Salman // 萨尔曼 @ForBo7_

「Open to Projects」 • Dabbler • Learner • Explorer • https://t.co/jTudwv3AAp student • https://t.co/7FBysT47M6 solver • Dabbling in Embodied AI • 自学中文 // Teaching myself Chinese forbo7.github.io Hong Kong/Shenzhen Joined September 2022

Tweets

1K
Followers

257
Following

989
Likes

380

𝐀𝐆 @AGkorthos

5 days ago

Nobody tells you this, but you’re allowed to stick a humanoid torso on basically anything.

80 144 2K 291K 481

View Details

Johannes Tscharn @JohannesTscharn

6 days ago

Augmented Reality is the natural interface for robotics. Not only for giving commands, but also to visualise hardware data (like LiDAR) natively in 3D. @specs AR glasses, @UnitreeRobotics Go2 Lite quadruped, @dimensionalos OS for the robot

101 386 3K 379K 1K

View Details

Salman // 萨尔曼 @ForBo7_

7 days ago

functional programming

0 0 1 67 0

View Details

PicoCreator - AI builder @ 🇸🇬 🔜 🌉 @picocreator

a week ago

Surprised we have yet to see a robotics startup pitching a household rail system, which skips all the current hardware problems of batteries and legs This lets it focus on the MVP, doing chores (in house), and hands.

116 61 598 74K 260

View Details

Simon Willison @simonw

a week ago

The idea of "AI employees" feels so short-sighted to me - both disrespectful to humans and a complete misunderstanding of what these tools can do and how to best put them to work You may as well start adding Excel spreadsheets to your org chart

176 188 2K 125K 217

View Details

Salman // 萨尔曼 @ForBo7_

2 weeks ago

sorting a list of file paths with a one‑liner lambda in python

0 0 1 49 0

View Details

Salman // 萨尔曼 @ForBo7_

2 weeks ago

Polya's framework is potent if you properly follow it along and spend proper time on each and every question asked in the How To list

0 0 0 28 0

View Details

Salman // 萨尔曼 @ForBo7_

2 weeks ago

pdb in solveit is cool, because the LLM can follow the same trace you are following

0 0 1 41 1

View Details

Salman // 萨尔曼 @ForBo7_

2 weeks ago

discovered this neat trick as I was reading through @answerdotai's fastllm library github.com/AnswerDotAI/fa…

0 0 1 11 0

View Details

Salman // 萨尔曼 @ForBo7_

2 weeks ago

if `^` is used instead, it will center align it

1 0 0 16 0

View Details

Salman // 萨尔曼 @ForBo7_

2 weeks ago

`n:>30s` means take n, turn it into a string, and then left align it w.r.t. → with an appropriate amount of whitespace so there 30 characters to the left of →

1 0 2 54 0

View Details

Rohan Paul @rohanpaul_ai

2 weeks ago

The robot’s “eyes” just received a big upgrade. LingBot-Depth 2.0, a depth-completion model with half the depth error just dropped. 12/16 benchmarks topped. Glass, mirrors, and transparent objects are so easy for us humans, but so hard for robots, because they do not behave like ordinary surfaces in a camera pipeline. A robot that misunderstands a balcony window or a table edge, will have a completely false planning inside a false world. Huge implecation. LingBot-Depth 2.0 takes an RGB image plus a broken depth map from a sensor and then outputs a cleaner depth map and a usable 3D point cloud. Numbers on LingBot-Depth 2.0 - Excels on glass, mirrors & transparent objects — where traditional depth cameras fail - Training data: 3M → 150M (50x scale-up) - 12 out of 16 first-place rankings on depth completion benchmarks - RMSE cut in half: 0.132 → 0.062 on the hardest indoor scenes LingBot-Vision trained on boundaries, because object edges carry the geometry robots need. No human boundary labels are used, which makes this approach easier to scale. The open-sourced LingBot-Vision is the general vision backbone, and LingBot-Depth 2.0 is the depth model built on it.

Robbyant @robbyant_brain

2 weeks ago

🪞 Glass. Mirrors. Transparent objects. — The nightmare of every depth camera. We just solved it! Introducing LingBot-Depth 2.0: 150M-scale training, half the depth error, 12/16 benchmarks topped. Powered by LingBot-Vision — the visual foundation model behind Depth's

113 134 502 549K 252

4 14 87 26K 69

View Details

harrrshall @HarshalsinghCN

2 weeks ago

introducing tinyrouter i reverse engineered the routing architecture behind Skana AI's Fugu and built replication for open frontier models. it's a tiny ~10K parameter LLM router that learns which model to use and what role it should play for every question. the goal was simple: beat every individual model by routing each task to the right specialist instead of relying on a single LLM. github.com/harrrshall/tin… a few interesting findings: • routing only helps when the models have complementary strengths. if every model performs similarly on a benchmark, a router has almost nothing to optimize. • on MMLU, the router beat every individual model. on math, it matched the best model because the pool had very little observable diversity. • warm-starting the router and shaping the evolutionary reward improved training, but we don't claim a real gain yet because evaluation variance was too high. more rigorous experiments are needed.

42 97 1K 101K 1K

View Details

Salman // 萨尔曼 @ForBo7_

2 weeks ago

thinking/thought is inherently creation

0 0 0 25 0

View Details

Lucas Beyer (bl16) @giffmana

2 weeks ago

When frontier labs suddenly cut costs, or say "we found a way to dramatically cut inference memory!!" This is what they found. The secret sauce. Vision always wins! Of course, my very capable ex-colleagues who now work on Gemini and Claude already found out years ago:

International Cyber Digest @IntCyberDigest

2 weeks ago

This is absurdly clever: there is a way you can cut Fable 5 costs by up to ~70%. Just turn Claude Code context into an image and make Fable OCR them. 😂 github.com/teamchong/pxpi…

136 318 5K 647K 4K

41 120 1K 184K 1K

View Details

Asimov @asimovinc

3 weeks ago

Asimov 1 is an open-source humanoid robot you can build and customize yourself. Two ways to get one: 1) Source the parts yourself: docs.menlo.ai/asimov/1/bom 2) Get the DIY kit: asimov.inc/diy-kit The kit bundles every part as a group buy, cheaper than sourcing one by one, and you build alongside others.