Rajesh Kalane @rajkalane

How LLMs work, clearly explained with visuals:

Humans: nanotechnology will never work!!! Bacteria: hold my beer.

We trained a robot dog to balance and walk on top of a yoga ball purely in simulation, and then transfer zero-shot to the real world. No fine-tuning. Just works. I’m excited to announce DrEureka, an LLM agent that writes code to train robot skills in simulation, and writes more…

This may end up being a big deal: Usually LLMs just predict the next token in a sequence, one at a time, but if you have them predict the next several tokens at once you get significantly better performance, faster, and with no added costs. The gains are better for bigger models

@johnmdudley @stevenstrogatz Together with @math_martens we extended this experiment to two coupled populations — they can be in sync and go in phase, anti-phase or the counterintuitive broken symmetry state, with one in sync and the other desynchronised — a chimera pnas.org/doi/full/10.10…

The hidden mechanism behind AI regulations: 1. Claim that AI can do everything . 2. Raise tons of money from investors. 3. Tell governments that AI is very dangerous and that open source AI should be regulated out of existence. .... 4. Profits!

Bill Gurley @bgurley

4 days ago

What’s really happening in the background around AI regulation.

46 235 969 494K 304

Download Image

MLPs are so foundational, but are there alternatives? MLPs place activation functions on neurons, but can we instead place (learnable) activation functions on weights? Yes, we KAN! We propose Kolmogorov-Arnold Networks (KAN), which are more accurate and interpretable than MLPs.🧵

In 1977, the inventors of RSA presented a challenge ciphertext and offered $100 to anyone who could decipher it The decryption involved factoring a 129-digit number, which they predicted would take 40 quadrillion years By 1994, the challenge was solved 📝fermatslibrary.com/s/a-new-kind-o…

Astrophysicist and Nobel laureate Subrahmanyan Chandrasekhar used to travel 80 kilometers every week from the Yerkes Observatory to the University of Chicago, where he taught a course attended by only two students. When asked why he spent his time this way, the professor replied…

There was a wide consensus around eugenics in the West at the beginning of the XXth century. Cohen (2016) reminds us: “Harvard administrators, faculty members, and alumni were at the forefront of American eugenics—founding eugenics organizations, writing academic and popular…

While we stay looking for a fix to one issue we continue to place the onus on the individual, the child. We are asking them to specify what isn’t specific. We are not assessing the environment as a whole. Instead we need to consider it may be all or lots. 1/

I'm always so envious when I meet people who haven't read any Wodehouse novels, because it means they have them all to enjoy for the first time. That seems the ultimate accolade for a writer of fiction: when the people who've read your books envy those who haven't.

Today our doctor, an older woman with 7 kids, gave us some of the best media advice: she said that a lot of the videos she has of her older kids were long-form videos taken with old camcorders, 20-60 minute shots. You get to see so many different dimensions of a child's…

imo in order to attract a mate, more people would be better off cultivating their lifeforce vs. improving their physical appearance, status, and finances with diminishing returns. healing men "doing the work" is the most attractive thing to 100% of my girlfriends. we've…

My webinar on helping children with severe anxiety is tomorrow lunchtime. It’s aimed parents of those young people who have completely withdrawn and who may not be leaving their homes or bedrooms. It is recorded if you can’t make the time. Please share if you know families who…

Raman Sundrum, a theoretical particle physicist University of Maryland, is part of a cadre of particle physicists who hope to establish a connection between gravitational wave signals and high-energy particle physics. quantamagazine.org/hopes-of-big-b…

I always wondered how autonomous vehicles could work in the absolute chaos of driving in India- people don’t always follow the rules, humans and animals are often on roads, etc. This co only raised $3M & is in a tier 2 city (Bhopal) not known for tech. Very impressive.

@AlexisPolyzos This is annoyingly true. And something that hasn't been discussed enough in the founder community. We are encouraged to solve something easy first to "show traction." If the easy thing doesn't show traction, you need to market harder. In many cases, it is the other way…

A few years ago, my dad gave me the best dating advice I've ever received. He told me, “Alex, you need to stop dating people who want the same career as you and want to compete. Start dating people who are passionate about something but are down for the ride, soothe you, and can…

fearlessness of first time founders usually trumps the experience of second time founders because you need a certain level of foolish belief to invent the future — you only have it *once* in your entrepreneurial life.. the first time! there are rare exceptions, as expected.