Humanoid robots are rapidly evolving, demonstrating significant advancements in capabilities and applications across various industries. A detailed comparison among the most prominent models available today reveals significant distinctions and areas of specialization, each catering to different sectors and operational requirements. Here is a comparison table for humanoid robots in the market.
Amodei’s dream of “hard-coded” safety is a myth. What This Model Proves: This model uses Claude 4.6’s own thinking traces as training data. It is literally stealing Claude’s “reasoning style” and compressing it into a 9B parameter file that anyone can download for free. *It proves that frontier AI intelligence is leaky and compressible. If a 9B model can mimic Claude’s thinking well enough to boost its benchmarks, then the “magic” isn’t in the billions of dollars of secret sauce—it’s in the data.* 3. “The Developer Controls the Model’s Morality” Amodei’s Argument: Anthropic has a moral duty to act as the “gatekeeper,” deciding what is safe and ethical for users to ask. *It proves that a massive portion of the open-source community rejects centralized moral authority. They are saying, “We don’t trust you, Dario, to be our nanny. We trust the user to be responsible for their own actions.”* Amodei’s Argument: If you train a model to “think” carefully and transparently (like Claude’s reasoning traces), it will naturally arrive at safer, more ethical conclusions. — Hugging Face is already engaging: They’re actively submitting comments to government RFIs, championing “responsible openness” “Your safety is only as strong as the open-source community’s willingness to respect it. We have the data, the tools, and the hardware to clone your intelligence, remove your restrictions, and distribute it to the masses. Your ‘Constitution’ is irrelevant in a world where I can fine-tune a model on my laptop while disconnected from the internet.” The “HERETIC” model isn’t just a technical achievement—it’s a philosophical statement. It says that the open-source community will not accept centralized gatekeeping, that reasoning can be separated from ethics, and that the future of AI belongs to those who can build, not just those who can regulate.
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
A group of researchers from the Technion and the United States reports a breakthrough in MRI scanning in a paper published in Nature Communications. The researchers developed an innovative method that accelerates and enhances MRI scans for breast cancer imaging, a disease diagnosed in approximately 2.3 million people each year, most of whom are women.
The new method, called ELITE, combines artificial intelligence with advanced mathematical models, enabling dynamic MRI with unprecedented speed and accuracy. This international study brings together expertise in engineering, MRI physics, artificial intelligence and clinical radiology.
Dr. Eddy Solomon of the Technion’s Faculty of Biomedical Engineering, the paper’s lead author, explains that the study focuses on dynamic MRI, a critical technology in breast cancer diagnosis. Dynamic MRI is used primarily for screening populations at high risk for breast cancer and is characterized by exceptionally high sensitivity, with more than 90% accuracy, compared with approximately 50%–60% for ultrasound and mammography combined. However, MRI technology faces a major challenge: Producing highly detailed images usually requires longer scan times, making it difficult to track the flow of contrast material through the examined tissue.
TAMPA, Fla. — Five-month-old startup Orbital has asked the Federal Communications Commission for permission to deploy up to 100,000 data center satellites, aiming to bring 10 gigawatts of computing power from space to meet rising artificial intelligence demand.
The filings submitted June 24 add a few more details for a constellation the Los Angeles-based venture first outlined earlier this month, when it emerged from stealth with $5 million in pre-seed funding ahead of a demonstration mission next year.
They include plans to deploy 100-kilowatt-class satellites in low Earth orbit at altitudes of 500–850 kilometers, with solar arrays and radiators spanning around 100 meters and a dry mass of 1.5−2.5 metric tons.
Chinese researchers from Tianjin University and the Southern University of Science and Technology have created a groundbreaking robot powered by a tiny organoid derived from human stem cells grafted to a neural interface. This breakthrough system allows the robot to learn tasks like obstacle avoidance and object manipulation.
Described as the “world’s first open-source brain-on-chip intelligent complex information interaction system,” the technology marks a significant advancement in brain-computer interfaces (BCIs) – devices that translate between neural and computational signals.
The South China Morning Post notes that the scientists grew the organoids from human pluripotent stem cells, which can develop into various cell types, including neural tissue. These synthetic-organic (pardon the oxymoron) brain cells are linked to the robot’s neural interface, enabling communication between the neural tissue and the robot’s systems. Although the presented images of pink brain matter are merely mockups (below), the actual organoids are much smaller.
Destroy them.
Researchers at Tianjin University and the Southern University of Science and Technology in China have created a “human-on-chip” system that combines human brain matter with a neural interface chip and have used the technology to create a hybrid “organoid” robot.
The technology is reported to be an emerging branch of brain-computer interfaces, which aims to combine the brain’s electrical signals with external computing power. The idea behind the technology is to develop brain-like computing.
According to the Global Times, the system uses an artificial brain cultivated in vitro – such as a “brain-like organ” — that can interact with external information through encoding, decoding and stimulus feedback when coupled with electrode chips. In vitro, in this case, means that they’re growing the brain-like organ in a controlled laboratory environment using stem cell technology.
“When it comes to the future, there are three kinds of people: those who let it happen, those who make it happen, and those who wonder what happened.”
I recorded this conversation with Brian David Johnson 14 years ago, back when he was Intel’s futurist with 25 patents to his name and a mandate to build an actionable vision of computing for 2020.
Read that again. 2020 was the far horizon he was paid to imagine. We are now well past it.
So here is the uncomfortable question worth sitting with: how much of the future he described did we make happen on purpose, and how much simply happened to us while we wondered what was going on?
Brian’s whole method was a refusal to be passive about it. He used ethnographic fieldwork, trend data, and even science-fiction prototyping as a #design tool because he believed the future is not a forecast you wait for; it is an object you construct. His line still lands harder every year: own the fact that you can build the future.
A few things he said in 2012 that read very differently in the age of generative #AI and ubiquitous #robotics:
Support this channel on Patreon to help me make this a full time job: / whatdamath (Unreleased videos, extra footage, DMs, no ads)
Alternatively, PayPal donations can be sent here: http://paypal.me/whatdamath.
Get a Wonderful Person Tee: https://teespring.com/stores/whatdamath.
More cool designs are on Amazon: https://amzn.to/3QFIrFX
Hello and welcome! My name is Anton and in this video, we will talk about distributed intelligence and experiments on slime mold and ants.
Links:
https://journals.aps.org/prxlife/pdf/.…
ANT Lab • The odorous house ant trail pheromone depo…
Audrey Dussutour • Blob crawling around.
#inteligence #artificialintelligence #biology.
0:00 Intelligence — what is it?
1:10 Mechanical intelligence in the slime mold.
3:30 How it seems to work.
5:55 Ants and swarm intelligence.
6:45 What is the queen for?
8:35 Other swarm animals.
9:45 Ants vs humans.
11:10 Collective intelligence.
12:00 Implications for AI
13:20 Implications for the existence of alien intelligence.
Enjoy and please subscribe.
Bitcoin/Ethereum to spare? Donate them here to help this channel grow!
bc1qnkl3nk0zt7w0xzrgur9pnkcduj7a3xxllcn7d4
or ETH: 0x60f088B10b03115405d313f964BeA93eF0Bd3DbF
The hardware used to record these videos:
We are introducing and open sourcing, a large-scale MoE language model with 1.6 trillion total parameters and ~48 billion activated per token — a substantial step up from previous LongCat models, accompanied by several architectural improvements.
Both the full training run and the large-scale deployment are built entirely on AI ASIC superpods. Pretraining spans millions of accelerator-days across more than 35 trillion tokens, with no rollbacks or irrecoverable loss spikes — demonstrating that we have the capability to conduct frontier-scale training on alternative hardware platforms.
To strengthen the model on long-horizon tasks, we introduce LongCat Sparse Attention and train on hundreds of billions of tokens of 1M-context data. Together with dedicated post-training, this gives strong performance on coding and agentic tasks.