Your curated collection of saved posts and media
@jxnlco @ajambrosino https://t.co/HrDuT6pbM5
@jxnlco https://t.co/PrbH2B41f0
@jxnlco https://t.co/PrbH2B41f0
@jxnlco https://t.co/XGWaXwzkGP
@jxnlco https://t.co/XGWaXwzkGP
BREAKING: Starlink will bring live sports streaming to Virgin Atlantic flights. β’ The service will be free through Virgin Atlanticβs high-speed Starlink WiFi. β’ FIFA World Cup 2026 matches will be available live during flights. β’ Passengers can watch live sports on personal devices. β’ Starlink is already installed on around one-third of Virgin Atlantic aircraft. β’ Virgin Atlantic plans to complete Starlink installation across its fleet by 2027 Another example of how Starlink is transforming the passenger experience at 35,000 feet.
@vincent_koc As you can see me repeating our number 1 rule, earlier today https://t.co/VPwRRirPnU
@vincent_koc And you can see right now that our mods became aware of this, he too is banned: https://t.co/B8Z6zHSHGU
Slowly figuring out ports and such, although it'll be easier with the Amazon order arriving tomorrow π TIL micropipettes make great CA glue applicators. Also TIL cover slips are fragile ππ·πΏ https://t.co/UhZ1vjMabB

https://t.co/cD8D44aGFh
@jxnlco https://t.co/I0gqekEvmO
@jxnlco https://t.co/I0gqekEvmO
https://t.co/PfoFRerWtk
https://t.co/cD8D44aGFh
https://t.co/PfoFRerWtk
The hardest part of VFX used to be making it look like it belongs. That That bar just got cleared. What used to require tracked plates, manual roto, and a lighting team now starts with a ComfyUI workflow and an auto-mask. Creator seungho__yeo ran ComfyUI + Beeble AI Canvas to blend AI-generated objects directly into live video: β Objects generated in ComfyUI, dropped into real footage without breaking the scene β Auto-masking that holds the original depth, lighting, and spatial feel β Per-mask motion control β follow the camera or move independently β Generate β Mask β Composite β Done The dolphin composite alone shows what's possible: different mask regions, different motion behaviors, one coherent shot. This is what VFX looks like when the pipeline fits in a single creator's hands.
my codex billboard submission https://t.co/dQXQjz1ovf
my codex billboard submission https://t.co/dQXQjz1ovf
Yay, I'm one of them! ππ€ Let's make the best AI agent even better together! https://t.co/0bN7beC2XJ
Seedance 2.0 in native 4K. Skin texture, glass rendering, directional lighting, and a real micro-expression. The kind of frame professional filmmakers would cut into a production. Half our team are professional filmmakers. This is their verdict. https://t.co/jEkyWnSJMV
Olivia Rodrigo on her Daisy Chain Fields festival! βAll of the artists who are doing it are making no profit. I just feel like we need something really positive to do and see, and young girls need awesome role models who are supporting other women and who are engaging in something thatβs really joyful and musical and community-oriented.β

Update - a TLDR video π https://t.co/51YKG6ErmG
Sakana Fugu Technical Report https://t.co/6e6WuA8FVB Release Notes: https://t.co/7xWGpOicFN https://t.co/g2yaZvex35
Learning to Control Self-Assembling Morphologies: A Study of Generalization via Modularity They co-evolve a bunch of limbs to self-assemble into a body to perform a given task. They show more limbs generalize to some extent. Fun work by @pathak2206 et al! https://t.co/jWXXhD5wyy https://t.co/AZhkSuw31N
@pathak2206 Thanks a lot, Deepak!! One day, we'll have self-assembling multi-agent LLM systems like your work: https://t.co/qhi40ctLaa
Learning to Control Self-Assembling Morphologies: A Study of Generalization via Modularity They co-evolve a bunch of limbs to self-assemble into a body to perform a given task. They show more limbs generalize to some extent. Fun work by @pathak2206 et al! https://t.co/jWXXhD5wyy
The thing that made Fable so impressive was its creative problem-solving and good judgement calls across long-running projects You can see this when I had it make a self-aware Snake game. I gave it no design feedback, just "make it better" Worth trying: https://t.co/KOIp52aGzg
Snake games are a bad test of AI beca- "Claude 3.7, make a snake game, but the snake is self-aware it is in a game and trying to escape and interesting things happen as a result" This is all AI (one prompt + a request to make special things happen faster). Matrix mode at 0:55 https://t.co/1ZtlC1wIVB
Here's Sonnet 3.7 doing the same thing a year ago. https://t.co/ahssk7uYgB
Snake games are a bad test of AI beca- "Claude 3.7, make a snake game, but the snake is self-aware it is in a game and trying to escape and interesting things happen as a result" This is all AI (one prompt + a request to make special things happen faster). Matrix mode at 0:55 h
This is one of the most absurd things I've read. I didn't think people could even had such a mindset! How can you seriously be against saving people's lives?! I've worked in medical AI for the past 8 years. Medical AI is one of the few applications even many AI haters I've talked to are excited by. They would definitely welcome a cure to cancer invented by AI. Seeing people want to slow this down is a first for me. Let me actually talk about the article itself. The thing is the author actually doesn't believe AI can cure cancer. So the logic of this whole article is kinda nonsensical. On top of that, the argument for why the author doesn't believe AI can cure cancer is also poorly constructed. Basically the author's argument is the data does not exist yet for AI to cure cancer. I wish the author would consider that there could potentially be novel creative ways to collect relevant data to feed to AI. The author says "There are, in short, many barriers to curing cancer beyond a lack of intelligence." I would respond: there are, in short, many solutions to these barriers, if only we refuse to be limited by a lack of imagination. The author then points to the chaotic release of Fable 5 as a reason for why AI needs to be slowed down which is so odd to me, I don't even know how to respond. Fundamentally this feels like a very zero-sum view of AI progress. Just because there are important questions and challenges that need to be addressed by society as a whole doesn't mean using AI to cure cancer should be slowed down! The article also says a variety of other odd things. For example the author says "I would neither spend months struggling with a research problem I knew AI could solve instantly, nor find as much pleasure in the answers it provided." So if AI found the cure to cancer, the author wouldn't find pleasure in it?! Idk I have a fundamentally different viewpoint about research. I want to know the truth no matter what tool finds it. The author then ends about how writing is a process of self-discovery and human connection and how her wife consoled her as she reflected on these things, but like all of that can happen with AI too. You don't need to use AI to write if you don't want to! You don't need to replace your wife with an AI therapist! Once again, a very zero-sum view of technology. I find this whole article just absolutely revolting. We should not be denigrating the use of AI to cure diseases and save lives, which I personally believe is one of the most important missions on this planet. ------ As a side note, I will clarify I don't actually think AI is going to find a singular magic drug to fully cure cancer in everyone with a single prompt to GPT-7 or whatever. I expect a cure to cancer looks more like preventative health, early diagnostics, and personalized medicine all powered primarily with advances and analyses by frontier AI systems.
everybody loves computer https://t.co/IKP1SAliHC
Our mission is to benefit all humanity https://t.co/Qwwu6Eat0O

everybody loves computer https://t.co/IKP1SAliHC

it's official, I'm on community at NousResearch. I'll be shipping tested walkthroughs, answering questions and carrying what the community builds back to the team. if you're on Hermes Agent, the Discord is where we all are, come through. https://t.co/584RcsJ8Sm
The new /goal command in Grok Build is a huge update Until now, most coding agents have worked like enhanced chatbots You ask. It responds. You review. You guide it. Repeat /goal changes the entire paradigm Instead of micromanaging every step, you hand Grok Build a single engineering objective, and it plans the work, breaks it into tasks, executes them, verifies the results, and keeps going until the goal is complete It can: β’ Create an execution plan β’ Track progress with a checklist β’ Review code β’ Inspect webpages β’ Execute scripts β’ Verify its own work β’ Recover from failures β’ Continue autonomously while you supervise You can monitor and control it with commands like: /goal status /goal pause /goal resume /goal clear This is a shift from βhelp me write codeβ to βown this engineering objectiveβ Thatβs a completely different category of coding agent...and a glimpse of where AI-assisted software development is headed
Introducing /goal in Grok Build. Execute long-running tasks autonomously, with multiple rounds of subagents implementing and verifying a single goal. https://t.co/D36aiwuXS5