I think somebody should take all the leaked and copyrighted code that exists out there for various LLM Coding Agents, and train a model with a special focus on reproducing that type of code. Then they ask the model to build them something like Claude Code but works with local models or whatever.

I want to watch Anthropic have to argue in a court of law that LLMs reproducing stolen code which they remembered in their weights is a copyright violation of the original authors of that stolen code. I want to see them have to argue that, or else have to admit that anyone is allowed to steal their code.

Hell, twist the knife in the wound by asking it to build a "clean-room re-write" of Claude Code.

I want to see this happen, despite the fact I probably would not even use this vibe coded software almost ever. I want to see Anthropic have to pivot to handle this.

#AI #LLM #noAI #copyright #OpenSource

Open Source AI for Enterprise: Data-Backed Strategic Evaluation from a Futurist Who Spoke to 1,000 Leaders

Book Ian Khan for Your Next Event Looking for a keynote speaker on this topic? Check availability → The conversation around open source ai for enterprise is stuck in a cycle of superficial analysis and generic recommendations that do not survive contact with real organizational complexity. I have sp

🔗 https://www.iankhan.com/open-source-ai-for-enterprise/

#AI #Futurist #Technology

📰 OpenAI Discloses Flawed Test Cases Undermine SWE-bench Verified Benchmark

OpenAI has officially withdrawn support for the SWE-bench Verified benchmark after discovering that at least 16.4% of its test cases contain critical flaws, rendering them unreliable for evaluating advanced AI coding systems. The revelation has sparked widespre...

#AINews #AI #Teknoloji #MachineLearning #Haber

🔗 https://aihaberleri.org/en/news/openai-discloses-flawed-test-cases-undermine-swe-bench-verified-benchmark

Linux Mint 22.1 “Xia” released!

https://blog.linuxmint.com/?p=4793

“Your Car, My Breath”

Quite a powerful picture 👇👇

By Marta Frej