Kalhan

How Much Energy Does It Take to Power Billions of AI Queries? All generative AI queries could hit 329 billion per day by 2030. See the big picture on AI's energy use and how it's reshaping our world.

Today, OpenAI's ChatGPT handles 2.5 billion queries per day; the electricity used is equivalent to fully charging 14,000 electric vehicles. Get the big picture on energy use as we scale up to all generative AI queries per year, both now and in 2030.

This thread from @jaschasd at @AnthropicAI is a major wake-up call for early-career researchers, especially in #robotics.

We must build things that scale with intelligence, not compete with it.

Thread: x.com/jaschasd/sta...

#AGI #ComputerVision #PhDLife #EmbodiedAI

Most of the "90% of code written by AI" claims come from vendors selling AI tools and lack credibility as a result

Armin (creator of Flask, Jinja, Click) is different - when he says 90% of a new significant infrastructure project he's building was AI generated that's worth paying attention to

*NOT Pro, Plus.

And also there's the definition of novel too, we could define it in many ways. Some may say that what I was given is a novel idea.

It's what I experienced using Gemini Pro and ChatGPT pro versions to create a robotic mapping system. I could always find the papers from which the Chat bots took the ideas it suggested for me. It's useful but not truly novel. You may be referring to something else.

AI ideas are generally not really novel. It's more like combining different frontiers of state of the art.

Life-saving drones tested to reach patients faster than ambulances UK researchers have successfully tested drones carrying defibrillators to cardiac arrest emergencies, potentially doubling survival chances.

Researchers at the University of Warwick have trialled a drone-delivery system to transport defibrillators directly to the scene of a cardiac emergency, dramatically cutting response times and increasing patients' chances of survival. 🤖 interestingengineering.com/innovation/u... #Robots #Robotics

3D and 4D World Modeling: A Survey

tl;dr: in title

arxiv.org/abs/2509.07996

SpatialVID: A Large-Scale Video Dataset with Spatial Annotations

Jiahao Wang, Yufeng Yuan, Rujie Zheng, Youtian Lin, Jian Gao, Lin-Zhuo Chen, Yajie Bao, Yi Zhang, Chang Zeng, Yanxi Zhou, Xiaoxiao Long, Hao Zhu, Zhaoxiang Zhang, Xun Cao, Yao Yao

tl;dr: in title

arxiv.org/abs/2509.09676

🤖 Haven't re-upped this in... Like ever!

If you are interested in the robotics community in Colorado check out this list 👇

If you do robotics in CO and aren't on the list, I can add you!

I would like a box like robot to do my work without even me realising it's done. A very very smart dishwasher integrated into my home.

I would like a box like machine to do my work without even me realising it's done. A very very smart dishwasher integrated into my home.

I thought I wouldn‘t be one of those academics super into outreach talks, but I just put together something about understanding LLMs for laypeople and I get to talk about results that I don’t really focus on in any of my technical talks! It’s actually really cool. I made this lil takeaway slide

Humans don’t fully trust the #AI #agents that are coding for them, and with good reason. New research shows that AI agents still fall short when it comes to huge codebases, extended context over many lines of code, and long-term planning on structure and design. spectrum.ieee.org/ai-for-coding

Nauto’s AI Dashcam Outperforms Rivals in Driver Safety Tests Nauto's dashcam offers more than monitoring—it's about shaping safer driving habits with real-time alerts and coaching.

What if your vehicle could predict accidents? Nauto's AI dashcam offers real-time alerts to keep drivers safe and informed.

Prototyping has become so easy.

Yes

I have similar experiences with ROS 2 code for robotics 🤯

Neuromorphic Programming: Emerging Directions for Brain-Inspired Hardware The value of brain-inspired neuromorphic computers critically depends on our ability to program them for relevant tasks. Currently, neuromorphic hardware often relies on machine learning methods adapt...

Curious about #neuromorphic computing? 🧠💻
We want to revolutionalize the way we program brain-inspired systems and are plotting a course in a new publication with Steven Abreu: ieeexplore.ieee.org/abstract/doc... (or open access arxiv.org/abs/2410.22352)
Let's build better neuromorphics together! 🚀

Covariant spatio-temporal receptive fields for spiking neural networks - Nature Communications Neuromorphic computing mimics brain efficiency but lacks theoretical guidance. Here, authors develop a computational foundation for processing signals in space and time in spiking neural networks that...

New paper on covariant #neuromorphic networks!

We're connecting decades of work in computer vision with decades of work in spiking networks. And, in an event-based vision task against regular ANNs of similar complexity, spiking networks are doing much, much better!

www.nature.com/articles/s41...

@ UKCI 2025 @edinburghnapier.bsky.social university.

Do People Really Want Humanoid Robots in Their Homes? Investors are pouring billions into humanoid robots, but do people actually want them in their homes?

While some prefer humanoid robots for certain tasks, people generally prefer special purpose robots over humanoids.

Juho Hirvonen and Jukka Suomela: Distributed Algorithms 2020 A free online textbook — introduction to the theory of distributed algorithms

Aalto University course "Distributed Algorithms" starts tomorrow! All course material (including an online textbook and short videos) are freely available here: jukkasuomela.fi/da2020/

A Wearable Robot That Learns New control algorithm personalizes user experience for stroke, ALS patients

Engineers at Harvard have developed a soft, wearable robot that uses machine learning to respond to each user’s unique movements, which could provide personalised assistance to patients with neurodegenerative disease or people recovering from a stroke. 🤖 www.wevolver.com/article/a-we... #Robotics

I can't* fathom why the top picture, and not the bottom picture, is the standard diagram for an autoencoder.

The whole idea of an autoencoder is that you complete a round trip and seek cycle consistency—why lay out the network linearly?

Watch robot crab 'Wavy Dave' get attacked in claw-waving contest with real crabs Researchers have built a robot crab that can compete in claw-waving displays with real fiddler crabs, but "Wavy Dave" doesn't always come out unscathed.

Researchers in Scotland created a robotic fiddler crab to study the claw-waving contests that males of the species use to attract a mate. But the robot, nicknamed "Wavy Dave", didn't come out of the experiments unscathed. 🤖 www.livescience.com/animals/crus... #Robots #Robotics

Robot in between rows of blueberries

Robot in between rows of blueberries

Field day with #AIgro #robotics #agriculture. Robot is taking pictures to predict yield and spot diseases, plagues etc.

Major milestone in Reinforcement Learning 🚀

In under just 1 hour of training, our AI Agent can now automate tasks requiring complex physics and object interactions. Done for the first time on 100$ robot arms.

Watch as it recovers when the cube gets stuck 🤯

Episode 113 – Soft robotic hands - Robot Talk Claire chatted to Kaspar Althoefer from Queen Mary University of London about soft robotic manipulators for healthcare and manufacturing.

📢 New Episode Alert!

This week, I chatted to Prof. Kaspar Althoefer from Queen Mary University of London about soft #robotic manipulators for healthcare and manufacturing. 🤖 🖐️

Listen now: www.robottalk.org/2025/03/14/e... #Robots #Robotics #Prosthetics #Surgery #Manufacturing