Arjun Karanam
@QuantumArjun
Followers
811
Following
6K
Media
7
Statuses
192
research @ , @StanfordHAI | interested in shaping embodied and collective intelligence | @stanford
SF
Joined September 2014
ive always held that full immersion is sliiiightly overrated - any screen can feel immersive if you're locked in enough - sent from my iphone while laying in bed
Today I'm releasing PortalVR 2.0, which lets you play the entire SteamVR ecosystem without wearing a headset. You plug in your Quest, put it on your desk, and play on your monitor. It supports 3D glasses and even 3D monitors! 😎 Get it now: https://t.co/dXqsCjllwt Thread 🧵
0
0
4
The streets feel like GTA V SF edition
A large power outage is impacting San Francisco - only call 9-1-1 for life safety emergencies, avoid non-essential travel, treat down traffic signals as four way stops, keep refrigerator and freezer doors closed, and turn off major appliances to prevent surges. Never use gas
0
0
3
My hot take is that continual learning is as much a product problem as it is an algorithm problem You can't learn from what you can't see. And measuring "did this actually help?" requires long-horizon signals that most model calls never get The hard parts: - Thumbs up ≠ the
Google is taking Continual Learning very seriously, having recently published papers about it (Titans + MIRAS, Nested Learning). This makes me think it's likely that @GoogleDeepMind will be be the first to develop and deploy Continual Learning AI, giving them a major advantage.
0
2
22
Amongst the hype beast-y tweets, a good technical recap from Michael
Here are some research directions I enjoyed in #neurips (will compile some more soon!) Bootstrapping long‑horizon reasoning: Recent work [1, 2] shows we can train LLMs on short-step problems and curriculum them into much longer chains. By composing simple problems into
1
0
6
Imo calling theater 'outdated' misses something fundamental: we don't gather in dark rooms to watch stories because it's efficient. We do it because experiencing a story together creates meaning that watching alone just can't. The collective gasp, the room holding its breath -
When asked if Netflix is destroying Hollywood, co-CEO Ted Sarandos said “No, we’re saving Hollywood.” “We’re in a period of transition. Folks grew up thinking, ‘I want to make movies on a gigantic screen and have strangers watch them [and to have them] play in the theater for
1
0
2
Probably the most talent dense place I’ve worked, and after launch, it was so obvious some of the folks were itching to go build their own thing
Thinking about it, ex Apple’s Vision Pro department people are out there building and shipping amazing stuff e.g. @benhylak at @raindrop_ai , @gravicle and @albtaiuti at @LumaLabsAI.
0
0
4
it's that time of year again! Lmk if you'll be at neurips, would love to chat about all things RL, continual learning, and the future of how we build software (...I've also spent the past 2 months content farming, so can just share some fun stories)
0
0
8
We’re backkkkk - apply now!!
Announcing: Rabbitholeathon 6.0 We invite you to a weekend retreat dedicated to curiosity, exploration, and diving down rabbitholes with other ambitious thinkers. Oct 24-27, 2025
0
0
7
At WWDC we introduce a new generation of LLMs developed to enhance the Apple Intelligence features. We also introduce the new Foundation Models framework, which gives app developers direct access to the on-device foundation language model. https://t.co/SnjCXrIyYj
machinelearning.apple.com
With Apple Intelligence, we're integrating powerful generative AI right into the apps and experiences people use every day, all while…
89
108
490
love to see cool ways of visualizing what actually goes on in government (it took me way too long to understand SF’s mayor vs board of supervisors distinction)
Government is going to change in 2025 and good tools can help us do it right. Introducing v2 of the SF Government Graph - a complete map of San Francisco’s government 🌉📈
1
0
7
🚀🚀🚀Want to develop a cutting-edge video generation model towards Sora? Please dive into Apple’s latest recipe and studies for scalable video generation models🔥🔥🔥. In this work, we aim at providing a transparent and detailed recipe 📖 for model architecture, training
9
46
130
rabbitholeathon never misses, so grateful that i still get to be involved, and learn so much every time
no matter how many times it happens im still in awe of the magic of rabbitholeathon each time. 5th retreat in and holy curiosity is contagious, it’s truly the best being surrounded by ppl who inspire you keep learning just cause
0
0
8
Obligatory “I’ll be presenting at NeurIPS next week” tweet! First ML conference and i’m so so excited - DM if you’ll also be there, would love to catch up!
1
0
13
can't wait for the inevitable LLM reasoning paper titled "Concept of a Plan"
0
0
4
The Venetian method for electing a Doge was literally insane, or genius
171
1K
11K