sameer

co-founder and CTO, iRage
loves playing sports, coding.

[sameer @the_sameerk]

@singhofstanplus IIRC, back in 2008, one of my yahoo! colleague started @exotel because he got too overwhelmed setting up phone connections for his initial startup office :)

[Prof. Krishnamurthy V Subramanian @SubramanianKri]

I’ve long maintained (since the Global Financial Crisis to be precise) that the unraveling of the West has started.

Mark my words: Future generations will ask why their great grandparents didn’t have the wisdom to see where the West was headed & chose to leave India to move there.

[sameer @the_sameerk]

@vinodchendhil Interestingly, we did recently got melanoma skin cancer vaccine, with llm help.
It takes stages, and trust in llm models to share IP as well.
Navier stokes was solved by stealing IP rights of two mathematicians working on the problem for more than decade.

[sameer @the_sameerk]

@SriniVega https://t.co/GwgaMh8VtX

[sameer @the_sameerk]

@thsottiaux @FelipeFr1702 Shastra

[sameer @the_sameerk]

We are limited only by our own imagination.

[Ornith @ornith_]

Aloha! 🌺Introducing Ornith-1.5, a family of open-source LLMs spanning 9B Dense, 35B MoE, and 397B MoE, trained with self-improving strategies.

It achieves state-of-the-art performance among open-source models of comparable size and delivers performance comparable to Claude Opus 4.8 across reasoning, agentic, and coding tasks:
✅Terminal-Bench 2.1 (86.1)
✅SWE-Bench (86 on verified, 65.1 on pro, 79.6 on Multilingual)
✅DeepSWE (56)
✅HLE (44.6)
✅ClawEval (81.4)
✅Tool Decathlon (71.2)

Ornith-1.5 takes a major step toward training foundation models through end-to-end self-improvement, extending the self-scaffolding strategies introduced in Ornith-1.0 into a more complete self-improvement loop: the model proposes new tasks, generates task-specific scaffolds, and produces solution rollouts for reinforcement learning, continuously creating new learning experiences from which it can improve.

All models, along with their quantized versions (FP8, GGUF, MLX, and NVFP4), have been released under the MIT License, enabling unrestricted commercial and research use.
📘Tech Blog: https://t.co/OZ63scRWLB
🤗Huggingface: https://t.co/mGJLwhrQOM

[sameer @the_sameerk]

@ttunguz Today I had a conan build failure. Qwen 3.8 figured out the issue, while opus 5 was still struggling (I cancelled it after 10 min)
Unbelievable!

[sameer @the_sameerk]

@dhh This is good. I just hope your measure hasn't become the benchmark which is maxx'ed out by the models.

[Alex Lieberman @businessbarista]

I stole this idea and now use it with every single employee.

It’s the best illustration I’ve seen of teaching someone to be high agency.

It says there are 5 levels of work:

Level 1: “There is a problem.”

Level 2: “There is a problem, and I’ve found some causes.”

Level 3: “Here’s the problem, here are some possible causes, and here are some possible solutions.”

Level 4: “Here’s the problem, here’s what I think caused it, here are some possible solutions, and here’s the one I think we should pick.”

Level 5: “I identified a problem, figured out what caused it, researched how to fix it, and I fixed it. Just wanted to keep you in the loop.”

Using this framework, here’s what I say to every new employee…

You will live at Level 4 from Day 1 and as we build trust you will rise to Level 5.

Being high agency doesn’t just mean tackling problems in this way. It means your entire way of working should be oriented to being a Level 4+ employee.

Plz feel free to steal it as well.

And ty @stephsmithio for the framework!

[Thorsten Ball @thorstenball]

Some thoughts on ownership I shared with the team this morning

[sameer @the_sameerk]

@_mekin More sad part is that we aren't able to retain even that 1% in India.

[sameer @the_sameerk]

@Ankit_Quant @graizada @Souvik131 may not during entry, but what about stoploss scenarios ?
Do you think you would have a better exit (lower slippage) at the beginning of `injection candle`, if you knew you are going to hit stop-loss 900 millisecond faster than other users, and avoid the stampede ?

[sameer @the_sameerk]

@AzhagesanS7 @Souvik131 That's exactly what it is.
volume = aggresive bid ( btv ) + aggresive ask (atv)

[sameer @the_sameerk]

@akgoyal You are in for a ride. Your kids will enjoy this :)

[sameer @the_sameerk]

@akgoyal Equity/Bond returns from Indian market in INR got depressed by at least 5%, and hedging cost went up.
You can make up anything in finance 😁

[sameer @the_sameerk]

@AzhagesanS7 @Souvik131 Volume = volume traded at bid + volume traded at ask, all 3 calculated from Tick by Tick data, and updated for each single trade.

[sameer @the_sameerk]

@dbytesmith @Souvik131 Exchange (NSE) gives MTBT (real-time mcast TBT) or broadcast/snapshot ( MBO,MBP at fixed interval, usually every 980 ms ).

This data is created(snapshot) from MTBT at 50ms interval.

People who know about "injection candle" can appreciate this 900 ms faster data advantage.

[sameer @the_sameerk]

the more I know, the more i know what i don't know.