Harsh Agrawal

Computer Use Agents @AnthropicAI | Previously Vercept, Apple MLR, PhD @Georgia Tech, CloudCV, EvalAI.

[Andrej Karpathy @karpathy]

This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November), peaking especially for long problem-solving sessions on very difficult problems. You can give it a lot more ambitious tasks than what you're used to, the model "gets it" and it will just go, and it's never felt this tempting to stop looking at the code at all (but don't do this in prod!). The model still has quirks that people will run into and the safeguards are configured to be a little too trigger happy for launch, which can hopefully be tuned over time.

I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing substantially. You can ask for anything - explainers, visualizers, dashboards, bespoke single-use apps (e.g. a full wandb that is hyper-specific just for your project), you can 10X your test suite, auto-optimize code, run giant research projects with custom HTML for the results, anything! "Free your mind" (Matrix ref). Really looking forward to all the things people build!

[Claude @claudeai]

Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.

Its capabilities exceed those of any model we’ve ever made generally available.

[Kiana Ehsani @ehsanik]

When we were deciding between joining @AnthropicAI or continuing to build independently at @Vercept_ai , we wrote down the pros and cons. One con everyone kept raising: acquisitions slow you down. You get dragged into bureaucracy. There's no way you ship as fast.

It's been less than four weeks since we joined, and with the team here behind us and joining forces, we just shipped our first product launch.

That speed comes down to the culture here. Everyone moves fast, everyone is incredibly smart, humble and supportive, and it's really easy to get things done. I'm very bullish on this: Anthropic's moat is its people.

Starting today, Cowork now supports full computer use, and Dispatch lets you control your machine remotely from anywhere.

Four weeks down, many more to come. We're just getting warmed up. 🔥

[Luca Weihs @LucaWeihs]

It's been 26 days since Vercept joined Anthropic ➡️

Claude is now able to use your mouse and keyboard to control your computer in Cowork and the Claude Code Desktop app 🖱️⌨️💻

I'm so excited for everyone to play with what we've been building!

[Felix Rieseberg @felixrieseberg]

Today, we’re releasing a feature that allows Claude to control your computer: Mouse, keyboard, and screen, giving it the ability to use any app.

I believe this is especially useful if used with Dispatch, which allows you to remotely control Claude on your computer while you’re away.

[Simon Taylor @sytaylor]

Can Anthropic just CHILL for like a day so we can process this stuff?

Shitting hell man

[Kiana Ehsani @ehsanik]

This is a long post, mainly because I have a lot to say, but in case you are too busy:

TLDR: @Vercept_ai is joining @AnthropicAI! We shared a mission, so we joined forces to accelerate it into reality. Couldn't be more excited!

Why Vercept was started

In 2024, AI coding tools were already becoming magical for developers, but other industries were ages behind. It felt insane that when my mom had IT issues, I still had to hop on a call and walk her through it step by step. Insane that sending a simple email took so many clicks.

That's why we started @Vercept_ai : Build something that acts for users instead of telling them how to do it. Two goals: 1) help people do tasks they didn't know how to do, and 2) handle the zero-brainpower tasks so people spend more time on creative work. As simple as scheduling meetings, as complex as reconciling messy financials before tax season. Ultimate goal was to have people spend less time behind screens and more time walking in nature. (Very Pacific Northwest mission 😁)

The ride

The journey of building an AI native company in this day and age was wild. Going from researcher to founder meant trading “reviewer number 2” for business partners and users, but surprisingly a lot of the same paradigms applied. Come up with a hypothesis, design an experiment, analyze user behavior, change the model and product based on the findings, wash, rinse and repeat.

There are some differences though. The pace and the adrenaline. Lows are low, highs are high. We were constantly being challenged and learned at a pace we had never learned before. NEVER! If you are an adrenaline junkie like we are, it's a blast. The joy of the startup adrenaline rush is truly underrated.

Why Anthropic

We raised more than $50M, had a comfortable runway and a successful product, were building full steam with a small team, and were truly enjoying every minute of it. But that's when the opportunity came to join forces with Anthropic. We already knew how great Anthropic was at building models and we admired their mission, but then we learned more about the vision.

We went on hours of walks, had long conversations, talked to members across different orgs, and learned more about Anthropic's vision and commitment to core beliefs which were very similar to ours. The more we talked, the more we realized we had been working on the same mission but from complementary perspectives. We realized that joining forces meant we could build something much much bigger together.

And beyond the mission, I am now a big believer that Anthropic's real moat isn't its best model. It's the people. Incredibly talented folks who genuinely care about mission and real impact over hype. A zero-ego culture obsessed with building something meaningful. The choices were clear: we could build independently and work toward the same vision as two separate versions of it, or join forces with an incredible team and accelerate that vision into reality. The decision became an easy choice.

What's next for our mission

Mission continues, just got a bigger stage and an expanded team. The goal is still to expand AI beyond just a chatbot, to enable non-technical users to leverage it just as much as technical ones. We're just getting started.

It takes a village

This journey wouldn't have happened without the people who made it what it was. First and foremost, my cofounders @LucaWeihs and @inkynumbers . Best people I could've wished for as cofounders. We never once got into an argument, always had communicative discussions and as a cherry on top shared the same sense of humor! I feel blessed and grateful to have these two in my life.

Thankful to our team for trusting in the three of us and showing up day and night.

Grateful for @sethbannon , our board member, lead investor, great mentor and the person whose energy is so infectious that whenever we were having a down moment we would say "channel your inner @fiftyyears energy!"

And to our wonderful investors and supporters: @chrija and @PointNineCap , Yifan and Jacob and @ai2incubator , and @mattmcilwain and Ted Kummert from @MadronaVentures . Couldn't have done this without you.

Onward 🐜

[Luca Weihs @LucaWeihs]

We've just announced that Vercept is joining Anthropic [1]. I’m beyond excited and wanted to provide some personal color on why now and why Anthropic.

There are many reasons why I chose to start Vercept with my co-founders. Some prosaic: I was ready for something new. Some personality-driven: I love building things with a focused group of brilliant people, working with my co-founders and everyone who joined us has been some of the most fun I've had in my life. And some more grand: AI is set to radically change the world and building a startup felt like one of the few ways to have a say in what this future will look like.

So what changed?

When we started this company, the best models at the time were just barely useful for coding. A year later, Claude writes code that is almost always better than my own. I thought we were at least another year or two away from an Opus 4.6 level model: I was wrong.

I don't think that the general public has caught up to these changes. When I speak to some of my brilliant friends who work outside of AI, they tell me about their middling experience with some copilot tools or how they have fun speaking with chatbots but they're too unreliable to be valuable beyond entertainment.

As someone who works with these models constantly, this attitude surprises me. Seeing this last year's progress, I can now easily imagine a future world in which Claude completes knowledge work tasks at the same level it codes today: better than myself, most of the time. What that really means is that how people work with their computers may fundamentally change, and getting that experience right matters as much as the underlying models. The benefits can be transformative [2], but only if we focus on the details: building AI that works alongside you on your actual computer, understands how you work, and acts transparently on your behalf.

When I look at the landscape of frontier AI labs, Anthropic stands above all others in its dedication to a mission that values human flourishing and a deep respect for its responsibilities [3]. I did not expect to feel this written mission echo through every conversation I've had with those at Anthropic; the culture is really something special. At this moment, I feel that Anthropic is the best steward of AI: I want it to succeed.

The above changed my internal calculus about how I can have the most leverage on the future to build safe and widely beneficial AI. With Vercept joining Anthropic, we have the opportunity to shape how the best AI models use computers to bring real value to the world. Vercept was on its way to the same goal but, with tremendous change coming in the near future, joining forces feels like the way to make the biggest difference when it matters most.

I am so excited to join forces to build a world where we keep thinking.

[Christoph Janz 🕊 @chrija]

Congrats, @ehsanik, @inkynumbers, @LucaWeihs and the entire @Vercept_ai team on joining Anthropic! And congrats, @AnthropicAI, on the amazing addition to an incredibly talented team.

It's been a huge pleasure working with the Vercept team and @sethbannon over the last year. I'd be lying if I didn't mention that this is a bittersweet moment – I would have loved to continue working with y'all on Vercept for the next 10 years! But if there's one company to join, it's Anthropic.

From a @PointNineCap perspective, we're thrilled to become a shareholder in what I believe is going to be THE company of the decade.

[Anthropic @AnthropicAI]

Anthropic has acquired @Vercept_ai to advance Claude’s computer use capabilities.

Read more: https://t.co/YVz1zh7Pqq

[Vercept @Vercept_ai]

Vercept is joining Anthropic.

When we started Vercept, we set out with a bold vision: to move beyond the mouse, keyboard, and touchscreen and build technology that truly expands human potential. ✨

That mission hasn't changed. It's found a bigger home.
We're thrilled to announce that the Vercept is joining Anthropic to help shape the future of AI that's safe, accessible, and built for everyone. 🎉

What drew us to @AnthropicAI is the alignment in our values: a shared belief that technology should be safe, accessible, and built to empower people, not exploit them. A belief that responsible development matters and that the future we're building should serve everyone, including communities that are too often left behind.

To our incredible team: this is a reflection of the culture you built, one rooted in transparency, collective accountability, and genuine care. You made this possible.

To the community that believed in our vision from the start: thank you. We're carrying that energy forward into this next chapter. 🙏

We couldn't be more excited about what's ahead. Let's build. 🚀

[Kiana Ehsani @ehsanik]

The FOMO when I’m not in the office is actually painful.

Yesterday was one of those days where every few hours someone went “do you guys wanna see what I built / how the new model is doing / these new benchmark numbers?” and we’d all gather around a desk for a demo. Not every day is like that. Some days are more like “I’m stuck on this, please help”, and “no clue why everything broke today, we gotta fix it”, and I kind of love both. It just sucks to miss those little “come see this” moments when I’m not physically there. At this point whenever any of us is out of office we have serious FOMO, or as @inkynumbers calls it, AMO: actually missing out.

Will be in the bay for 24 hours. Pingponging between MTV and SF. Text me if you’re around and wanna meet.

[Jared Isaacman @rookisaacman]

Well it is official now..

I want to again express my sincere appreciation to President Donald J. Trump @POTUS for nominating me to lead NASA, and to the United States Senate--and Chairman Cruz @tedcruz - for their diligence and fairness throughout the confirmation process. I am grateful to Secretary Duffy @SecDuffy for his leadership as Acting Administrator during this transition, and to my wife Monica, my family, my friend Senator Sheehy @TimSheehyMT and everyone who offered their support along the way.

As I step into this role, I make these personal commitments:

– Mission: I will champion the bold objectives of human space exploration, scientific discovery, and a thriving orbital economy that ensures America’s leadership in space. We will never again give up our capabilities to reach for the stars, and we will never settle for second place.

– Integrity: I will serve responsibly, transparently, and without personal gain, covering every cost I am legally permitted to, and fully adhering to my ethics agreement. My loyalty is to my country, my President, and the space agency that has inspired me since I was a child.

– Urgency: I will intensely focus the agency on achieving the near-impossible, the very reason NASA was established in the first place. We will eliminate the bureaucracy that impedes progress and empower the best and brightest to take ownership, move quickly, accept smart risks, and act with a relentless focus on mission success.

– Inspiration: Every launch, every scientific breakthrough must inspire the next generation to dream bigger, to reach higher, and believe that anything is possible. In addition to my existing philanthropic efforts, I will donate my salary as Administrator to the U.S. Space & Rocket Center’s Space Camp to help prepare the pioneers of tomorrow.

I am humbled by this opportunity, proud to serve, and ready to work alongside the most talented minds in America as we continue the greatest adventure in human history.

Sincerely,
Jared Isaacman
NASA Administrator

[Harsh Agrawal @harsh_092]

The team worked incredibly hard to bring the cross-platform app to life. This is just the beginning. With this core in place, we’re gearing up to launch a wave of exciting new features in the coming weeks.

What an incredible first few weeks at Vercept.

[Vercept @Vercept_ai]

We have been heads down on something for the past 5 months (which is a very long time in startup time). Today we are finally sharing it.

Vy, our AI agent that uses your computer for you, is now cross platform.

It runs on both Mac and Windows, sees your screen, moves the mouse, types, scrolls, and clicks through apps. No APIs or custom integrations needed.

[Harsh Agrawal @harsh_092]

@AriX @ChatGPTapp Congrats Ari!! Very exciting news!

[Harsh Agrawal @harsh_092]

Excited for everyone to try this out!! The magic of Vy is learning how it thinks. Build that “theory of mind,” and you can stop babysitting—Vy will run for hundreds of steps and get things done!

[Vercept @Vercept_ai]

🚀 Big day today! We’re teaming up with @_ai_collective + @MadronaVentures (and the amazing organizers 👏) to kick off the #IASummit by demoing Vy 🎉

For the first time ever, we’re unveiling our new cross-platform app (Windows + Mac) 💻✨! Until now, only a select few had early access. Still pre-release, but today we’re opening a small waitlist!

The event is fully booked with a long waitlist 🔥 but if you’d like a chance to join, hop on the list now! Link 👇

[Harsh Agrawal @harsh_092]

@jasonwuishere @PurdueCS Amazing news!! Congrats!!

[Harsh Agrawal @harsh_092]

@deviparikh Yeah, exactly! And if the paper is printed on both sides, the way different designs interact could be quite fun.

[Harsh Agrawal @harsh_092]

@deviparikh Tutorial Link: https://t.co/yE1iZrvHxM

[Harsh Agrawal @harsh_092]

@deviparikh I couldn't find pictures of the final thing, but I once made origami fireworks to introduce origami to a bunch of school kids. I remember one kid thought it resembled a kaleidoscope.

I do have pictures of an intermediate attempt though.