Complex agentic workflows remain the next frontier for open-source models.
Context rot, infinite loops, hallucinated completions, MCP bloat, handoff loss, reward hacking, and fragile browser navigation are just problems that will be solved in the future.
I'm not worried.
Showing your pictures like this, greyed out, with a big scary warning, is right up there with Facebook using your friends and family faces on the "ARE YOU SURE YOU WANT TO LEAVE?" cancel screens of old. Insanely manipulative.
Jensen's return is not just in the best interest of NVIDIA, but the world. The board of directors best get moving on extending that invitation. Chop chop!
M4 MacBook Pro Omarchy update:
Now working and matching macOS:
- XDR is always on.
- True Tone adjusts the colour.
- Chromium hardware video encode is on.
- Low Power mode matches macOS.
- The speaker amps and headphone chip use macOS's settings, with 0 errors.
Sleep:
- Light sleep draws 0.98 W
- Deep sleep now gets all the way into sleep, but waking up resets the Mac 62 s later. A fix is ready for the next kernel I’m working on.
Beautiful, matching icon sets for Omarchy is coming. Drawn with svg, so they can easily be themed with matching colors. Open for you to install your own. Just like custom themes. https://t.co/m5gD0D1EVJ
The Derp State is out to ban open weight models, cartelize labs with frontier models, make sure no one can go off the grid. This is as dumb as it sounds, it can't work. But they will try, because they're stupid and evil.
That’s 17 PRs in less than 48h for Folio. Syntax highlighting, RAW preview support, faster search, search-as-you-type, support for all Omarchy opacities, cleaner chrome and sidebars, Markdown rendering, font previews, full vertical preview support, and a bunch of performance tweaks.
Everyone’s praying for Folio’s downfall. We’re just getting started.
A sad day for NVIDIA, developers, America, and all humanity. Jensen founded and ran one of the few companies to reach a massive scale while driving massive continued innovation and ubiquity, while others fell into the trap of ever-increasing extraction.
Well, that was fun.
As I said in the video three months ago, Flea was built to one day be the default file manager for Omarchy. Last night, Folio was merged as Omarchy's new default.
There's not much I can say other than thank you to everyone who supported the project. This has been one of the most wonderful experiences of my journey in tech so far.
I made all of this happen on a very tight budget. Meals were skipped, sleep was lost, and the journey took a real toll on my mental and physical health. I was solo the entire time, maintaining Flea and pushing the limits of what a file manager on Linux could be. We broke every benchmark we could find and hit unheard-of speeds. It was a generational and extremely fun run.
The good news is that Folio is C++ and Qt, so most of the front-end optimizations from Flea 0.3.12 can be ported over, which means it can still be the second-fastest file manager on Linux. I'll send as many PRs as I can so the file manager does justice to the wonderful project that Omarchy is.
The bad news is that I can't keep up this cash and mental burn for Flea now that it can never be the default. That was the main pitch to sponsors, and it's what drove me to push as hard as I did these last few months. I really tried my best to keep that pace going, but it isn't sustainable anymore.
Flea is not stopping. Starting with 0.4.3, it moves to a much slower pace. Bug and security fixes come first, and new features will land when time and budget allow. I had up to 16 more releases planned, and they will still happen, just over a much longer timeline. On performance, as of this writing there is simply nothing else out there that compares to Flea, and that is a win I will carry forever.
"Why don't you just open it up to other distros?" To be completely honest, I'm only interested in developing for Omarchy. I have no doubt it's the best the Linux desktop gets, and it's hard to develop for things you don't use. If I can't promise you something I'd use myself, there's no point.
If there's enough interest, I may pick the pace back up. For now I'm still grieving the default dream, so I'm taking it one release at a time.
Thank you to all 120+ contributors, and to superwhisper and zeroramp for sponsoring the project.
The sprint is over, but Flea isn't.
In June @NaderLikeLadder emailed Jensen about a crazy idea we had over dinner to make Local AI 10x better on NVIDIA hardware.
Jensen replied in 7 minutes. He saw this as important work, and approved a war room for us to work on this at NVIDIA HQ.
For a month, NVIDIA HQ was the @exolabs satellite office. We were overwhelmed by the support we got. We were back to back the first week with smart NVIDIANs coming in to our war room to help us. Creators of CUDA, Dynamo, people who built NVIDIA.
Long story short with NVIDIA we managed to improve their Local AI playbooks by 9.6x (almost the 10x goal we set out with). We’ll be publishing a whitepaper on this work.
We were just a small startup with a crazy idea. Jensen took a bet on us. Jensen cares deeply about Open Source AI and Local AI.
I don't believe Jensen is replaceable, he's a 1 of 1 human, but I hope NVIDIA will continue to aggressively support Open Source AI and Local AI, and continue to take bets on small startups with crazy ideas.
Thank you Jensen