Abhijit Mahabal

There is more than one way to see it™. Also, Knowledge Architect at Pinterest.

[Abhijit Mahabal @amahabal]

@aryaman2020 Are you falling prey to Sturgeon’s law? https://t.co/ycGzXlS353

[Abhijit Mahabal @amahabal]

@ShriramKMurthi Applies to other religions as well, e.g., can an American reading “the hitchhikers guide to the galaxy” truly understand it without knowing cricket?

[Abhijit Mahabal @amahabal]

@ShriramKMurthi I thing therefore I am.

[Abhijit Mahabal @amahabal]

@RichardSSutton Human knowledge can mean “what humans know” and also “how humans come to their knowledge”. Are you only cautioning against the former but not the latter?

[Abhijit Mahabal @amahabal]

@andrewgwils But… today’s computers are also benefiting from that same evolution (as Dennett suggests in from bacteria to Bach and back): evolution made humans who made computers. And to the extent what I say is indirect, so is the transfer of samples from parent to child

[Abhijit Mahabal @amahabal]

@ShriramKMurthi Doug is at IU, yes, but GEB is his grad school work and predates IU.

[Abhijit Mahabal @amahabal]

The best take I have seen on whether AI can replace mathematicians.

[(((ل()(ل() 'yoav))))👾 @yoavgo]

AI cannot replace mathematicians, because the only reason to do math work is for making mathematicians happy. so either the mathematicians will do math with help of AI and be happy, or no one will do math at all.

[Abhijit Mahabal @amahabal]

@xzzzxxzx Even the formula is incorrect.

[Ritesh Banglani @banglani]

English: customer retention

Urdu:
Humko unse wafa ki hai ummeed
Jo nahin jaante wafa kya hai

[Abhijit Mahabal @amahabal]

@rao2z @Nature How does this work? I thought the only choices available were existence and non existence, not existence after time t.

[Abhijit Mahabal @amahabal]

@yoavgo Yup. These are, however, not “solving” hard problems; they are approximation algorithms (and very P). Agree that it is amazing how such simple methods very reasonably approximate immensely complex tasks.

[Abhijit Mahabal @amahabal]

@banglani False dichotomy. In reality they are like the prehensile tail of a new world monkey

[Abhijit Mahabal @amahabal]

@ShriramKMurthi This, but one hand holds an eraser instead

[Abhijit Mahabal @amahabal]

@karpathy "potential is 🔥" Pun intended?

[Abhijit Mahabal @amahabal]

@yoavgo @autonomy_future Indeed. If you give a kid a cookie but give their sibling two cookies, it is not a reward for child 1

[Abhijit Mahabal @amahabal]

@cosminnegruseri Would love to!

[Abhijit Mahabal @amahabal]

@cosminnegruseri There is one math. Presto and sparksql and God knows what else look similar but aren’t. Even sparksql is unlike sparksql (this is not a typo): one uses 0 indexed arrays, the other uses 1 indexed.

[Subbarao Kambhampati (కంభంపాటి సుబ్బారావు) @rao2z]

What is mind-boggling here is that googling for one of the possibly fake citations that R2 highlighted gives an AI overview that makes it look like the paper existed, but with a circular reference back to the @iclr_conf paper citing it.. A glimpse of our afactual world to come..

[Abhijit Mahabal @amahabal]

Given today’s AI inundation, I just realized, an old pun (two meanings) has now become a double pun: On being asked “can you pun on the zodiac?” the punster replied “By Gemini, I can sir”.

[Abhijit Mahabal @amahabal]

Wonderful exploration of a math problem using several LLMs and why it ain’t slam dunk that they will accelerate research

[Find me on bsky @colin-fraser.net @colin_fraser]

Let me sum up the episode that took place in this thread, because I think it's instructive and microcosmic of at least one way that I expect LLM-"assisted" research to progress in the real world. Anecdotal and you know I'm predisposed as a hater but I think it's a good case study