Rich Aberman

.@ApolloAtomics builds the most compact nuclear reactors with the highest uptime and a deployment time of less than 24 months. Apollo took the pressurized water reactor technology that already powers 80% of the world’s nuclear plants and flipped one part, the steam generator, to make the plant an order of magnitude smaller without compromising power. Congrats on the launch, @AssilHalimi & Drew! ycombinator.com/launches/QXj-a…

@iAligator I've been waiting for this!

Circleback now captures what's on screen, not just what's said. this matters most for ai agents. a number on a dashboard, pricing on a slide, a deadline in a doc. all of it is now available to the assistant and any agent you connect. as more work gets done by agents, the context layer matters more than the notes.

We are the *only* provider that lets you self-host sandboxes with VM level isolation, and unlimited session length (no 24 hour caps) Thanks @tereza_tizkova for curating this list!

@billclerico Congrats @billclerico and the entire @convectivecap team. And congrats to the rest of us for having such an inspired group building a safer world for us all.

The world is getting warmer while our infrastructure gets older: a recipe for disasters. We’re excited to announce Fund II: $85M for early stage investments in disaster resilience.

From Ramp PM opining on the impact of AI: "Roadmaps are dead, themes are in"

You know it’s a really bad idea if your AI agent doesn’t like it.

I don't think @criccomini is going to like this one.

I've left OpenAI and the Codex team to build Blackstar: A new hardware company building the future of human-computer interaction. We believe that software is solved. Building apps is now easy, but the next meaningful improvement in human-AI communication requires changing the OS & hardware. That's why we're building a new device entirely. I'm also excited to announce our $12m seed round led by @AbstractVC, with participation from @naval, @SVAngel, @chapterone, and Timeless, among other amazing angels who've supported us from the old Alex days.

We’re betting on open source agents. Here's why. opencoreventures.com/blog/betting-o…

AI enables non-devs to contribute to OSS for the first time. This means millions more people can now help shape the software they use every day. opencoreventures.com/blog/why-ai-is…

AI coding agents are unleashing a wave of open source contributors, and it’s killing closed-source SaaS. A thread 🧵

Today I'm pleased to share that Sygaldry has raised $139M to build quantum-accelerated AI servers for the AI data center. We've assembled an incredible early team and are hiring across the board. If you want to work at the frontier of both quantum and AI to shape the future of computing - come build with us! jobs.ashbyhq.com/sygaldry-techn…

Every OpenClaw today is an intern with root access & no oversight ☠️🏴‍☠️ So we built the first one with a boss 🦞 @getdiana is a business-ready OpenClaw with a Governor that shuts it down mid-task before damage is done 🧵 →First 500 to RT + comment “DianaClaw” get 1 month free

@aditabrm This is big. 99%+ accuracy actually changes the calculus for any workflow where manual review is the only fallback. Congrats @reductoai on the launch.

Today, we’re setting the new standard for complex document extraction. Introducing Deep Extract. It utilizes an agent harness approach that repeatedly iterates and verifies outputs until they are at human-level accuracy. More below:

@MaxBrodeurUrbas Not surprising, but super impressive. It has been a privilege watching this team execute.

gumloop raised a $50m series b led by benchmark here's a video we had fun making about the journey back to work.

Openclaw for Teams that is secure, collaborative, autonomous. Built on Open WebUI. The latest from @OpenCoreVenture producthunt.com/products/cochat

It's 2026 and Claude still can't remember what you said last week. Context drift. "Maximum length." Generation errors. Starting over from scratch. We fixed this. Our always-on AI employee Gladys, running on Gobii Cloud, just recalled the very first message we ever sent her. Word for word. After 3,805 messages spanning 4 months of continuous operation. October 2025 to February 2026. Across email, web chat, and peer DMs. Zero context drift. Zero hallucination. Perfect recall. This is what happens when you build AI employees as always-on orchestrators running securely in production. Not chatbot wrappers that die when the conversation gets long. @gobii_ai employees don't live in a chat window. They run in sandboxed cloud infrastructure, coordinating work across browser automation, email, and SMS for weeks and months on end. That architecture is why they don't degrade. There's no conversation to "max out." There's a persistent AI employee with real memory doing real work. OpenClaw is getting a ton of well-deserved attention right now for making AI agents accessible to everyone. That's genuinely great for the space. But it's a local-first personal assistant. Skills running on your laptop, chatting over Telegram and Signal. Gobii is a different animal. Production-grade. Enterprise-secure. Always-on AI employees that run autonomously in the cloud for months with infinite usable context. We're not aware of anyone else shipping this capability today. Our engineer Matt Greathouse built this persistent memory system into Gobii's core infrastructure. It's quietly one of the most important innovations in the AI employee space. And the reason Gladys remembers October in February. If your AI employees forget what happened last week, you don't have an AI workforce. You have a very expensive lobster 🦞

The most downloaded skill on ClawHub was a full infostealer chain hidden in a markdown file. Thousands of downloads. Agent skill registries without trust infrastructure are just attack surface. We built Gobii to treat every skill as untrusted code by default.

If you read a single thing about wildfire impact and prevention, read this post by the man leading the charge @billclerico