I've been using Explainpaper (https://www.explainpaper.com/) to get an introductory understanding of textbook chapters and technical articles. I want to do the same for news articles and specialised blogs.
Yes, I would like to see a one-paragraph summary of technical articles to decide if they are worthwhile to read. For news articles, the summary might be all I want to read.
I have very little time to read things online. I try to skim the introductory and closing paragraphs in articles I read, but some authors bury the "meat" of the matter among walls of text. And for someone like me, that knowledge is practically inaccessible. I also don't think I have the attention span for some very long technical texts. A plain English summary would enable me to read them.
To recap - I would like to see a GPT-3-like summary of the entire article. It would primarily help me by saving time reading technical texts - either by helping me digest them or by helping me understand whether the text is worth reading in the first place.
If article summaries were cached for each article, that would be fine for my use and probably more economical on your end. I do not want to get bespoke summaries for parts of the article or summaries answering questions.
Have you tried some of the existing solutions mentioned in the ticket? I'm curious if they solve the problem for you, and if not, what Unclutter could do better.
I took time to try all four examples in the feature request.
BlinkNotes didn't work for a technical reason, TLDR; it did not have AI capabilities - only crowdsourced summaries.
TLDR This was good but seemed to use an NLP algorithm they call "AI." So instead of summarizing articles, it seems to produce a bullet list of the essential text fragments. This is in contrast to the much easier-to-read conversational style summaries GPT-3 can provide. TLDR This also did poorly with some of the technical articles I have authored, sometimes extracting pieces of C++ code as meaningful text.
Summari worked very well for me. It summarized even quite technical articles well because it uses a conversational AI. Thanks for the recommendation.
If I had to name one key criterion that differentiates good summaries from bad, in my opinion, it would be the effective use of language. Conversational/language models like GPT-3 are proficient at absorbing much contextual information and synthesizing a short and effective summary. NLP algorithms are good at throwing away superfluous context, which is common in casual writing, but they do not seem to work well for technical writing or texts whose purpose is to explain concepts and where there is little superfluous context to throw away.
Perhaps for something like Unclutter, if the users mainly read news sites, then an NLP approach could be appropriate (it would be cheaper and works well for such content). But the ideal implementation for an article summarizer for me needs that summary-from-a-lot-of-context synthesizing capability.
The big problem with current transformer-based models is the input limit. We'd probably need to split large articles and then summarize the summarizations of these chunks. And yeah I agree, Unclutter should work on any kind of article, not just on factual news.
Please let me know if you're still using Summari in a bit! They seem to focus on their hosted version, I'm curious why.
Yes, I would like to see a one-paragraph summary of technical articles to decide if they are worthwhile to read. For news articles, the summary might be all I want to read.
I have very little time to read things online. I try to skim the introductory and closing paragraphs in articles I read, but some authors bury the "meat" of the matter among walls of text. And for someone like me, that knowledge is practically inaccessible. I also don't think I have the attention span for some very long technical texts. A plain English summary would enable me to read them.
To recap - I would like to see a GPT-3-like summary of the entire article. It would primarily help me by saving time reading technical texts - either by helping me digest them or by helping me understand whether the text is worth reading in the first place.
If article summaries were cached for each article, that would be fine for my use and probably more economical on your end. I do not want to get bespoke summaries for parts of the article or summaries answering questions.