Abhishek Singh

[Abhishek Singh @MnaAbhi]

@AnkurGarg9 @EurekaforbesIN I also bought their AMC in Bangalore and no one visits on time. When I call their call center, they cancel the old request and create a new one with a later date. Very pathetic service! Would have been better if I would have got this from urban company or somewhere else!!

[Thomas Wolf @Thom_Wolf]

we've just passed $1,000,000 in sales for Microduck

[Prayank Swaroop @prayanks]

I have been trying to learn Kannada for a very long time, but I haven't been able to get to a course / book which uses the structure of Hindi to teach Kannada - every course / book / offline classes use English .. but Hindi and Kannada have a very similar language construct - Subject Object Verb structure - so using Hindi to learn Kannda is more natural - at least for me. Hopefully others will also find it useful.

This is one of my fun AI projects.

Showcasing https://t.co/kuBTQ1WBvx

Thanks to:
Hermes agent - I'm using it maintain the course regularly using Hermes Agent (@NousResearch / @Teknium) - I can literally add chapters while in a cab.

@grok / @elonmusk - the work horse for this project is Grok4.5 xHigh

@Microsoft for Edgetts - https://t.co/YLI1wPmRaD for creating all the audio examples

@mattpocockuk - the whole course was created using his teach skill

@github - for enabling hosting so everyone can access it

[Tinker @tinkerapi]

Sorting which financial docs are worth an analyst's time is surprisingly hard for frontier LLMs. With an expert-labeled dataset and on-policy distillation, Bridgewater fine-tuned a model to do it reliably and cheaply.
https://t.co/gyYzXq15zd

[Garry Tan @garrytan]

https://t.co/XTqXqWuQn5

[dominic dyer @domdyer70]

Happy world Donkey Day

[Tanay Kothari @tankots]

We offered 5 people a Porsche 911 GT3 RS if they could get @WisprFlow to make a mistake

It's the fastest and most accurate AI voice dictation app that's 3x more accurate than ChatGPT, Claude, or Siri.

Today, we’re finally launching on Android. Download now: https://t.co/TJhnUhDSLv

As a part of the launch, we’re giving away 6 months of Wispr Flow Pro for free.

Like, retweet and comment ‘Wispr Flow’ to get it. Enjoy.

— Written with Wispr Flow

[Abhishek Singh @MnaAbhi]

@calcsam Book

[Patrick Brauckmann 🕉️ @vonbrauckmann]

Remember, the lamps are different but the light that shines is One.

[Abhishek Singh @MnaAbhi]

@orchidsapp If you could help me with credits please..

[Abhishek Singh @MnaAbhi]

@mischa_vdburg Yes pls

[Abhishek Singh @MnaAbhi]

@kadalcippi Mine please

[Abhishek Singh @MnaAbhi]

@kadalcippi Hello, whenever you have some time, please help me with my astro reading. Would really appreciate it, thanks.

[Tai Lopez @tailopez]

A lot of wealth comes from just being organized. It’s basically impossible to make money when your brain is scattered like most humans.

[Nick Dobos @NickADobos]

Telegram CEO is never getting out
Gonna rot in jail for the rest of his life

If you do any kind of messaging encryption at scale, shitty people will use it, and if you don’t bend the knee to government, expect most countries on the planet to jail you

Privacy is dead

[Historic Vids @historyinmemes]

Telegram CEO Pavel Durov: "I don't own any big assets."

[Andrej Karpathy @karpathy]

These 94 lines of code are everything that is needed to train a neural network. Everything else is just efficiency.

This is my earlier project Micrograd. It implements a scalar-valued auto-grad engine. You start with some numbers at the leafs (usually the input data and the neural network parameters), build up a computational graph with operations like + and * that mix them, and the graph ends with a single value at the very end (the loss). You then go backwards through the graph applying chain rule at each node to calculate the gradients. The gradients tell you how to nudge your parameters to decrease the loss (and hence improve your network).

Sometimes when things get too complicated, I come back to this code and just breathe a little. But ok ok you also do have to know what the computational graph should be (e.g. MLP -> Transformer), what the loss function should be (e.g. autoregressive/diffusion), how to best use the gradients for a parameter update (e.g. SGD -> AdamW) etc etc. But it is the core of what is mostly happening.

The 1986 paper from Rumelhart, Hinton, Williams that popularized and used this algorithm (backpropagation) for training neural nets:
https://t.co/f52IcDNitR
micrograd on Github: https://t.co/GaTd16jRnB
and my (now somewhat old) YouTube video where I very slowly build and explain:
https://t.co/EPGG6kd5Yz

[Dylan Freedman @dylfreed]

New open source OCR model just dropped! This one by Microsoft features the best text recognition I've seen in any open model and performs admirably on handwriting.

It also handles a diverse range of vision tasks. You can play with it here: https://t.co/gCLes81Th3

[Abhishek Singh @MnaAbhi]

@thekabirgrover @IAF_MCC @PMOIndia

[Tereza Tizkova @tereza_tizkova]

✨ AI agents database updated ✨
🔗 Link in comment

What's new:
🤖 3 open-source Devins
🤖 @InvictaAI
🤖 @ask_pandi
🤖 GPTSwarm by @MingchenZhuge
& more.

Check ~200 agents on the list, such as @HyperWriteAI, @Kompas_AI, @gitwitdev, Maige by @RubricLabs, @OpenInterpreter, @superagent_ai