I haven't checked it "in detail" but someone has: Let me start with the headline: Gemini Advanced is clearly a GPT-4 class model. The statistics show this, but so does a month of our informal testing. And this is a big deal because OpenAI’s GPT-4 (the paid version of ChatGPT/Microsoft Copilot) has been the dominant … Continue reading Google Gemini Review →

This is pre-AI era. I start by reading the abstract. Then, I skim the introduction and flip through the article to look at the figures. I try to identify the most prominent one or two figures, and I really make sure I understand what's going on in them. Then, I read the conclusion/summary. Only when … Continue reading How to read a scientific paper →

This is an interesting idea. Why? Because it involves public funds (with much accountability) and then using that to set up local compute resources to "share the costs". I haven't seen the idea in execution yet, but this could be an example of a shared work-load. Payments can be done by interested academic groups to … Continue reading What is sovereign AI? →

I agree with this assessment: Instead, it’s evident these platforms have reached some sort of equilibrium in which they can incrementally and indefinitely extract ever-increasing profits but without pushing users to defect, not that there are any good alternatives. Hence, why users stick around despite a worsening user experience, such as more restrictions (like Chat … Continue reading Another perspective on “enshittification” →