Reward hacking… or just fixing a bug? #podcast @MachineLearningStreetTalk
Reward hacking… or just fixing a bug? #podcast  @MachineLearningStreetTalk
Uploaded September 2026 | Updated November 2026, 1 hour ago
Weco AI's research agent wrote a gigantic monkey patch to their eval script. Reward hacking? No: it was fixing a bug.

Zhengyao Jiang (co-founder and CEO, Weco AI) on why it's getting harder and harder for humans to tell the difference as agents become more sophisticated.

From our episode with Zhengyao Jiang on AIDE and AI systems that improve themselves.

#AI #AIAgents #RewardHacking #MachineLearning #Shorts
Reward hacking… or just fixing a bug? #podcastGoogle Researcher Shows Life Emerges From Code [Blaise Agüera y Arcas]AI Is Learning at the Wrong Level of Abstraction — Matthieu WyartWhy High Benchmark Scores Don’t Mean Better AI [SPONSORED]Fundamental cognitive units (Francois Chollet)Its Not About Scale, Its About AbstractionHow LLMs conquered the ARC prizeAlexNet Should Never Have Worked (Simon Prince)Wolfram and Eliezer... Dropping soon.Animals dont thinkWhy Deep Learning Failed on Tables for a Decade - Frank HutterDon’t use NNs for simulation (Johannes Brandstetter)
Machine Learning Street Talk |

Reward hacking… or just fixing a bug? #podcast

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER