LLM that loops instead of Doing Chain-of-Thought @bycloudAI
LLM that loops instead of Doing Chain-of-Thought  @bycloudAI
Uploaded July 2026 | Updated September 2026, 1 week ago
Need to fine-tune a model without the hassle? Try out Crusoe's serverless fine-tuning today! crusoe.ai/contact-sales/serverless-preview?utm_source=bycloud&utm_medium=influencer&utm_campaign=serverlessfinetuning

my latest project: Intuitive AI Academy
We just wrote a new piece on Optimization!!
https://intuitiveai.academy/
limited time code "LOCKIN" for 35% off yearly plan


Chain-of-Thought is ugly, so what if we remove it? In this video, I am diving into the latest architecture experiment that is Looped Transformer. So LLMs think with words, but here Looped Transformer simulates thinking by looping the same few layers.


My Newsletter
mail.bycloud.ai

My Patreon
patreon.com/c/bycloud


Loop, Think, & Generalize: Implicit Reasoning in Recurrent-Depth Transformers
[Paper] arxiv.org/abs/2604.07822

A Mechanistic Analysis of Looped RLMs
[Paper] arxiv.org/abs/2604.11791

Parcae: Scaling Laws For Stable Looped Language Models
[Paper] arxiv.org/abs/2604.12946

Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation
[Paper] arxiv.org/abs/2507.10524



Try out my new fav place to learn how to code scrimba.com/?via=bycloudAI

This video is supported by the kind Patrons & YouTube Members:
🙏Spam Maj, Alex, Chris LeDoux, DX Research Group, Poof N' Inu, Deagan, Robert Zawiasa, Ryszard Warzocha, Tobe2d, Louis Muk, Akkusativ, Kevin Tai, Mark Buckler, NO U, Tony Jimenez, Ângelo Fonseca, jiye, Anushka, Asad Dhamani, Binnie Yiu, Calvin Yan, Clayton Ford, Diego Silva, Etrotta, Gonzalo Fidalgo, Handenon, Hector, Jake Disco very, Michael Brenner, Nilly K, OlegWock, Daddy Wen, Shuhong Chen, Sid_Cipher, Stefan Lorenz, Sup, tantan assawade, Thipok Tham, Thomas Di Martino, Thomas Lin, Richárd Nagyfi, Paperboy, mika, Leo, Berhane-Meskel, Kadhai Pesalam, mayssam, Bill Mangrum, nyaa, Toru Mon, Lame Plane, Matej Macak, Len Mo, saylikhapekar, ZyanSheep, THEVIERAOS, Ricardo Raphael Corona-Moreno, superchordate


[Discord] discord.gg/NhJZGtH
[Twitter] twitter.com/bycloudai
[Patreon] patreon.com/bycloud
[Business Inquiries] bycloud@smoothmedia.co
[Other Inquiries] bycloudai@gmail.com
[Profile & Banner Art] twitter.com/pygm7
[Video Editor] @aduckchicken2
Manim Animations created with Manimate manimate.ai
[Ko-fi] ko-fi.com/bycloudai
LLM that loops instead of Doing Chain-of-ThoughtWhat Is Yann LeCun Cooking? JEPA Explained SimplyThe REAL AI Architecture That Unifies Vision & LanguageAnthropic Just Proved Reasoning AIs Would Silently CheatOpenAI o1s New Paradigm: Test-Time Compute ExplainedDeepSeek V3.2 Just Broke SoTA Again… But How?AI Tech CEOs In A Cage Fight??? [The AI Timeline #6]The AI Bubble Will Be Impossible To Pop Thanks To... AI Video?!Kimi K2.5 Brought Us 3 brand NEW LLM Frontier!?Forcefully controlling AI’s ideology doesnt really work... (yet?)OpenAIs Sora Got Some New Results... And Theyre Way Too GoodMetas Reasoning AI Model... Reasons Without Using Words?
bycloud |

LLM that loops instead of Doing Chain-of-Thought

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER