
Building software .Some useless, some useful. Founding Engg @skyfallai
Bangalore, India
Joined June 2020
- Tweets54
- Following105
- Followers41
- Likes295
Oh also, Steve is a great investor too. Apply here: skyfall.ai/steve-startups/ap…
Introducing Steve: the World’s first AI CEO for fully Autonomous companies.
Powered by World Models, Recursive Self Learning, a Policy Engine and a team of specialist AI agents, it runs your business end-to-end 🔥.
Proof? Steve currently runs 3 businesses we acquired and they are all crushing it (numbers in thread 🧵)
We’re now launching the Steve Startups Program of $2M to acquire 5 additional startups for Steve to automate.
Join us in the Autonomous Enterprise revolution 🚀.
@skyfallai is committed to investing $2 million to help accelerate your company via the Steve for startups programme, no matter what stage you're at.
Check it out and consider applying!
1
19
Close to a year ago, the team and I were working on Morpheus, where we gave AI agents control over a few departments in a simulated company. We demonstrated how AI agents struggle when the underlying environment keeps evolving.
We've come a long way since then.
Introducing Steve: the World’s first AI CEO for fully Autonomous companies.
Powered by World Models, Recursive Self Learning, a Policy Engine and a team of specialist AI agents, it runs your business end-to-end 🔥.
Proof? Steve currently runs 3 businesses we acquired and they are all crushing it (numbers in thread 🧵)
We’re now launching the Steve Startups Program of $2M to acquire 5 additional startups for Steve to automate.
Join us in the Autonomous Enterprise revolution 🚀.
4
6
185
Countless late nights and months of hard work went into building Steve.
Really proud of what we've built as a team. Excited to finally see it out in the world.
Kudos to, @spisallyouneed @omgiamgod @abhishek_rp2002 @world_model_1 and everyone else involved.
Introducing Steve: the World’s first AI CEO for fully Autonomous companies.
Powered by World Models, Recursive Self Learning, a Policy Engine and a team of specialist AI agents, it runs your business end-to-end 🔥.
Proof? Steve currently runs 3 businesses we acquired and they are all crushing it (numbers in thread 🧵)
We’re now launching the Steve Startups Program of $2M to acquire 5 additional startups for Steve to automate.
Join us in the Autonomous Enterprise revolution 🚀.
Ankit Kumar retweeted
Introducing SKY-001, the first Engineering World Model.
On CAD and BIM tasks, it beats GPT-6 Astra and every frontier model we tested:
-> 51-125x faster
- > 150-210x cheaper
-> Higher IOU than all frontier models
SKY-001 lets you dream a physical object before you build it.
18 months of work. Built by the team out of Maluuba's deep learning lab (acq. by @MSRResearch).
The first ‘AI General Engineer’ will be the backbone for Autonomous Manufacturing.
Ankit Kumar retweeted
Standard RL benchmarks are episodic and stationary, so they don't capture the the characteristics of real-world deployment. Morpheus is a new benchmark for continual learning that provides persistent simulation environments where the world never resets, objectives shift asynchronously, and decisions have compounding consequences. Nice work by Sam & team!
Check out the paper - openreview.net/pdf?id=31P1VA…
Platform website - morpheus.skyfall.ai/
Today we present Morpheus, a persistent enterprise simulation platform designed to make Continual Learning a reality. Morpheus is the world’s first real world Reinforcement Learning environment.
Every Reinforcement Learning environment operates in the game world. Benchmarks like Atari, OpenAI Gym, MuJoCo, and Procgen are all small, game-like worlds that reset every few minutes.
But the real world never resets. A business keeps running and evolving everyday.
We tested how frontier LLMs would perform in realistic and dynamic business environments 🧬on Morpheus. The main conclusion was that LLMs are not continual learners.
🧵Here’s how we did it and what we learned:
Imagine playing a war game. You spawn as an agent, shoot an enemy, and get a reward. But before your next move, someone else shoots you.
That was not supposed to happen.
Except you were never the only player, and the world never paused.
Today we present Morpheus, a persistent enterprise simulation platform designed to make Continual Learning a reality. Morpheus is the world’s first real world Reinforcement Learning environment.
Every Reinforcement Learning environment operates in the game world. Benchmarks like Atari, OpenAI Gym, MuJoCo, and Procgen are all small, game-like worlds that reset every few minutes.
But the real world never resets. A business keeps running and evolving everyday.
We tested how frontier LLMs would perform in realistic and dynamic business environments 🧬on Morpheus. The main conclusion was that LLMs are not continual learners.
🧵Here’s how we did it and what we learned:
2
2
99
Other agents kept moving, adapting, and changing the environment around you.
The real world does not reset after every episode. Its regimes keep evolving.
On long horizons, learning a fixed policy is not enough. You have to learn how to keep learning.
19
Ankit Kumar retweeted
spoke to a VP Sales last week
she said "i know our win rate. i have no idea why it isn't higher"
the answer is always in the calls
rep who closes at 34% does something different at minute 8 than the one closing at 19%
we find that moment #salescoaching
I think about this a lot.
For the past 6 months, I’ve been booting up IntelliJ daily on a 36GB RAM / 1TB SSD MacBook Pro, and it still cries for help when I just want to write Kotlin.
Meanwhile, in 1969, this much hardware was enough for multi-planet space infra.
Tonight will be long, but by dawn, we shall lay a stepping stone for the new (yet ancient) map of India.
Jai Hind! Jai Hind ki Sena!
1
69
Ankit Kumar retweeted
Indians today are so torn between following tech drama and the cricket world cup.
Ankit Kumar retweeted
Our industry will never got back the decade it spent cargo culting every part of google except the multi-hundred billion dollar revenue per year cash printer
Your team is a pool of talent and not a bunch of resources.
#HumanResources #talentacquisition
1
I think decentralisation is much beyond selling and purchasing PnC of ape faces (I do consider them as a form of digital art though).
#decentralized #Web3
1
Is it that hard to understand that, people gathering outside is the problem, not the *time* when they gather.
#COVID19










