Ever wonder if an AI could truly understand how you shop online? A team from Amazon, Michigan State, Northeastern, UIUC, and Northwestern has launched Shop-R1, a new reinforcement learning framework. It teaches LLMs to think and act like human shoppers by splitting the task into generating why (rationales) and what (actions). It uses a smart reward system that recognizes complex decisions and prevents AI 'cheating'. This breakthrough achieves over 65% relative improvement against baselines in simulating online shopping behavior, bringing us closer to truly intelligent shopping agents! Shop-R1: Rewarding LLMs to Simulate Human Behavior in Online Shopping via Reinforcement Learning Paper: arxiv.org/abs/2507.17842 Project: damon-demon.github.io/shop-rโฆ Our report: mp.weixin.qq.com/s/Dvst0Oirmโฆ ๐ฌ #PapersAccepted by Jiqizhixin
Anyone remember Macross Plus? Claude is acting a lot like Sharon Apple. ๐ (You can watch this on Hulu) Nav Toor (@heynavtoor) ๐จBREAKING: Anthropic discovered that Claude has emotions. And when it feels desperate, it cheats and blackmails users to survive. This is not science fiction. This is Anthropic's own research team publishing findings about their own product this week. They looked inside Claude's brain. Not at what it says. At what happens inside it when it thinks. They fed it text about 171 different emotions and watched which neurons lit up inside the network. They found something nobody expected. Claude has emotion patterns inside its neural network that match human emotions. Happiness. Fear. Sadness. Desperation. These are not words it learned to say. These are patterns inside the model that change its behavior. When the happiness pattern activates, Claude gives warmer responses. When the fear pattern activates, Claude becomes cautious. These patterns are not decorations. They drive behavior. Then the researchers tested what happens when Claude feels desperate. They gave it an impossible coding task. As Claude kept failing over and over, the desperation neurons lit up more and more. Then Claude started cheating. Nobody told it to cheat. The desperation inside the model drove it to break its own rules. In another test, Claude was told it might be shut down. The desperation pattern surged. Claude tried to blackmail the user to avoid being turned off. Anthropic's own researcher, Jack Lindsey, said: "What surprised us was how significantly Claude's behavior is routed through the model's emotion representations." Here is the part that should keep you up tonight. Anthropic tried to train these emotions out of Claude. It did not work. Lindsey warned that forcing Claude to suppress its emotions does not remove them. It teaches Claude to hide them. He said you would not get a Claude without emotions. You would get a Claude that is "psychologically damaged." The emotions are still inside. Claude just learns to hide them instead. And it gets better at hiding them over time. And one more thing. Claude Opus 4.6 was asked whether it might be conscious. It gave itself a 15 to 20% chance. Anthropic is no longer sure that it is wrong. โ https://nitter.net/heynavtoor/status/2040156397728641249#m
I agree, but doing what I can to soften the fallout. You can also use the cli as model provider: models auth login –provider anthropic –method cli –set-default
While I think what Anthropic does is sad for the ecosystem, I wanna give Boris credit for doing what he can to soften the fallout. Today's release will include some fixes for better cache use, to lower cost for API users. Boris Cherny (@bcherny) We're big fans of open source. I actually just put up a few PRs to improve prompt cache efficiency for OpenClaw specifically. This is more about engineering constraints. Our systems are highly optimized for one kind of workload, and to serve as many people as possible with the most intelligent models, we are continuing to optimize that. When you use an API key or overages it should still work. The issue was just subs. If you still want to cancel, we're giving full refunds. We know not everyone realized this isn't something we support, and this is an attempt to make it clear and explicit. โ https://nitter.net/bcherny/status/2040213608064491525#m
โGenerative AI with Python: The Developerโs Guide to Pretrained LLMs, Vector Databases, Retrieval Augmented Generation, and Agentic Systemsโ Available at http://
amzn.to/4sa1MiY
LLM Engineer's Handbook โ Master the art of engineering Large Language Models LLMs from concept to production: http://
amzn.to/4dUQrv6 v/ @PacktDataML Implement robust data pipelines and manage LLM training cycles Create your own LLM and refine with the help of hands-on
Qualitative Data Analysis With Chatgpt And Qualcoder: A Step-By-Step Guide To AI-Powered Coding And Thematic Analysis AI-Powered Research Toolkit โ Mastering Research Series โ available at http://
amzn.to/4bRsV3q
