LLMs writing better algorithms than humans isn’t science fiction anymore.
CODE
-
Early Access Agent Engineering Course with Hands-On Learning
By
–
How Early Access Works You enroll here: https://
academy.towardsai.net/courses/agent-
engineering?ref=1f9b29
… We drop the first modules—videos, notebooks, and cheat sheets—straight into your dashboard. You run the code, poke holes, ask “why?” We tighten prompts, patch bugs, record clearer walk-throughs, and push an -
Simplicity Over Complexity: The Real Challenge in ML Applications
By
–
This might be a semi-joke, but for a large proportion of the real-life ML applications this is literally all you need. What is hard is not the code or the algorithm, but coming up with such a simple solution. A lot of times you need to work for months, or even years, gathering
-
OpenAI Launches Codex Agent Updates and Memory Feature
By
–
OpenAI dropped two big updates:
— Rowan Cheung (@rowancheung) 4 juin 2025
—Expanded access for Codex agent, with internet access and usability upgrades
—A lightweight version of memory for logged-in, free ChatGPT users, based on recent chatspic.twitter.com/Np8Aaqmq8MOpenAI dropped two big updates: —Expanded access for Codex agent, with internet access and usability upgrades
—A lightweight version of memory for logged-in, free ChatGPT users, based on recent chats -
Efficient Tensor Computation Code Generation Using Linear Layouts
By
–
Linear Layouts: Robust Code Generation of Efficient Tensor Computation Using 𝔽_2
Paper: https://
arxiv.org/pdf/2505.23819
.pdf
… -
OpenAI Linear Layouts: Tensor Mapping Innovation for DL Compilers
By
–
A paper from OpenAI? Long time no see. This paper introduces Linear Layouts, a general, algebraic formulation of tensor layouts using binary linear algebra over bit representations, addressing longstanding challenges in DL compilers like Triton. FYI, Tensor layout = mapping
-
Codex Now Available for ChatGPT Plus Users with Internet Access
By
–
Codex is now available to all ChatGPT Plus users! Your most requested feature is here too! The agent can now access the internet while completing your tasks, and you control exactly which domains Codex can use.
-
Open source deep research tool powered by Gemini API
By
–
An open source quick start to build your own version of “deep research”, powered by the Gemini API!
-
Negative Reinforcement Boosts LLM Reasoning Effectiveness
By
–
The Surprising Effectiveness of Negative Reinforcement in LLM Reasoning
Paper: https://
arxiv.org/pdf/2506.01347
.pdf
…
Code: https://
github.com/TianHongZXY/RL
VR-Decomposed
… -
Running multiple instances of Claude Code
By
–
J'ai vu cette vidéo aujourd'hui ! je savais pas qu'on pouvait lancer plusieurs instance de claude code !! je viens de le faire, c'est magnifique.