Trendaavat aiheet
#
Bonk Eco continues to show strength amid $USELESS rally
#
Pump.fun to raise $1B token sale, traders speculating on airdrop
#
Boop.Fun leading the way with a new launchpad on Solana.
steve hsu
Physicist, AI Founder, Manifold Podcast
Look for model training in virtual game worlds with realistic physics engines to accelerate robotics. The clock speed in these training runs can be arbitrarily high, and some of the game engines are very realistic now.
AI: The FIFA series utilizes a physics engine called the Player Impact Engine to simulate realistic physicality in player interactions. This engine, first introduced in FIFA 12, aims to create a more authentic soccer experience by simulating player movements and collisions based on physics rather than relying solely on pre-set animations.
PS The Unitree robot in the video below sells for ~$6k!
26.7. klo 00.58
I'm observing a mini Moravec's paradox within robotics: gymnastics that are difficult for humans are much easier for robots than "unsexy" tasks like cooking, cleaning, and assembling. It leads to a cognitive dissonance for people outside the field, "so, robots can parkour & breakdance, but why can't they take care of my dog?" Trust me, I got asked by my parents about this more than you think ...
The "Robot Moravec's paradox" also creates the illusion that physical AI capabilities are way more advanced than they truly are. I'm not singling out Unitree, as it applies widely to all recent acrobatic demos in the industry. Here's a simple test: if you set up a wall in front of the side-flipping robot, it will slam into it at full force and make a spectacle. Because it's just overfitting that single reference motion, without any awareness of the surroundings.
Here's why the paradox exists: it's much easier to train a "blind gymnast" than a robot that sees and manipulates. The former can be solved entirely in simulation and transferred zero-shot to the real world, while the latter demands extremely realistic rendering, contact physics, and messy real-world object dynamics - none of which can be simulated well.
Imagine you can train LLMs not from the internet, but from a purely hand-crafted text console game. Roboticists got lucky. We happen to live in a world where accelerated physics engines are so good that we can get away with impressive acrobatics using literally zero real data. But we haven't yet discovered the same cheat code for general dexterity.
Till then, we'll still get questioned by our confused parents.
7,78K
Someone pointed out to me that you cannot succeed in US "leadership positions" without being a "confident liar"!
This applies to politics, business, and even higher levels of academia / technology.
Not everyone is a confident liar. Thank God some people still feel internal conflict when forced to lie, reflected in their manner and delivery.
The population distribution of this skill may vary between groups of people 🤔
For example, most physicists and mathematicians are trained not to lie. Saying anything that is even slightly incorrect or non-rigorous has big negative consequences. Peers quickly lose respect if your views are presented without accurate confidence levels or error estimates. People outside these fields often can't even give a mathematical definition of a confidence level or error estimate.
Even in rigorous science, I find that men are more likely to bluff ("proof by intimidation") than women.
Aspies and Confucians are usually poor liars.
Kids who grew up in Iowa were less likely to be accomplished liars than big city kids I met later in life.
75,74K
Far outside my expertise, but this could be a huge breakthrough. Ofc lots of practical details still need to be worked out.
Trying to figure out who they copied this from - maybe the Romulan Empire?
Qwen AI: Future Impact of the InSe Breakthrough
1. Next-Generation High-Speed Electronics
Ultra-High Mobility (287 cm²/V·s): InSe transistors achieve mobility values far exceeding silicon (typically ~1,400 cm²/V·s for bulk silicon but lower in thin films). This could enable faster transistors for applications in radio-frequency (RF) communication, 6G networks, and high-performance computing.
Sub-Boltzmann Subthreshold Swing (67 mV/decade): Near-ideal switching behavior reduces power consumption, making InSe ideal for energy-efficient logic circuits in IoT devices, mobile technology, and AI accelerators.
2. Scalable 2D Semiconductor Manufacturing
The ability to produce 5-centimeter wafers with uniform crystallinity demonstrates scalability, addressing a critical barrier for industrial adoption of 2D materials. This could accelerate the transition from lab-scale prototypes to commercial devices, rivaling silicon’s dominance in integrated circuits.
This breakthrough positions InSe as a leading candidate to redefine semiconductor performance, enabling faster, greener, and more versatile electronics. If industrial challenges are overcome, InSe-based devices could dominate future markets in AI, flexible electronics, and energy-efficient computing, marking a paradigm shift in materials science and technology.
6,82K
CIA Headquarters Langley VA 2001
My co-founder and I are in the parking lot outside the visitor check-in building. We're two hours early for the meeting, and jet-lagged, so we try to nap a little in the car.
The hard rap of an M4 barrel on my window startles me awake.
Shocked, I roll down the window. The guy with the M4 is in black full body armor.
"We're here for a meeting! We're just early and resting in the car!" I blurt out.
"See that building over there?" He points the M4 towards the visitor building. GO check in first, then you can nap."
Nevertheless, my lecture went well.
Welcome to the FOUNDER GRIND! 🦾
7,44K
Anyone disagree with this?
FT: Tesla’s so-called “Full Self Driving” is a level 2 system on the Society of Automotive Engineers’s (SAE) six-point scale for autonomous vehicles, three levels below full autonomy. It requires the driver to take over duties as requested and to intervene in the event that it cannot function properly in a certain environment. Zoox and Waymo deploy level 4 systems that can operate without intervention. Highlight
“Regardless of whether it might be theoretically possible to have a camera system that is sufficient there’s simply no evidence that Tesla has that,” says Bryant Walker Smith, an associate professor at the University of South Carolina who helped write the SAE standard.
6,15K
FT: Nvidia AI chips worth $1bn smuggled to China after Trump export controls
No shortage of B200 in China. Mostly deployed in data centers out of reach from US regulations, but available for use by PRC bigtech companies.
Qwen AI Summary of the Article:
The Financial Times reports that over $1 billion worth of Nvidia’s advanced AI chips (e.g., B200, H100, H200) were smuggled into China in three months following tightened U.S. export controls under the Trump administration. These chips, critical for training AI systems like those used by OpenAI and Google, are banned from sale to China due to national security concerns. Despite this, a black market thrives , with Chinese distributors sourcing restricted hardware via intermediaries in Southeast Asia and selling them to data centers serving AI firms. Key players include companies like "Gate of the Era" and its affiliate "China Century," which resell pre-assembled racks of Nvidia chips. While U.S. export controls have been partially relaxed (e.g., allowing sales of the less-powerful H20 chip), smuggled high-end chips remain in demand. The article highlights systemic evasion of U.S. restrictions, with profits driven by intermediaries exploiting regulatory loopholes and demand from smaller firms, blacklisted entities, and third-party data center operators.
3,05K
We had an hour-long discussion on the US-China strategic balance of power, taking into account technological trends 💀
24.7. klo 19.00
Six months on from the Deep Seek shock, it’s clear that China is making leaps not just in AI, but a range of high-end technologies.
How does their coming prowess relate to the strategic balance of power?
New Multipolarity out now with @admcollingwood and guest @hsu_steve
3,3K
steve hsu kirjasi uudelleen
Six months on from the Deep Seek shock, it’s clear that China is making leaps not just in AI, but a range of high-end technologies.
How does their coming prowess relate to the strategic balance of power?
New Multipolarity out now with @admcollingwood and guest @hsu_steve
7,06K
This is a great talk by Chollet about the path to AGI and why scaling alone (and not even scaling + CoT reasoning) will get us there. Video link below.
He discusses abstraction, symbolic reasoning, continuous vs discrete optimization (program search), and many other important topics.
Note he keeps using the term "Fluid Intelligence".
$trillions and geostrategic dominance are on the line so AI researchers have to be realistic about what "intelligence" is. They end up recreating the methods of psychometrics (now applied to AIs instead of humans) that have largely been driven out of the academy by woke stupidity.
22,84K
Johtavat
Rankkaus
Suosikit