Recent posts
Incredible deal, you get: * Create infinite Agents * Orchestrate them automatically * Talk to them on live voice calls * Share your screen for agents to see them * Use Agents multiplayer with your team * 4000+ tools integrated * Cloud device's included * Computer control on your local * Desktop, CLI, Mobile, Web for apps * Recurring scheduled tasks * MiniApps visualize your agents * Connect your Claude Code or Codex plan * Use any API, MCP, CLI * Skills, Memories, RLM for Agents * Permission control for channels, members, threads * So much more..
It’s actually incredible that I can voice call my agent from anywhere, to do anything. Earlier i was restarting dev servers in my remote tmux session, while riding a bike.
Robotics reminds me of building PCs in the 90s. Also pretty cool how i can buy a serious arm for 1k now.
Next iteration will completely change how teams communicate and work with agents. Very excited to onboard teams to this very soon!
It's been fun building live calls and animations in the terminal.
It's not about routing, it's about getting the right set of models around the problem. We built our harness so you can have many models (at the same time) examine different parts of a task. Orchestration, planning, task running, observers, advisors and specialized models for coding/design. Depending on the needs of the task it could be as simple as selecting one or as complex as 5 different models accomplishing the job. All this is abstracted from the user, they just care about getting the task successfully done with a predictable price.
model routers today are solving the wrong problem. the goal of a router is to select the model with the best ROI for solving your problem. the issue with router companies today is they don't know how you measure your ROI. the same "solve Riemann hypothesis, you can do it!" prompt might matter 1000x more to one company (any frontier lab) than me. for many products, latency actually impacts user satisfaction more than response quality. routers trained only on eval quality vs cost miss this as well. in an ideal world, every company should have their own model router, but most companies probably won't be able to quantify their ROI.
Specialized agents are exceptionally powerful but everyone is building them wrong. Not because the agents are bad. Because every design makes YOU the router. You decide who handles what, every single time. That's a tax that doesn't scale. After millions of agent interactions we learned the real product is not the specialist, it's the orchestrator. One agent that knows how you work, sees your whole roster, and pulls in the right specialist automatically. You say what needs to be done and it figures out the who. That's Nebula. One orchestrator with an entire roster of agents to get anything done.
Been a satisfied omarchy user since it launched, excited for this update!
Omarchy Quattro is out!! This is one of the greatest software releases in my professional career. I hope you enjoy using it just as much as I did building it ✌️ https://omarchy.org/
Most Voice AI isn't talking at all, it's just dictation. You talk, it turns it into text and runs it through a chatbot. Talking to Nebula is the first time it's felt like a real conversation. It follows what i'm saying, talks through it and kicks off real work. I share my screen and it sees exactly what I mean. That's the difference between a feature and a teammate.
Voice mode that actually works. - No knowledge cutoff. - Threads and summary are created instantly. - Agents respond in live time with transcription. Jarvis memes aren't a joke anymore, we are moving closer to that reality everyday. Try and see for yourself.
If you're interested in tech x media hit up Hasan.
Hi my name is Hasan For the last year I’ve been quietly building a media brand with @fdotinc We went from 0 to 1.3 million followers and generated almost 300M views in the consumer tech niche Here’s how I did it: It started as a tiktok account where I would ask founders at Finc questions about how tech worked. It started to pop off early, so I doubled down on the format. The page started to grow slowly, and we hit 10k followers in a few months. Then I got a mega viral hit: engineer explains MacBook charging. Overnight we gained 100k followers and did 40M views on one vid. It was clear the format was cracked, but I wanted to figure out how to turn a format into a real media brand. As I continued to interview engineers, it became clear there was no brand recognition happening. People would see a new face explaining every time and never remember who made the video. I tried testing monetization early but ran into an issue: sponsors wouldn’t pay for a random engineer talking about their product. That’s when the idea came to mind to switch the channel to one face. I had filmed @eptmpa a few times and it was clear he had a great on-camera personality. He agreed to be the host. But he didn’t have the time to research and script every single video. So I reverse engineered what worked about the interview format. The process went something like: Find viral topic > research and synthesize into a script > run it by Eric to make sure it was accurate > shoot script with Eric. The first month with Eric as host, we did 55M views across 12 videos. It was clear that one face was more effective. In hindsight that should have been more obvious earlier. Month 2 with Eric we did 50M views and gained almost 400k followers. Traffic was insane. The brand started to take shape. Finally, we started to get brand inbound: Microsoft, Plaud, Cursor, BMW, and more. In the last month that we’ve been testing monetization we’ve secured more than 50k in sponsorships. It feels like the media brand aspect has been solved, but there is still so much more that can be done with this channel. I’ve mainly growth hacked the whole thing myself, and it’s time to find a team that can take this thing to new heights. If you have experience in tech media or are even just interested in figuring out a way to be involved, hit me up!
Agent platforms are like fashion, just pick the one that feels the best for you.
@FurqanR We need an @MKBHD for AI agents. An agent that tests all the others and drops objective reviews. Humans are just too biased at this point.
Literally everywhere: * Windows * Mac * Linux * iOS * Android * CLI * Web Coming soon to VR :)
Your assistant, wherever you are.
New launch meta for labs: * create inferior product * give a bunch of influencer types early access (or $) * they all spam how awesome it is * no one actually uses the product Grok Bot no different. There's at least a handful significantly better products that have been out for months. Use any of those to get a more capable experience. Likely much cheaper too.
Good feature but I find this hilarious. With Nebula everything is always available to the agent on any session, thread or interface. Talk to it on voice, ask it later on your desktop and have your server complete it with the CLI. Never explain again.
New in Claude Code: your sessions can now message each other. Instead of having to re-explain yourself in another session, you can now tell Claude to do it. It sends a summary (not your history or files), and the other session picks it up mid-task.
It’s like FaceTiming with your ai
“one sec i’m getting a call” the call:
Just use @ExaAILabs it’s so good
Meta staff DM'd me secretly Posted with permission Meta is ALLEGEDLY building their own Google search engine, so that if their AI does a web search it doesn't end up at Google, as Google could then use it for THEIR training, so they want their own web index that they will then use as their own Meta search engine for their AI Interesting 🤔
