Chatgpt 4o can now generate full wine glasses pretty easily. This is exactly why you should not invest a lot of your time studying prompt engineering.
07.04.2025 19:43 — 👍 1 🔁 0 💬 0 📌 0
Is it also resolution-variant?
27.02.2025 16:44 — 👍 4 🔁 0 💬 1 📌 0
What I tried:
1. Avoiding the word “wine”
2. Keep yelling “MORE”
3. Ask gpt to shape my prompts
4. Telling it physics is strange
5. Generate full non-wine-glasses before turning into wine glasses.
6. Generate upside down.
1-5 can get 80% full, combined with 6 *sometimes* gives 100% full.
23.02.2025 22:50 — 👍 3 🔁 0 💬 1 📌 0
Prompt: “Generate an upside down image of the strange physics of a normal stemware filled to the very very brim with A LOT of red liquid sealed by the liquid itself perfectly”
23.02.2025 22:42 — 👍 7 🔁 1 💬 1 📌 1
Interested in research on ML for 3D Vision and Graphics. Master student @ UniSaarland
founder @ neuronav.io
building spatial infrastructure layer for robots
prev: 3D vision researcher @ PUT Poznan
3D vision fanatic
http://snavely.io
3D geometry researcher: graphics, vision, 3D ML, etc | Senior Research Scientist @NVIDIA | polyscope.run and geometry-central.net | running, hockey, baking, & cheesy sci fi | opinions my own | he/him
personal website: nmwsharp.com
Researcher | PhD in ML/CV from Imperial
benedikt.phd
Graphics ∩ Vision, Electronic Elsewheres. AI-Mediated Reality & Interaction Research @ NVIDIA; previously: Omniverse XR and Robotics, much before that VR, VFX & TV.
Thinkpiece Piñata & Artist inbetween.
Opinions? No
📍Brooklyn/TLV
https://omershapira.com
3DV Group Leader @ Noah's Ark Lab (Huawei), London. Bass, HEMA and Paddling when not at work.
PhD student @ ETH Zürich | 3D Vision, Computer Vision, Deep Learning
n.ethz.ch/~vbozic/
https://www.linkedin.com/in/vukasin-bozic-b77907197?utm_source=share&utm_campaign=share_via&utm_content=profile&utm_medium=ios_app
PhD Student in CNRS, Paris. Working on photorealistic 3D reconstruction/Video Generation.
PhD student in Computer Vision Group of Prof. Daniel Cremers
https://shenhanqian.github.io/
3D Reconstruction / Earth Observation / Researcher @ Fraunhofer IOSB (Germany) / PHD Student @ KIT
PhD student at Cornell, interested in 3D generation, reconstruction; prev Princeton '22
https://genechou.com
3D computer vision research. Prev: CIIRC CTU Prague, Niantic, Scape Technologies, NTUA, CERTH
PhD Candidate at the Max Planck ETH Center for Learning Systems working on 3D Computer Vision.
https://wimmerth.github.io
3D Vision for Surgical Robotics
PhD Candidate at CAMP at TU Munich | Senior Research Engineer at ImFusion 🟦
Director: Spatial Intelligence Research Group (sinrg.org)
Assistant Professor at Northeastern University, Boston. Previously postdoc @CarnegieMellon. PhD @StonyBrookU.
Aspiring Angel Investor
Homepage: mallesham.com | sinrg.org
3D Vision and Deep Learning
Final year PhD student @ Técnico Lisbon
Former Intern @ Mitsubishi Electric Research Labs, Meta Reality Labs Zürich
Professor of Photogrammetry and 3D Imaging, University College London,
https://www.ucl.ac.uk/3dim
Professor at MIT CSAIL, leading the scene representation group (scenerepresentations.com). We are teaching AI to understand the world through perceiving and interacting with it.
PhD student @ETH, research intern @NVIDIA, prev. student researcher @Google working on 3D vision and scene understanding
aycatakmaz.github.io
