teams who add Viktor never stop using it.
Give Viktor your worst task. $100 in credits
*This is sponsored
THROUGH THE VALLEY
1/ Anthropic just gave Claude your mouse and keyboard
On Tuesday, Anthropic rolled out a new feature that lets Claude control your computer. It can click, type, scroll, open apps, and move across your system on its own.
What’s happening?
Claude can now interact with your computer directly, not just through APIs or integrations.
It can access files, use your browser, and work inside tools like Slack or Google Calendar. And when something isn’t supported, it doesn’t stop there. It just uses the screen itself, like a human would.
Why does this matter?
Most AI tools break the moment an integration is missing.
Claude doesn’t.
Instead of waiting for native support, it works through the interface itself. That makes it far more flexible in real-world use.
Who’s in trouble?
Tools like OpenClaw.
They built their entire product around autonomous agents. Anthropic is adding the same capability directly into Claude.
That means users don’t need to switch. They just get the upgrade.
What else is new?
It connects with Dispatch, so you can assign tasks from your phone and have Claude complete them on your computer while you’re away.
Any risks?
Yes.
Anthropic says it has added safeguards against things like prompt injection, but the feature is still early and not meant for sensitive data yet.
What’s the real story?
Anthropic is quietly shifting Claude from an assistant into an operator.
And once that works reliably on desktop, moving to mobile is the obvious next step.
And some sources say that they are actively working on a Phone Use feature, so Claude will be able to make calls and execute tasks on your mobile device.
2/ “We’ve achieved AGI”
On a recent podcast, Nvidia CEO Jensen Huang said, “I think we’ve achieved AGI.”
What does he mean?
Not the sci-fi version of AGI.
He’s talking about how AI is already being used today. People are building systems that can automate tasks, write code, run workflows, and even manage small projects.
In that sense, AI is starting to handle real work, not just answer questions.
So is it actually AGI?
Not really.
The definition of AGI is still unclear, and today’s systems are far from reliable across everything. They work well in some areas, but still fail in others.
What’s the catch?
Most of these AI systems don’t last.
People experiment with them, but very few turn into stable products or businesses. It’s easy to build something that works once, much harder to make it consistent.
What’s the takeaway?
This isn’t a clear “AGI is here” moment.
But it does show how close AI is getting to doing useful, independent work, even if it’s not fully there yet.
3/ Elon Musk’s AI chip factory could rival a city
Elon Musk says Tesla and SpaceX plan to build a massive chip factory called “Terafab” in Texas.
The scale being discussed is hard to wrap your head around.
What’s being claimed?
Terafab is expected to produce up to 1 terawatt of compute annually, which Musk says is roughly 2× the current US capacity.
Some projections around it go even further:
Targeting output comparable to a large share of global chip production
Chips designed to power Optimus robots and AI satellites
A facility as large as a major city (~100M sq ft)
Heavy use of space-based compute, powered by solar energy
Not all of this is confirmed, but it shows the direction Musk is aiming for.
Why is this such a big deal?
Right now, almost every major AI company depends on a few chip manufacturers like TSMC.
That creates a bottleneck.
If Tesla and SpaceX can build even part of this capacity themselves, they stop competing for supply and start controlling it.
What’s the real bet here?
Not just better AI.
More compute than anyone else.
Musk is betting that future AI systems, robots, and even satellites will require orders of magnitude more compute than what exists today.
And instead of waiting for suppliers to catch up, he wants to build it himself.
What about the space angle?
Part of the vision includes running AI workloads on satellites.
In theory, space-based systems can use constant solar power and operate without many of the limits data centers face on Earth.
It sounds extreme, but it fits SpaceX’s long-term direction.
Is this realistic?
Parts of it, yes.
All of it, unclear.
The scale being discussed is far beyond what any single company has built before. Even hitting a fraction of these targets would be significant.
TRENDING TOOLS
Layers.com > Turn your Code into Customers. Full-stack agentic marketing automation for apps, SaaS and ecom (sponsored)
Luma AI Uni-1 > A new image model that understands text and visuals in one system, reasoning through tasks before and during creation
PlayerZero > An engineering world model that puts debugging, fixing, and testing your code on autopilot
Dimension > Connects to your work apps and handles briefings, meeting prep, emails, and action items while you sleep
Gamma > Introduces AI-native templates that let you update any presentation with prompts while keeping the original style intact
Dreamer > A platform to discover, build, and explore agentic apps
WHAT I'M CONSUMING
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.
HOW WAS TODAY'S NEWSLETTER |
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”.





