a simple example of how Recommendation Algorithms works, this 1y ago video is surfacing now after the events of sbta.
Algorithms don't show you the reality but reflects what we think reality is, and the more you think of negative/positive things the more it reflects back to you
I think there’s a shocking lack of imagination in a lot of the CEOs/influencers pushing open models rn.
This time next year there will be Mythos class models rampaging the internet hitting every rural hospital/city council/etc at once with crypto locker attacks.
Benjamin Netanyahu giving direct orders to Secretary of War Hegseth and Secretary of State Rubio today.
Aggressive finger pointing, serious faces, and “yessir” seen from Rubio and Hegseth.
Wow.
I worked through my own birthday party this year. I also worked through a Vegas trip while playing the World Series of Poker. And I still hit every KPI I set for myself.
I gave Codex a real task from my phone, put the phone down, and went back to the table. It spent the session building, testing and fixing while I played. I checked in between hands.
This is the first time in my life that having fun has not interfered with my work.
Coding nowadays is mostly communicating. If you can describe what you want clearly, the machine doing the typing does not have to be in front of you. I keep a machine running at home and connect to it remotely, so my phone is the remote control. I start the session, hand off the task, and the work happens on real hardware while I am somewhere else.
The second half is choosing the right work. An agent running unsupervised is great at building something new and self contained and terrible at careful surgery on a codebase you have to live with. So I save the greenfield stuff for trips. New project, clear goal, no half finished context in my head that I need to explain. Then I write the task like I am briefing someone competent who doesn't need to ask me follow up questions.
For me this replaces endless scrolling on Instagram. Instead of opening the app at the table, I open a session and start something.
Releasing the model weights and technical report of Kimi K3.
Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.
New model architecture: 2.5x the intelligence per unit of compute, not just more params.
Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.
Model weights: https://t.co/7m7eEg6Y0B
Tech report: https://t.co/yeu6cjpMCT
Tech blog: https://t.co/YTfiMSNM1f
Piers Morgan: “Iran is no threat to the US?”
John Mearsheimer: “No.”
Piers Morgan “They fund groups Israel doesn’t like”
John: “You’re baking the idea that Iran is a threat to Israel, and that any threat to Israel is a threat to the U.S. I don’t buy it.”
@YittaYasmineEl As you ve said, Sam never gaved a damn about security
They are using this for fearmongering to create their "mythos is so powerfull we can't give or to the public"moment