Bhargav Modha

ML Engineer @labs_cactus

NLP || Backend || Diploma CSE MU'21 || ITNU CSE'24

Tweets about Programming and return (Random.Tweet.randomtweet());

[Bhargav Modha @TheBhargavModha]

💔💔

[Bhargav Modha @TheBhargavModha]

“Fake it till you make it” backfires when it becomes a cover for not admitting what you don’t know.

The fastest learners aren't the best pretenders.

They’re comfortable saying: “I don’t know this yet.”

[Pedro Domingos @pmddomingos]

AI now is like databases in the 60s: widely used but an ugly pile of hacks, waiting for Ted Codd to bring order to it with the relational model.

[Bhargav Modha @TheBhargavModha]

In the future, it would be literally a "skill" issue.

[Bhargav Modha @TheBhargavModha]

@SidRangnani @RailwaySeva

[Balaji @balajis]

Too many are interested in “tech” (by which they mean money) but they aren’t interested in tech (by which I mean math).

[vicki @vboykis]

writing code for a living is fun because you spend all day solving little mysteries that have you questioning your competence as a human being and it turns out that the issue was like you forgot to include the slash in the path and then when you fix that the next thing breaks

[Hiten Shah @hnshah]

Growth doesn’t feel like growth while it’s happening.

From the outside, people see a story. Inside, it’s chaos that hasn’t found its form yet.

Speed magnifies everything. The faster you move, the louder the cracks sound. Teams outgrow processes that barely existed. Systems that worked last quarter start fighting each other.

Great companies aren’t clean at the start. They are just better at surviving the disorder long enough for patterns to appear. The mess is what progress looks like before it gets a name.

Scaling means everything is temporary. Roles shift. Priorities blur. Decisions that made perfect sense yesterday look wrong today. It can feel like confusion. It’s evolution in fast-forward.

People who struggle expect stability too early. They want the manual before the machine exists. The ones who thrive can hold two truths at once. That things feel broken, and that the breaking is how you find what works.

You can tell a strong team by how their environment handles the in-between. Weak teams chase polish. Strong ones chase truth. They shorten the feedback loops and move through the noise without pretending it’s quiet.

Fast growth always looks messy because clarity lags behind creation. You can’t optimize what you haven’t built yet.

If it looks perfect from the inside, you’re probably late.

[Jerry Tworek @MillionInt]

My daily mantra

Wake up early
Be optimistic and courageous
Fix whats broken
Simplify whats giving you trouble
Be more afraid to aim too low than to fail
Nothing is ever perfect
Speed of iteration beats prior knowledge

[Jerry Tworek @MillionInt]

AI should be the most intuitive technology to use, but we’re still not there.

For now you need to learn how to use every new model

[fish @fishPointer]

trillions of dollars spent on AI and the only interface is still a chatbot. still no branching canvas, no mindmap, no dimensionality, no forking, no inversion, no merging. no custom functions. no style and no creativity

[Bhargav Modha @TheBhargavModha]

Grateful, always.

[Varun Mayya @waitin4agi_]

3.6

The number after 3.5 is 3.6

[Matt Henderson @matthen2]

to tell if a maze is solvable, just hang it by its corners! The first maze stays in one piece, so there is no path from the entrance at the top to the exit at the bottom. The second maze splits apart along the solution.

[Philipp Schmid @_philschmid]

Llama 3 released! 🚨🔔@AIatMeta just released their best open LLM! 👑🚀 Llama 3 is the next iteration of Llama with a ~10% relative improvement to its predecessor! 🤯 Llama 3 comes in 2 different sizes 8B and 70B with a new extended tokenizer and commercially permissive license! ✅

Blog: https://t.co/VrceFpjI1o
Models: https://t.co/SdkOeWURuM

New and improvements to v2✨:
🔠 Trained on 15T Tokens & fine-tuned on 10M human annotated samples
🧮 8B & 70B versions as Instruct and Base
🚀 Llama 3 70B best open LLM on MMLU (> 80 🤯)
🧑🏻‍💻 Instruct good at coding 8B with 62.2 and 70B 81.7 on Human Eval
✍🏻 Tiktoken-based tokenizer with a 128k vocabulary
🪟 8192 default context window (can be increased)
🧠 Used SFT, PPO & DPO for alignment.
💰Commercial use allowed ✅
🤗 Available on @Hugging Face
🤝 1-click deployments on Hugging Face, Amazon SageMaker, Google Cloud
🔜 more model sizes & enhanced performance

Massive kudos to Meta for continuing its commitment to open AI. Honored to partner with Joe and team! 🤗 The gap is melting. 🧊

[Manish Pandey (मनीष पाण्डेय) @join2manish]

the closer you are to death, the more alive you feel - every second that passes in our life is taking us near to our deaths

[Bhargav Modha @TheBhargavModha]

Heard it. Over. Everything happens for a reason.

[Bhargav Modha @TheBhargavModha]

Increasing Dimensionality

[Bhargav Modha @TheBhargavModha]

Soon. 🖨️

[Bhargav Modha @TheBhargavModha]

Imagine a curious child eager to explore the vast realm of knowledge. Would they plunge into the boundless internet, absorbing massive volumes of information randomly?

[Sudeepa Kolli @i_sudeepa]

Job search routine: 3 cold emails, 2 tailored cover letters, 10 career site searches, 15 easy applies, career fairs biweekly, 1-hour exercise + walk, 7 hours sleep. More importantly, track your applications and send follow-ups.

A brief thread:

[Bhargav Modha @TheBhargavModha]

@yashginoya_ Commenting for better reach
It truly hurts 💔

[Milan Chheda @milanchheda]

Paperpal Copilot: Generative AI for Research: Rewrite/Generate Titles, Outlines, Abstracts & More

https://t.co/0tCWEEjCME

#researchers #AcademicTwitter #academicwriting #GenerativeAI