omg that’s so cool!!! how long did this take?

haha omg tqdm is so great!! i felt the same when i found it. it took away so much anxiety from just waiting >24 hrs for my mcmc code to finish running and also bye to the avalanche of print statements! 😅

i’d like the option to get notified when a friend/someone I follow posts (and now I want a twix bar 😄)

the way my entire body just yelled HIT THE SPACE BAR so it goes into that curvy thing 😂

reasoning benchmarks (which have improved over time and were updated with a new graduate-level one last year!) but to say that it “absolutely cannot” feels like an overstatement.

Also, how you word your prompts and how much information you give it is important. When probed, it does acknowledge its mistakes and corrects itself. Now, I don’t know if that’s reasoning like how we understand reasoning. For that, all we have are the mathematical, scientific and commonsense

engineering, work projects, and yesterday when my nvidia geforce app stopped working (which it successfully debugged!). It often returns accurate and usable code and it can analyse data and be methodical BUT domain knowledge is still important and not trusting it blindly (at least not yet) is key.

so great to see so many nods to 12 angry men!! That and rear window are two of my faves - fascinated by bottle movies. A more recent film is parasite *chefs kiss* 10/10

this is so flipping cute!!! the animation is so warm 😌

well, if you ever need an enthusiastic user review, 🙋🏽‍♀️

yes!! It feels so much more human-like and nuanced in the way it works with you. A lot of the times ChatGPT feels stiff, like it just wants the conversation to end.

This is an incorrect calculation of the area of a triangle by a young Babylonian student at school around 1900-1600 BCE, excavated from the ancient city of Kish.

It’s a moment in the history of math, of education, of children, of mistakes, of people being people. I love it.

love it, thanks!! also, can I just take a moment to geek out over how much I love hugging face😄 the NLP course is what got me into LLMs and the docs/hub made it so easy to train/deploy models. I rave about hf to anyone who asks where to start. so huge shoutout to yall for making ml more accessible!🙏

i love the colours!! SO pretty 🤩

👋🏽 id love to know the same thing!! 👀

missed opportunity for a git init 😝