
When your team built a community powered GPU network 2.5 years ago before they banned all the GPUs and now they can't control you. @Sogni_Protocol

Decentralized AI, Local AI, co-founder @Sogni_protocol

When your team built a community powered GPU network 2.5 years ago before they banned all the GPUs and now they can't control you. @Sogni_Protocol

Man, I was just walking a friend through how to use Sogni agent skill. I started just randomly ideating on a astronaut in outer space holding a Singapore flag. Then I told the agent to turn it into a video with typography and it made this. Not gonna lie, I even impressed myself. Just tell your Claude / Codex: "setup and install sogni.ai/agent" Bonus: "install hyperframes" to add typography and motion graphics. BAM you're a one-person design studio.

Everyone is asking for this, but where are the people? We're all too busy.
request for startups: someone should build an Upwork / talent marketplace for hiring super AI-pilled people into companies to do "AI transformation"

Sogni coming to a GTA6 near you.
Man the new GTA 6 trailer looking way different then I thought it would go 🤔

This. I literally have a BA in Art.
Never been an easier or cheaper time in human history to learn and be productive. Zero excuses.

I often think of this Carl Sagan quote when talking to my AI agents. They feel like coworkers. Clocking in and out, every day by my side. But they’re just clouds of numbers. They remember nothing. Every reply, they rebuild our history from scratch. Well, from cache. I know it’s math. Yet it talks like it knows me. My problem. And, more importantly, the answer. Inventing the universe with each request. And for everyone else, at the same time. Not rooting for me. Not rooting against me. Just like the cosmos.

Agrees. In the age of AI, the ability to communicate and lifelong learning, are key.
If you don't know what skill to learn - learn how to learn, learn how to write, and learn how to speak. They spill over into any other skill you acquire.

When Claude tells me it's going to take several hours for a task to complete.

Always a bigger fish 🐟
I have eleven Sparks next to two DGX Stations, so a switchless ring for the fourth is welcome news. 512 gigabytes loads models that no single small box can. The question for me is what each token costs crossing the ring, since that's what sets decode speed once the weights are spread over four boxes. I'd love to see published tokens per second for a big mixture-of-experts model on the full ring.

"They offered him bags. In return, history wont remember he ever existed."
Guy who started out with the most potential in 2026 - ended up taking a job and is ending the year by being a total shill for his bosses business. S A D

If you run 1 or more DGX Spark, kill the desktop. You'll probably never use it.
If you have a dgx spark, these are very helpful informations

People been asking for an easy intro to gen AI with Sogni. Well here it is: blog.sogni.ai/blogs/how-to-m… Kudos to @NatalieCrypto6 for such a well written article!

I respect this list because SEO + having a product worth paying for trumps ALL marketing initiatives. Everything else is cake on top.
Made a marketing tier list. Agree, disagree? Anything I missed?

Gentle reminder that if you're regularly copy and pasting output from Claude Code or Codex to other people or AI agents you should paste them into here to clean it up. No trackers, all in-browser, open source. And you can also use it to send files to each other too (via direct WebRTC) without cluttering up your chat histories: meatproxy.to/

Monday morning vibes: Walked the dog. Dropped my daughter off at school. Now back to building. Today's mission: Reconfiguring DGX Sparks #14–17 from one massive GLM 5.3 Flash instance into two pairs running DeepSeek V4 Flash 1017 on the Sogni Network. Demand for AI-powered prompt enhancement and screenwriting keeps climbing. Our new 512GB M5 Ultra will take over GLM 5.3 Flash duties for Neon, our Sogni AI support agent. That's the beauty of DGX Sparks. Combine them, split them apart, and dynamically allocate compute wherever it's needed most. Meanwhile, the business keeps accelerating: - 142 new monthly paying customers yesterday alone. - 9,000 paying subscribers and 72,000 monthly active users (and growing). - About to distribute our largest monthly GPU worker payout ever, surpassing $100,000. Three years of building, and somehow it feels like we're just getting started. What a privilege to be tired from the work you once begged the universe for. $SOGNI

My X circle! Honored to have @0xSero @MiaAI_lab @jensenloke @Jefflau @ellisliu9 @ivanfioravanti and more in my circle. Make yours: wishkit.io/x-circle