prasun /prʌˈsun/

[prasun /prʌˈsun/ @p_r]

@StuFlemingNZ

[prasun /prʌˈsun/ @p_r]

@ShriramKMurthi https://t.co/B9Qf8207HG

[prasun /prʌˈsun/ @p_r]

@ShriramKMurthi https://t.co/iz4hPqtGlF

[prasun /prʌˈsun/ @p_r]

@ShriramKMurthi https://t.co/B9Qf8207HG

[Ethan Mollick @emollick]

@karpathy We wrote a paper on this last year based on data from a randomized controlled trial calling this the "jagged frontier" of AI. https://t.co/vTbYZryoJI

[prasun /prʌˈsun/ @p_r]

@ShriramKMurthi https://t.co/iz4hPqtGlF

[Andrej Karpathy @karpathy]

Jagged Intelligence

The word I came up with to describe the (strange, unintuitive) fact that state of the art LLMs can both perform extremely impressive tasks (e.g. solve complex math problems) while simultaneously struggle with some very dumb problems.

E.g. example from two days ago - which number is bigger, 9.11 or 9.9? Wrong.
https://t.co/dUrR6wm8GC

or failing to play tic-tac-toe: making non-sensical decisions:
https://t.co/XarwfUBtod

or another common example, failing to count, e.g. the number of times the letter "r" occurs in the word "barrier", ChatGPT-4o claims it's 2:
https://t.co/xpffK2r0pv

The same is true in other modalities. State of the art LLMs can reasonably identify thousands of species of dogs or flowers, but e.g. can't tell if two circles overlap:
https://t.co/HCXxBxosAu

Jagged Intelligence. Some things work extremely well (by human standards) while some things fail catastrophically (again by human standards), and it's not always obvious which is which, though you can develop a bit of intuition over time. Different from humans, where a lot of knowledge and problem solving capabilities are all highly correlated and improve linearly all together, from birth to adulthood.

Personally I think these are not fundamental issues. They demand more work across the stack, including not just scaling. The big one I think is the present lack of "cognitive self-knowledge", which requires more sophisticated approaches in model post-training instead of the naive "imitate human labelers and make it big" solutions that have mostly gotten us this far. For an example of what I'm talking about, see Llama 3.1 paper section on mitigating hallucinations:
https://t.co/pjuxoIOJCY

For now, this is something to be aware of, especially in production settings. Use LLMs for the tasks they are good at but be on a lookout for jagged edges, and keep a human in the loop.

[prasun /prʌˈsun/ @p_r]

@stevenstrogatz @m_j_wiener > see his tweets below

I wasn't able to find the tweets without dinner digging. Should be easier to find by following backwards from this tweet

https://t.co/DRoCucENdH

[Steven Strogatz @stevenstrogatz]

@MarcosCarreira @m_j_wiener You guys realize what a precious thing it is to be able to do math together. I love it that we can understand each other, admit our mistakes if and when we make them (nobody’s perfect!) and figure out the truth together.

[prasun /prʌˈsun/ @p_r]

@alex1craig @doodlestein @ryancarson You ask AI to do it

[prasun /prʌˈsun/ @p_r]

$AMD 80.86 #AMD

[prasun /prʌˈsun/ @p_r]

$AMD 486

[prasun /prʌˈsun/ @p_r]

@lauriewired One of the reasons that we have silicon transistors is that silicon could run hotter than germanium

[prasun /prʌˈsun/ @p_r]

@BlrLitFest Jupiter Jones, Pete Crenshaw, and a Robert Andrews

[prasun /prʌˈsun/ @p_r]

@tennisabstract What do these numbers look like now

https://t.co/gePnrmKaZq

[prasun /prʌˈsun/ @p_r]

@tennisabstract Rafa's going to win this, isn't he

[prasun /prʌˈsun/ @p_r]

@tennisabstract Rafa's going to win this, isn't he

[prasun /prʌˈsun/ @p_r]

@tennisabstract Rafa's going to win this, isn't he

[prasun /prʌˈsun/ @p_r]

@patrickc When John Templeton said that, he was talking about learning from your mistakes instead of thinking "this time it's different" and making the same mistake again.

[prasun /prʌˈsun/ @p_r]

@kaushikcbasu Roger Penrose discovered Penrose tiles in the 1970s. These tiles use simple shapes to achieve a non-repeating pattern ("aperiodic tiling").

People observed later that this concept of aperiodicity had appeared centuries earlier with Girih tiles used in the darb-i imam shrine.

[prasun /prʌˈsun/ @p_r]

@ChShersh Michael J. Flynn, creator of the foundational 1966 Flynn's taxonomy for classifying computer architectures (SISD, SIMD, MISD, MIMD) based on instruction/data streams, passed away on December 24, 2025, at age 91.

[prasun /prʌˈsun/ @p_r]

@Thom_Wolf https://t.co/ffItSnuFeX

[prasun /prʌˈsun/ @p_r]

@kaushikcbasu There's an app for that

https://t.co/AG0dt5X3zD

[prasun /prʌˈsun/ @p_r]

@unixterminal ~94 ms
94 million ns x ~3 cycles/ns x ~3 insts/cycle
~1 billion instructions

[prasun /prʌˈsun/ @p_r]

@jimkxa https://t.co/BMMjZahbCD