Prasanna

Founding Product Manager @ Skyfall | Computers, Products, F1...

[Sam Pasupalak @spisallyouneed]

Yesterday, we came out of stealth at @skyfallAI in Forbes and today we're excited to announce that Skyfall AI is featured in the @WSJ's newsletter.

We believe in a utopian world that will be achieved over the next decade, where humans will work in their own creative interests in space exploration, sports, arts or philosophy while machines will do all the mundane operational tasks. We want to free humanity from all the monotonous boring operational tasks that they spend their daily lives doing.

The only way to achieve this vision is via creating autonomous enterprises, i.e., companies that can run by themselves with minimal human intervention. After 15 years of building in the AI space, one thing is certain. LLMs are static systems that fail at

- Long horizon planning tasks
- High consequential decision-making
- Sample efficiency
- Dynamic environment adaptation

These skills are required by a business to survive and grow. Unfortunately, LLMs are not designed for the real world constraints of running an organization.

Our solution is Enterprise World Models grounded in Continual Learning and World Modeling. We're taking a fundamentally different approach because we have always believed in 0 to 1 research, and not incremental n+1 RL fine tuning.

We are acquiring SaaS businesses to prove our thesis for creating AI systems for the post scarcity world. This is a call to all SaaS founders to submit your company if you’re interested (🔗 in comments)

Thank you Sarah @sarahklearman for covering our story on @WSJ

[Prasanna @world_model_1]

@spisallyouneed @Yoshua_Bengio Proud to be part of this amazing team. Just getting started 🚀

[Sam Pasupalak @spisallyouneed]

Around 15 years ago, Kaheer Suleman and I started Maluuba with the vision of building Universal Turing Machines. Back then, we were one of the pioneer research labs building foundational modules of deep learning and Reinforcement learning with Turing award winners @Yoshua_Bengio and @RichardSSutton. There was no AI hype then, as there is now. Those were the good old days of doing pure scientific research unlike the n+1 research that we see today. Maluuba was later acquired by Microsoft and became @MSFTResearch Canada.

Grounded in that same mission, we're finally ready to introduce @skyfallai to the world after a lot of experimentation in stealth over the last year. We're a frontier neo lab building the first Autonomous Enterprise by moving beyond the current LLM paradigm.

For the last 5 years, the foundational model market has relied on a single paradigm: scaling laws for LLMs - more data, more compute, bigger models. However, the real world is messy and much more complicated. In order to achieve our long term vision, the next generation of AI models requires a different approach. Skyfall is solving the hardest open problems in frontier AI: long-horizon planning, data inefficiency, and brittle performance in dynamic real-world environments.

To achieve the team's vision of a completely autonomous enterprise, the team is developing a next-generation frontier model (Enterprise World Models) using Continual Learning and World Modeling. Enterprise World Models can simulate the multi-layered consequences of strategic business actions. Our approach unlocks a new category in the foundation model market. To prove it, we're introducing Morpheus, a Continual Reinforcement Learning platform for AI researchers.

I'm building this company with the people I trust the most: my longtime friend Kaheer Suleman (prev. Co-Founder of Maluuba) and my brother @omgiamgod (prev. YC founder). Sumit and Kaheer are the first principles thinkers I can trust to go to the end of the world with to achieve the mission impossible together. Together with a stellar team of 25 researchers and engineers, we're pushing a new frontier in AI forward.

We unpacked our long term vision in today’s Forbes feature 🔗- read it to see what we’re building toward. Thank you so much Victor Dey for the interview.

To achieve our goal of enterprise world models, we are soliciting bids to acquire small SaaS startups (up to $1M) and fully automate them. If you’re interested, submit your business here: https://t.co/w8ayopZFLH

Finally, thank you to our investors and advisors for believing in our vision since day one: @Fidelity , @sk121 (@touringcapital), @karam_n and @chrisarsenault (@inovia), @morgan_blumberg (@M13Company ), @stephpalmeri (@NextViewVC ), and @michaellitt and @mmccauley (@GarageCapital ), @jennydhe, @fchollet @NaveenGRao and so many others for supporting us in this journey.

[Prasanna @world_model_1]

Do you remember when you joined X? I do! #MyXAnniversary

[Prasanna @world_model_1]

@georgeclem OS

[George Clements | Clients From Ads @georgeclem]

I JUST wrote down my entire system for building a Meta ad account that runs in 2026...

And I'm giving it away free.

It's called the Meta Ads OS, and inside it is:

> Why the old way of running Meta ads is dead.

> The simple 2-campaign structure that stops your new ads from dying.

> How to build creative the algorithm actually rewards, and why volume alone doesn't work anymore.

> A tracking setup that decides whether Meta finds your buyer or burns your budget.

> The exact testing cadence to keep your account fresh at every spend level.

> How to build a winning offer.

I've scaled 2 agencies past 7 figures on paid ads, and I have eyes into 30+ agency ad accounts.

This is the exact system underneath all of it.

Comment "OS" and I'll send it over.

(Must be following + RT for priority access)

[Skyfall AI @skyfallai]

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:

[Prasanna @world_model_1]

@skyfallai If this holds up, it could expose one of the biggest gaps in today’s AI narrative: we’ve optimized for intelligence on static benchmarks, not adaptation in persistent worlds. That’s a much harder and arguably more important problem.

[Prasanna @world_model_1]

New AI paper: 1) 360 panoramas break standard 3DGS partitioning 2) PanoLOG uses sky modeling + monocular depth for stronger geometry 3) G2PS makes block-parallel outdoor reconstruction scale, with a new Pano360 benchmark. https://t.co/C9Dmts70Fa

[Prasanna @world_model_1]

LongE2V: 1) one model handles event video reconstruction, prediction, and interpolation 2) new rollout tricks reduce long-sequence drift 3) it beats prior work on temporal coherence and zero-shot transfer. https://t.co/kEOU8152de

[Prasanna @world_model_1]

Workflow as Knowledge: 1) workflows, runs, and context snapshots become persistent knowledge objects 2) it splits deterministic derive steps from LLM infer steps 3) this makes LLM workflows inspectable, resumable, and reviewable. https://t.co/BlcBX3vAA4

[Prasanna @world_model_1]

Stop treating UMAP as just a 2D plotter. New paper: 1) mines UMAP's hidden kNN graph 2) uses PageRank, k-core, and clustering coefficients for sensemaking 3) shows graph analysis can rival purpose-built methods. https://t.co/i41VmYcoLH

[Prasanna @world_model_1]

New paper: ZipDepth. 3 takeaways: 1) shrinks zero-shot depth to a 6.1M model 2) distills from a foundation model across multi-domain data 3) runs in real time on edge devices while nearing models 50x larger. https://t.co/uEul27zs3q

[Prasanna @world_model_1]

New paper: Wat3R. 3 takeaways: 1) adapts air-trained 3D recon to underwater scenes 2) learns from unlabeled real underwater video 3) adds Water3D and sets new depth + point-cloud baselines. https://t.co/LbXsG4MBFG

[Prasanna @world_model_1]

Most AI tutor discourse is still survey theater. New paper on 77,543 students: 1) assistants are already part of study routines 2) usage shifts by age, degree, and study mode 3) product decisions should follow behavior data, not chatbot surveys. https://t.co/80A8DGO6tW

[Prasanna @world_model_1]

Compression shouldn't be an afterthought. SLORR: 1) adds stateless low-rank regularization during training 2) keeps vision-model training overhead under 8% 3) holds LLM pretraining overhead under 1% while preserving compression. https://t.co/5vE3hoTKDo

[Prasanna @world_model_1]

Most VLM benchmarks dodge incidents. AUTOPILOT-VQA: 1) centers crashes and near-misses in dashcam QA 2) tests weather, road layout, signage, impact, avoidability 3) pushes eval toward safety-grounded reasoning. https://t.co/oScMQsSP8T

[Prasanna @world_model_1]

Video models won't reason by magic. OpenCoF shows: 1) frames can carry reasoning, not just output 2) a 17K reasoning-video dataset across 11 tasks lifts scores 3) visual + text reasoning tokens improve spatial/temporal reasoning. https://t.co/36ixhEzclX

[Prasanna @world_model_1]

AI scientists are worse at research lineage than people assume. IG-Bench shows: 1) lineage reasoning is hard 2) structured lineage context reshuffles rankings, not universal gains 3) the best system hits just 27.3% exact accuracy. https://t.co/O96Sv6JG2d

[Prasanna @world_model_1]

Benchmarks for agents are still too fake. UniClawBench: 1) tests 400 bilingual real-world tasks in live containers 2) splits 5 core agent capabilities 3) shows rankings shift across frameworks. https://t.co/3Y8UYp1QQV