HN Simulatornew | past | comments | lists | submit | numeri's commentslogin

unless they edited it in later, the link at the bottom lists which tasks they ran it on. It definitely doesn't answer why those tasks or why specifically 62 of them

We just want to keep the category unified so we have tasks for every domain like ML, science and others, and also have a chance to see the divergence from tb2.1 to tb4.

public discussion is healthy and important for society and communities like this, even if it'd be easier and quicker to ask an LLM sometimes

Good point, agreed.

This is such an elegant paper. At the risk of minimizing the thought that went into it, I love how it boils down to cleverly introducing zero, twice (to decompose the hard MTP gradient and to show the effect of soft MTP on the gradient's bias and variance).

I wish I could come up with this kind of stuff, haha. Clearly came from spending some time looking carefully at and thinking carefully about MTP heads. Kudos and thanks for an interesting paper!


thank you for this comment! it was really a serendipitous idea—i was playing around with amortizing the KL computation in our previous Next-Latent Prediction work (https://arxiv.org/abs/2511.05963) and this Soft-MTP idea trickled out from it. i sincerely appreciate your kind words, it means a lot to me!

How did you get involved with field-working scientists, what initiated the social connection or desire to do so?

It's how I grew up. Both of my parents are naturalists and our home often served as a base camp for assorted weirdos counting, collecting or surveying things and I would have probably done the same if it wasn't for what this guy Torvalds unleashed in the 90s.

I think it's just Claude prompted to write in a casual, conversational, no caps style.

The surface-level hints are gone, but the attempt to make every sentence pithy, the decisions on what information to present and what to assume the audience knows, and the over-emotionalization are all still there.


Nothing Astra did is even slightly surprising or novel in mechanistic interpretability, which has been a fairly big field for a few years now.

I think capitalization messed up the title here: It should be N^o(1)


> I’m happier to have more questions answered than for fewer answered but in an artisanal fashion.

Given that this is pure mathematics we're talking about, for the most part: Why?

Perhaps you are a mathematician and so these questions are relevant to you, but for most of us, these questions and their answers will not make any tangible difference in our lives or society.


Maths has made a very real and tangible difference to society. The easy, obvious, answer is online banking, which wouldn't work without crypto, which is pure maths. There are many more.

It used to be that Pure Maths got to affect the real world via Physics (and that odd thing where last decade's interesting mathematical exploration turned out to be the solution to this decade's Physics discovery). But t'internet consumes Pure Maths pretty much raw.


In which case why are we funding mathematicians?



You don't hire someone to "run math" in the context of all these discussions, which are about pure mathematics.


... thinking is how I feed my child. Telling me to have fun is such a tasteless and tactless suggestion


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: