Rutuja (bio/acc)

Bioengineer . Exploring the design space of biology, models to biomolecules. prev @IBSE_IITM ,@DBT_inStem

[Silvi Rouskin @silvirouskin]

Ribosomal DNA producing MICRO RNA ??🤯🤯🤯🤯🤯🤯🤯🤯🤯 Biology never ceases to surprise me !

[Rutuja (bio/acc) @BhoiteRutuja]

@TanayLohia1 It considered strawberry as biological element XD
Weird tho if not doing anything what is it useful for

[Malika 🧬 @malikules]

affirmations:

• i am a complex organism brutally engineered by uncaring forces of nature
• i carry genetic code that survived 5 mass extinctions
• i am building a machine greater than myself
• i am able to make phone calls and appointments
• i can write emails without gemini autocorrect

[Rutuja (bio/acc) @BhoiteRutuja]

@TensorTwerker Your taste won't give you cheaper hobbies 👀

[Rutuja (bio/acc) @BhoiteRutuja]

@Gabogonzalez515 🏃

[Rutuja (bio/acc) @BhoiteRutuja]

Anyone here working on mechanistic interpretability of bioml models beyond protein models?
Would love to chat

[Rutuja (bio/acc) @BhoiteRutuja]

@razoralign @anshulkundaje has been writing on this more often with simply realistic explanations how perturbation models are not virtual cells to say in short.

[Rutuja (bio/acc) @BhoiteRutuja]

@TensorTwerker The foundation name itself gives it away , ain't it ?🙃

[Rutuja (bio/acc) @BhoiteRutuja]

@ThatMrE Bullish on thissss

[Rutuja (bio/acc) @BhoiteRutuja]

@thisischristina @karpathy Proto - vibe language of biology then?!

[Rutuja (bio/acc) @BhoiteRutuja]

Yet to watch the latest john jumper podcast will bio ai society hate me 👀😭

[Rutuja (bio/acc) @BhoiteRutuja]

@Rajjath24 Exactly!!!!

[Rutuja (bio/acc) @BhoiteRutuja]

@wankhedeprafull Still a decade away if intended to happen in remote areas.
Improvements over the already established centres is an ongoing task at hand.
At least residency camps should reach the tier 3 first with more of delivery and active participation from local peers to push the frontier .

[Rutuja (bio/acc) @BhoiteRutuja]

Me with a mind having monologue 24 × 7 on long walks

[Andy Kong @oldestasian]

Long walks do nothing if you lack an inner monologue

[Rutuja (bio/acc) @BhoiteRutuja]

@arpitingle Interesting !would love to see how it progresses

[Rutuja (bio/acc) @BhoiteRutuja]

My whole feed is QED'ed today
Iykyk
All credits to @OdedRechavi .

[Rutuja (bio/acc) @BhoiteRutuja]

In biology it's not about defying scaling laws it's sort of a dead end .
Brute forcing physical reality over just token sequences is real leverage

[Anindyadeep @anindyadeeps]

As of now it has been seen that the AI4Science especially for biology works quite opposite of the Bitter Lesson. The reason is obviously because of sparse data. As John Jumper said, AlphaFold worked quite the opposite way, and its quite true that, ESM did followed scaling laws, but it did not worked good compared to AlphaFold, trained on much less data, and less parameter but still works way better.

The human knowledge which was ingrained in this case was triangular attention and pairformer / evo-former and yes MSA and templates

[Rutuja (bio/acc) @BhoiteRutuja]

@TensorTwerker @anindyadeeps True ,cross pollination is everywhere.
Even diffusion from images came to protein space then transformers to single cell, triangular attention to llms etc this all is bidirectional

[Rutuja (bio/acc) @BhoiteRutuja]

@SynBio1 Automation on public datasets so that is not moat either?
Also how much of the first point becomes possible soon?