uptownfunk

why is this flagged?

Can we build an OS version of this and make it easy for solo dev to self host / roll their own?

This is like looking at the kids who collect Pokemon cards out of the sheer love of the Pokemon and playing the game compared to the scalpers and “investors” / “speculators”. Yes they may end up with a lot of money but they will always lack that sweetness, purity, and wholesome joy that you see when a kid finds and enjoys something.

There is something cult like, mystic, almost magical about Ramanujan that almost transcends the raw intellectual horsepower (which he certainly had).

No hate here. And you’re very much not quoting anything I said so who knows where you got that from. Don’t be so insecure.

Ramanujan responsible for inspiring generations of mathematicians throughout the world. His life was a beautiful tragedy. One that leaves me in awe and also great inner sadness. If you come from a hardcore traditional br*hmin family, just to cross over the ocean by boat would risk you getting excommunicated. The culture which he came from makes the entire story all the more legendary. Just cutting off your topknot and forgoing the dhoti to wear a western suit. We don’t understand what he went through and what he gave up to give us his mathematics. What he had to sacrifice to practice his art.. to be.

Yes this is what I am referring. I’m not saying he is like the boy coder genius, he is good at becoming and staying king. And if you look at all the departures of some of today’s brilliant technology innovators, perhaps it tells you why. Don’t fly too close to the sun.

Sam is a scary good guy. But I’ve also learned to never underestimate Microsoft. They’ve been playing the game a long long time.

I love how out there this guy seems and kudos to his courage for sharing his thoughts and feelings and being able to go and do something to the beat of his own drum.

Solutions to major mathematics problems generally require the creation of entirely new objects and the appropriate machinery to analyze them. The behavior of said objects generally ends up providing the solution to the original question as a special case. A new architecture is required but one that is probably not far off from the original transformer in nature.

Hahaha awesome go for it

Most of these comments are just cursor folks selling the product

The very fundamental problem with LLM is there is no guarantee on any of the reasoning it gives you without a human there to give a thumbs up. They are working on solving this (alpha proof, lean agent etc) but getting this to run at inference time in an optimized way is what I would call one of the millenial prize problems of AI which will lead to a quantum leap in the path towards the singularity.

Curious why it ends up causing cancer..

One of the greatest entrepreneurs in the world. I hope they make a movie on him and his life.

To me this is more controversial than Geoffrey Hinton winning it for physics.

I only drink or smoke a little if I need to get out of my head to get mundane stuff done like laundry or house cleaning and stuff

It sucks, the people are everything. Integrity and execution are the main ingredients. The rest is luck. But no integrity or no execution and you implode. You literally just need integrity and execution.

Folks. This is what innovation looks like. Well done chaps

Fuck this is so cool love it

Do folks here know how expensive it is to develop a drug? How much work it takes to get it through the pipeline? How much time, heartache, effort, has gone wasted? How many patients given false hope? This is tragic on so many levels

This is so messed up.

Even if it does the problem will still be there

Will this fix the problems it claims to? The power of R is the rich package ecosystem. It caters to people who don’t want to think about engineering concerns but want a fast way to access the powers of computation rather than building a scalable system, two very different things. It excels at the former. A new language will not fix this, because this type of thinking has infected the entire package ecosystem. Frankly with code translation you probably don’t need a new language. Prototype in R and code translate to Python or whatever you want to use in prod. Or frankly just do code gen directly in Python so you can skip having to confirm if the results match.

To be clear, I love R, it excels in prototyping but I have seen too many real world struggles of folks trying to move to prod that I would say save it for EDA projects and one time analyses.

It’s marketing to get the best researchers. The researchers want the meta pay and they want to hedge their careers to continue to publish. That’s the real game, it’s a war for talent. Everything else is just secondary effects.

It’s for VCs to go ooh and aah and boost their valuation.

This hits the nail on the head. It is a consumer facing product not a technology to solve deep thinking.

Thanks for your work and also for your comments of AF3 and Chai-1. It sounds like you are implying there are potentially gross and subtle types of data set leakages taking place between the train and test which are resulting in what seem to be inflated performance metrics? These are pretty serious issues if so. Also I would agree with previous authors that marginal Improvement over sota is proof more that they have recreated something than really made significant new progress. But this has been an issue with LLMs for sometime now. But it sounds like they have some bright engineers from good brand name companies who are coming together with some VC backing of the team to try and do something in this space. I do appreciate that the weights are open. I would like to learn more about their future direction and their training methods

What happens when the new models come out and the data centers are full of old models to be decommissioned. Would love to buy a huge amount of h200 once they have become “obsolete”

Wow the end of an era. I first heard about it at Cal through college roommates. I haven’t used it much lately but still.. sad to see it go.