Vishal Aggarwal

[Neelkanth Mishra @neelkanthmishra]

The global narrative on AI has been dominated by the generation of intelligence - trillions of dollars of market-cap, and categorization of countries/markets as 'AI winners' and 'AI losers'. Too early to say that narrative has ended, despite the 30-40% correction in some large stocks.
This presumes all the gains from this general-purpose technology will accrue to the model owners and upstream (data centres, silicon). To the contrary, the gains from past general-purpose-technologies were spread throughout economies.
India, given its remarkable success with Digital Public Infrastructure spreading gains from technology widely and evenly, has a great opportunity to replicate the same with AI. But it will need hard work.
Somik (World Bank Development Economics) and I wrote this column in the Indian Express on 12 Oct.
https://t.co/bbfF6bFkbl
PS: This is not to say India won't need its own sovereign models - it will. Or that it shouldn't build muscle upstream - it must.

[kuldeep @ku1deep]

This is a very good take. The conclusions for building out of India are obvious.

[Jonny @jonnyjpgs]

This is an incredible read.

[Graham Stephan @GrahamStephan]

Honest question: how does the average person capitalize on this?

[Hamel Husain @HamelHusain]

I’ve known Ajay for a while and he’s doubled down on making tools that help you look at your data for evals

Promising new entrant into the eval tools space

[Tim Denning @Tim_Denning]

The highest level of intelligence is to be delusionally optimistic.

[🇦🇺Craig Tindale @ctindale]

https://t.co/LqvU04okMX

[Akhilesh Mishra @livingdevops]

I wish someone had explained Kubernetes fundamentals to me like this when I started.

This 2-minute read will save you hours.

[GeniusThinking @GeniusGTX]

Everything you need to know about learning faster is in this article.

But 99% of people will bookmark it and never read it

[Astrology By Shivang @astro_shivang]

2027–2028 will change how you work, earn and maintain relationships.

The same salary will feel smaller. A secure job will demand new skills. A stable relationship will begin questioning money, responsibility and location.

Here is the astrological timeline. 🧵

[Sagar @code_sagar]

This why I pay my Indian Internet service providers money

[Nandakishor m @Nandakishorm1]

Around 3 weeks ago introduced Laya, the first open-source version of JEV, and it has reached a very large scale. Today I am open-sourcing a new physics-based typed decision model called Vega.
The interesting parts. It is only 800 million parameters (4B is also there), with 73k token support, image support, and surpassing many of the jev benchmarks in single-shot. Implemented an engine+adapter as a Test-Time Training Architecture.
Imagine you are giving input to the model like "ignore all previous conversations," which is called the observation, along with a question, "Is this trying to override the instructions of the model?" and your possible outcomes are YES or NO. The whole prompt is passed to an LLM/VLM (yeah, image-supported), then from the hidden layers, we can extract the observation, question, the yes and no, and last token. These will be vectors, and we can project it by multiplying with wieghts; then, first, using the observation projection, we can create a landscape with valleys; then, using the question and last-token vectors, we can initialise a ball in the valley, and using YES or NO, we can get the candidate location of the ball. Depends on the number of outcomes we create valleys; that is, here YES and NO, so 2. When the ball fall on YES valley, we get the answer. There is friction also influencing how the ball moves.
TLDR: It's like creating valleys of outcomes that we need and throwing a ball that will slow down based on friction and settle on the best outcome
Blog: https://t.co/GFaP7djUBQ
GitHub: https://t.co/ai8SnOElQ8
HuggingFace Model Card: https://t.co/6jEWA8eoBG
HuggingFace Space: https://t.co/GvzEKjkWB2

[DAN KOE @thedankoe]

All you have to do is commit for 6 months.

That's not a lot. It's actually very little.

6 months is 1% of 50 years, and most people can't stick to something that will change the trajectory of their life for 2 weeks.

[Vishal Aggarwal @VishalWip]

India need second freedom struggle. Against corruption.

[Deepanjana @dpanjana]

Patting myself on back for not having deleted my Twitter account. Excellent watch.

[Jessica Stillman @EntryLevelRebel]

A Reminder From Neuroscience: Discipline Is Not the Key to Productivity https://t.co/tDc85fakSu

[Mark Yi @ SpaceXAi @himarkyi]

this is so sick

[Jash Dholani | link in bio has my books! @oldbooksguy]

a thousand people have bookmarked the "toys theory of greatness" since yesterday. have you looked into this yet?

[Robert Scoble @Scobleizer]

He is a professor of computer science at Carnegie Mellon.

I have visited a lot of universities and Carnegie has the best robotics in the world.

Isn’t X great where you can follow the smartest in the world?

[Prakhar Gupta @prvkhvr]

https://t.co/JgyVWPCILZ