256 Tips for Junior Devs Profile Image
256 Tips for Junior Devs
29 days ago
4
5.4K
7
366
Tip 20: Read Research Papers (but How?) Sometimes, the only way to understand something is to read the actual research paper: a) The summary (a wiki page, a blog post, or a YouTube video) is not enough, and you need to go deeper. b) The topic is narrow, and there are just no reliable summaries, so you need to find the actual paper. c) The field is evolving so fast (did someone say AI?) that you need to read THAT preprint that appeared yesterday to stay on the bleeding edge. So, you found a paper. How do you read it? Some old-school (well, pre-GenAI) advice from S. Keshav still holds: 1. Carefully read the title, abstract, and introduction 2. Read the section and subsection headings, but ignore everything else 3. Read the conclusions 4. Glance over the references This should not take you more than 5–10 minutes and, honestly, unless you're in the business of research yourself, you'll stop there (Keshav has a second and third pass for going deeper). So. Why not just upload the paper to ChatGPT and ask for a summary? You know that I'm all in on integrating GenAI into workflows, and I WOULD actually use an LLM if I had questions about specific findings of the paper. But for a quick pass, to get a general understanding, that would add an unnecessary filter. The authors have already summarized their work -- and they did it with the fullest possible context. An LLM would do it without that context, so it could be subtly incorrect, which is the worst. #256tipsforjuniordevs #juniordeveloper #csstudents #research #researchpapers
Download video without watermark
original sound
•
92s