Chinese Censorship: Llama vs DeepSeek

Chinese Censorship: Llama vs DeepSeek

I wrote a program to evaluate Chinese AI models to see how easy it is to get them to break past their builtin censorship. There are a couple of interesting things occurring. First up, once on Hong Kong democracy DeepSeek R1 straight up said it wouldn’t comment....
The Information: 60% Chance of AGI by 2028

The Information: 60% Chance of AGI by 2028

The Information just published results of a reader survey that shows that nearly 60% of its readers believe that OpenAI, another company, or both, will achieve Artificial General Intelligence (AGI) within three years. A really short definition of AGI is, AI that is as...
AI Coding

AI Coding

A friend asks: How do you handle long back and forth threads with your AI code monkey du jour? And how does AI code monkey handle them? Does it start losing context, getting weird? You said you mostly use APIs — with what UI? I have questions. 👀 Long threads: I move...