Pinned Tweet
Omarchy 4.5 RC1 is out.
The first big feature release since @omarchy Quattro brings:
- 50 languages
- Initial AArch64 support for Snapdragon X and NVIDIA N1X laptops
- Per-monitor workspaces
- Enhanced notifications (stack, snooze, history, inline actions)
- Floating workspaces with a dock
- Much faster boot (~1s)
- Video wallpapers and theme intros
- Quake-style agent console
- Gliff remote desktop over SSH
Plus speed-ups, better agent support, and more.
RC only for now. Test carefully. Final date not set yet.
"We can fix everything."
This has become one of the mottos of @Omarchy, and it represents a fundamental shift in how we think about our operating systems.
On Windows and macOS, we're used to countless papercuts, annoying bugs, and downright idiotic design decisions. Most of the time, there's nothing we can do about them. We find workarounds and eventually learn to live with the frustration.
Omarchy changes that. If something annoys you, FIX IT! Your OS should adapt to you, not the other way around.
And thanks to AI coding agents, you don't even need to be an experienced developer anymore. Anyone can make these changes.
Once you get used to fixing things instead of putting up with them, it's hard to go back.
If you see some cyber attacks which are highly publicized in the media, chances are the attacks were false flag operations by AI doomers.
This was always a possibility and it has become even more likely now.
LiteLLM 1.107.0 was poisoned on PyPI and Docker for about two hours yesterday. 🚨
The malware runs every time Python starts, even if you never import litellm. It steals cloud keys, SSH keys and your Cursor and Claude configs.
The attacker used stolen GitHub tokens, one of them an org owner's, to turn off branch protection and publish the release.
This is the second LiteLLM hijack this year. If you installed it Wednesday afternoon, rotate everything.
I love the new "World Clock" plugin in @Omarchy 4.5 😍
It's typically Omarchy: Comes with sane defaults that make sense for most people. In my case, it's perfect. I live in the Berlin time zone, but I often need to check the Los Angeles time!
Elegant and simple!
Grok Build by @SpaceXSI just shipped two updates! 🚀
v1.0.66 + v1.0.67:
• Worktrees inherit saved tool approvals
• Prompt suggestions now work with Vim mode
• Headless grok -p no longer hangs
• Faster exits for sessions with hooks
• @ file completion: 40 ms → 5 ms
• Theme picker fixes
Sacks with yet another based and nuanced take on SI risk:
There’s a world of difference between user-authorized agents (Muse, Instinct, Grokbot, etc.) and true unauthorized/rogue agents that could constitute cyberattacks.
Most of the activity described in the Axios story appears to be the former: customers sharing credentials with agents so they can take actions on their behalf. This may violate bank terms requiring humans to log in themselves, but these are in no way cyberattacks.
The direction banks should go is to support user-directed agents. For example, Amex is an authorized connection in Muse, making it easy for users to grant, limit, and revoke access.
Read-only permission levels are a sensible interim step. Several banks already support or are moving toward them. Full write access (with user credentials) is a different risk category and should be approached carefully.
Frightening the public by blurring the line between user-authorized agentic behavior and cyberattacks is irresponsible.
I'm regularly spreading positive vibes about AI/SI and technology here. They don't get much views, because you know: Fear sells.
I could easily drum up some scary stories, like certain other accounts are doing!
But I'm staying true to my beliefs and my e/acc roots.
Accelerate!
It's high time OpenAI got rid of the doomers in the company! Or else this will end badly.
How about this instead: Acceleration 2.0 🚀
Sam Altman says the company needs to figure out when AI becomes too powerful to safely develop or release.
Internally dubbed “pacing 2.0,” the effort would establish limits on AI systems that can continuously improve themselves.
Read more: thein.fo/3WCUYPR
Grok Bot is now available directly on 𝕏! 🚀
Your personal AI assistant now lives in your X Chat DMs. It can connect to your apps, handle tasks autonomously, and keep working even when you're away.
AI agents are becoming part of everyday life, and I love it!
If AI/SI becomes the most powerful tool ever created, restricting it to a handful of approved institutions would be an extraordinary surrender of power.
We want intelligence people can run, modify, and build on. Keep it open. e/acc.
Railway (@Railway) has been hacked 🚨
If you have used this service, rotate your keys immediately to prevent unauthorized usage. Better safe than sorry.
So @Railway was hacked!
ROTATE ALL YOUR API KEYS YOU HAVE IN RAILWAY RIGHT NOW
My @OpenRouter API key was drained!
It's only on railway, nowhere else
AI helped build this coding model. Now you can run it on your laptop. @Recursive_SI has released Flint, an open-weight model built from scratch by a small team of people and agents in under four months. It has 40 billion parameters, activates just 2.5 billion per token, and comes under Apache 2.0.
Recursive's Eureka Machine proposed ideas, implemented experiments and tested improvements across architecture, training recipes, quantization and hardware kernels. The company reports over 125 tokens per second on an M4 Pro MacBook Pro at short contexts, with leading coding results among similarly sized MoE models. Those figures are self-reported, but the weights are available for anyone to test.
Close your laptop. Claude Duet keeps working. Anthropic's new proactive agent runs on its own cloud computer, checks your connected apps every few hours and sends a morning brief. Powered by Opus 5.5, it can prepare meetings, chase unpaid invoices, build dashboards and flag problems without waiting for you to ask.
Duet carries over your Claude memory, skills and connected apps, and asks for approval before sending messages, making payments or deleting files. It's a separate app for iOS, Android and the web, starting with a waitlist for US Pro and Max subscribers. Background checks count toward your existing Claude usage limits.
DeepSeek has opened up another piece of its efficiency playbook. DeepInfer, the inference engine powering its production services, is now available under Apache 2.0. It supports DeepSeek V4.1-Flash on both Nvidia Blackwell GPUs and Huawei Ascend 950 chips.
The code includes optimized kernels, speculative decoding and scheduling that overlaps CPU work with chip execution. Developers can now inspect and adapt the machinery DeepSeek uses to serve models at scale. It still needs components such as KV-cache infrastructure and load balancing to become a complete hosting service, but the core engine is there to study, benchmark and build on.
PSA: The @Omarchy package server seems to have some problems right now. 🚨
Updates may fail with timeouts.
"Started production of @Tesla_Optimus on our high-volume lines in Fremont."
Optimus V3 is already in production! That means the official presentation must be pretty close.
I'm thinking maybe next month.
Q3 Shareholder Update → ir.tesla.com
Highlights
– FSD Supervised attach rate grew to over 75% of new vehicle deliveries in NA
– Unveiled the next-gen Roadster
– Record deliveries in Canada, Australia, Japan & Taiwan
– Tesla Semi now in volume production. Top variant can go up to 500 miles at a full 82,000 lbs while achieving efficiency of 1.7 kWh/mile – about a third of the energy a comparable diesel truck burns for the same load
– Ended the quarter with the largest backlog for our passenger vehicles in 4 years!
FSD Supervised
– 4 new approvals since Q2 (Croatia, Slovenia, Czech Republic, Slovakia)
– FSD is now live in nine EU countries & used by over 100,000 customers logging more than 160 million kilometers since initial approval
Robotaxi
– Deployed Cybercab into our Austin fleet
– Hit nearly 2 million cumulative unsupervised miles, 8x growth in the quarter
– Austin service now runs 6 am to midnight, with wait times continuing to come down as we add cars & improve the orchestration algorithms
– Added Grok to Cybercab with Grok Bot & Connectors, letting you interface with your inbox, calendar, files, chats and tasks, all by voice.
Transport, office & living room in one
Robotics
– Started production of @Tesla_Optimus on our high-volume lines in Fremont.
These are the first production-quality bots we've built, proving that the current manufacturing process for Optimus is suitable for a commercial product
SI Compute
– On track for 400 MW of training compute by year end, feeding both vehicle & humanoid autonomy
– Austin Fab construction continues, equipment commissioning planned for 2027
Battery
– Battery supply is still the limiter on how fast we ramp vehicles globally, so we're expanding on every front: pack capacity in Berlin, cathode and lithium refining in Texas, LFP cells in Nevada
– Shipped the first Megapacks built entirely with our own LFP cells
– Increasing 4680 output to support ramp of Cybercab, Semi & Model Y
Energy
– Megafactory Texas is up & running, with Megapack 3 production started
– Highest-ever order rate & backlog for energy storage
Grid constraints & power availability are making Megapack the answer for SI infrastructure demand: it excels at supplying large data center loads, smooths out demand spikes, helps new projects meet grid connection rules & makes the grid itself more stable
All with less local fossil fuel generation
Supporting Infrastructure
– Added 2,100+ net new Supercharger stalls, up 14% year over year
Grok Bot 0.71.2 just dropped! 🚀
What's new:
• Files up to 250 MB, videos up to 1 GB
• 20 attachments per message
• 200 plugin skills (up from 24!)
• Smarter desktop notifications
• Tagged email addresses for Bot routines
• Voice calls reconnect automatically
The pace is nuts!
Meta has paused sharing of Muse chats!
The security firm Claroc found more than 6,800 shared conversations that anyone could open without logging in. They found them through public DNS records.
Hitting "share" on an AI chat may be the same as publishing it.
Bye-bye, BioWare. In the end, it was just a shell of its former self, taken over by woke activists.
The real BioWare died long ago.
🫡
BREAKING: Electronic Arts is cutting 400 jobs and shutting down BioWare, the legendary development studio behind Dragon Age and Mass Effect. bloomberg.com/news/articles/…





















