AI @planet | geo | space | optimism | jhana dabbler | qualia surfer | latent space explorer | e/acc with care | here for vibes & AI, views are my own

Based in United States
Replying to @BrianRoemmele
Looks a lot like the Belousov-Zhabotinsky reaction
28
1,682
Continuous Space-Time Video Super-Resolution V³ reimagines video as a continuous 3D Fourier field, not pixels in frames. Huge potential for video enhancement and generation. 📄 arxiv.org/abs/2509.26325 ✨ v3vsr.github.io
1
8
245
Replying to @patloeber
Bananas
623
Ever wanted to jump into a satellite image? Here are some Nano Banana perspectives on maps 🧵
5
1
10
2,039
Replying to @RichardMCNgo
Not unlike org charts
1
11
664
AI and crows: "they can learn how to use tools, they can use tools to make tools"
1
2
4,921
Are you manifesting new software from your phone yet?
The idea of writing software on your phone would have been comical 2 years ago. Now with AI Agents this is a reasonable form factor. Because the interaction model of AI Agents is heavily task oriented, you’ll be able to do almost any type of complex work from anywhere.
210
You can count r in strawberry but can you count objects in satellite images… Anyone with o3-pro access wanna try this? Regular o3 is off by one here. Link to source image in thread 🧵
sumovabitch, they did it they really really did it
1
2
218
Replying to @petergyang
Is this a ladder? (I thing we’re more likely to have more of a donut hollowing/flattening than a full bottom dropout FWIW) intelligence-curse.ai/
1
9
2,389
Gotta love those SF street flyer GPUs
4
122
Writing AI code with my AI editor on my AI car ride with some EDM in the background
2
6
160
Classic purchasers vs. users
6
Letting go of unneeded legacy constraints. Calling in expansive freedom, authenticity, and creative potential 🦋
100
Replying to @ciguleva
Here’s a Grok-designed Grok logo
1
74
Replying to @Rainmaker1973
Reminds me a little of the tobacco mosaic virus self-assembled tube structure
1
173
Replying to @SawyerMerritt
Grok’s new design
2
77
VAR you ready? Visual Autoregressive Modeling just won best NeurIPS paper
1
518
One possible architecture
Just saw the first actual example of this in production in a 1,000+ people company, and it's amazing They built an AI agent that: 1/ makes phone calls to everyone in the co every week, asking for a status update 2/ sends a report back with a summary of all the updates 3/ has a convo so the CEO can double click on any issue Basically replaces the entire middle management layer of the business — no human needed to distill information up and down, with all the biases & politics & slowness involved in that. Just ICs, a CEO, and AI in the middle Also means that as an IC you don't have to spend hours preparing for your big "meeting with the CEO." You can just ramble on, add as much info as you want, and trust that the agent will summarize and focus on what the CEO needs We're adding the ability for Lindy to make phone calls, so this will soon come in a Lindy near you
1
127
7-day lead time for floor forecasting “Before all data is fed into the LSTM network, the outputs of the different embedding networks are combined given the data availability, which makes the model robust to individual weather products that might be missing in the operational setting.”
Floods are among the most widespread natural disasters. Here we describe our new flood forecasting model for state-of-the-art 7-day lead time prediction, expanded forecasts on ungauged watersheds for expert use, and a new reanalysis & re-forecast dataset. goo.gle/4hHDS9O
1
214
Replying to @RhysLindmark
It’s definitely eye-catching (especially because of the title IMO). Here’s my attempt to visualize “playing the kernel trick on a multipolar trap” with ChatGPT
1
16
As our LLM friends get to know us, they can offer interesting new tools for introspection and reflection. Ask one of yours to make a representative image of your life.
1
1
116
3D Vision Transformer for field boundary delineation from time series satellite imagery Uses spatiotemporal correlations to improve accuracy and robustness 📄paper: arxiv.org/abs/2409.13568 💻code: github.com/feevos/tfcl 🎵suno song: suno.com/song/3ff72217-2a85-…
2
29
151
9,041
Soft Cells in Nature
Using centuries-old mathematics, researchers have discovered a new kind of shape that describes forms commonly found in the natural world. go.nature.com/3XSlKBZ
1
198
Map with the original prompt from @binarybits modified with the suffix parts from this thread (and using o1-mini)
The prompt can matter a lot here. Got a valid but not quite optimal route from o1 mini with some additional guidance 🧵
1
103
Final results
1
66
Founder mode move from Andy Jassy at Amazon. Flat, empowered, high-agency teal organizations are the future.
Two big updates from Andy Jassy today: - Taking a buzzsaw to middle management - Return to Office 5 days a week $AMZN has been dead money for 4 years. Time to crack some skulls
1
1
279
Replying to @anpaure
Don’t forget the other kind of generalist
40
genReality
the Mechanistic Universe is a lie - everything is diffusion. atoms, electrical charges, the strong and weak forces - just heuristics (generated lazily) with high enough resolution that we mostly don’t notice.
107
These are great questions: it seems difficult to know from an image alone if it is generated or manipulated. Here is ChatGPT’s response to the prompt “Realistic satellite image of a volcano in Iceland, Sentinel resolution and visual colors”
1
74
Beautiful views of Transporter-11
1
2
4
374
X isn’t dead, it’s just multidimensional
1
2
183
Replying to @doganuraldesign
Ambivert with community with AI with AI with AI
2
301
Focus on your AI fingers. Are they really there?
1
17
Replying to @bo_wangbo
Mind the (modality) gap proceedings.neurips.cc/paper…
3
298
I see so many connections to Shlain’s Alphabet Versus the Goddess here piped.video/watch?v=2QQuD62R…
1
1
5
6,238
Vision-language geo-foundation models (VLGFM) survey Paper: arxiv.org/abs/2406.09385 Ever-growing repo: github.com/zytx121/Awesome-V…
New discovery: Towards Vision-Language Geo-Foundation Models: A Survey github.com/zytx121/Awesome-V…
1
588
Postgres party 🎉
4
43
Remote Sensing Change Detection with CDMamba is here for the SSM wave arxiv.org/abs/2406.04207
127
Replying to @JustJamieJoyce
“nothing more whole than a broken heart” SoundCloud: on.soundcloud.com/Lb6FvWXgrd… Spotify: open.spotify.com/track/74IGL… YouTube: piped.video/watch?v=LDo4390h…
1
3
65
Replying to @nickcammarata
Higgs potential
2
67
Palette name candidates: off-Stanford, Palo Alt identity.stanford.edu/design…
1
1,000
PEARL: Platonic Embeddings Align Representation Learning phillipi.github.io/prh/#evid…
1
4
226
Excellent vision for the future of Earth and climate foundation models from the trailblazing EO AI lab of @xiaoxiang_zhu multi-modal, multi-sensor, multi-task
🚀 Perspectives on Earth and Climate Foundation Models! 🌍 Check out our new paper, "On the Foundations of Earth and Climate #FoundationModels," where we define 11 essential features to guide research towards the ideal FM. 👉 Read it here: lnkd.in/dgujQHkj
2
448
Hey Xosphere, in downtown San Francisco there are two flags on top of a building: the German flag and another that says “Metropolis” - anyone know why? Is it a reference to the film? I think I remember an American flag being there before.
1
1
324
Legendary Affective Computing trailblazer @RosalindPicard was featured on the @lexfridman podcast in 2019 piped.video/kq0VO1FqE6I
Here's my conversation with @RosalindPicard, professor at MIT & founder of Affective Computing. Teaching AI systems to better understand human beings is a passion for me, so talking with Roz was an honor & a pleasure. She is a true visionary in the field. piped.video/watch?v=kq0VO1Fq…
1
85
Replying to @djbaskin
2024: cutecut flairstyle
1
104