Published inAI AdvancesWhy LLM watermarking will never workHave you heard the claim that “watermarking can help identify text generated by an LLM”? Or that “watermarking can help reduce harms like…6d ago86d ago8
ChatGPT with confidence scoresIn this post we’ll look at using the confidence scores available through the OpenAI API.Sep 15Sep 15
Integrating AI into Windows (a low-tech approach with AutoHotKey)Wouldn’t it be nice if you could press a shortcut and ask ChatGPT about something you see on your screen? Or select a bit of text in any…Jul 172Jul 172
Ensuring Fair Compensation for Content Creators in the Age of AI ChatbotsMuch of the content on the internet is created by humans, and these content-creating humans need food to operate. Food costs money, and…Jul 75Jul 75
The three types of timeHandling dates and times in code can be tricky. Understanding that there are three types of time — and knowing which one you’re dealing…Feb 310Feb 310
87 Lesser-known Python FeaturesThis post is for people who use Python daily, but have never actually sat down and read through all the documentation.Jun 29, 202311Jun 29, 202311
Published inBetter ProgrammingTeam GPT: Simulating Multi-agent Collaboration With Role-playing PromptsJun 11, 2023Jun 11, 2023
Published inBetter ProgrammingThe mean misleads, part 2: more data for the doubtersSome backgroundJun 4, 2023Jun 4, 2023
Published inBetter ProgrammingA Keyframe-style Learning Rate Scheduler for PyTorchWhen it comes to defining learning rate schedules in PyTorch, you have plenty of options. 15 different scheduler classes, to be exact.May 31, 20231May 31, 20231
Published inBetter ProgrammingThe mean misleads: why the minimum is the true measure of a function’s run timeImagine: you’ve got a function and you’d like to know how long it takes to run, on average. So you run the function a few times, measuring…May 27, 20237May 27, 20237