Hacker News new | comments | show | ask | jobs | submit login
Show HN: My startup - Summarizing popular HN articles (skimthat.com)
16 points by MichaelApproved 2001 days ago | hide | past | web | 28 comments | favorite



Is this really a startup or more of a side project? Do you plan to expand to other sites/blogs, sort of like the business book summary market?


I plan to expand this beyond HN but, at first, I want to focus on HN since I'm part of this community and it's a great place to get feedback. It's my first attempt at an MVP and I want to make sure I get things right before building out the entire site.


People who believe that they don't have time to read are fooling themselves. The real problem is that they find it increasingly difficult to concentrate amongst the distractions that a lot of us have allowed into our lives. The solution is not more technology (or smart services like this one) but rather a significant change in our lifestyles. Training yourself out of the ability to concentrate will have serious long-term consequences as you grow older and those who seek to invent great things will find it hard to do so without deep thought.


For me, it's not a matter of concentration. There are some articles where I'd prefer to just get the raw information.


Personally, Daniel Tenner's Swombat [1] is already fulfilling that need.

[1] http://swombat.com/


Daniel includes opinion and editorial. I plan on taking things down with no personal editorial.


How is this a startup? not all project are startup thia seems more like a [side] project


I'm mainly looking to get a working formula for now. There will be much more built out once I get that right.


Have you thought about the authors IP rights? For example, if they have a CC no-derivs license on their blog, wouldn't you be violating it?


I don't believe summarizing violates copyrights. Fair use should cover what I'm doing.


What exactly is your startup offering? Is someone going to paraphrase each article submission-by-submission? Personally, if I was told that Jacques Mattheij had an interesting blog post, I would rather read what he wrote rather than what someone else interpreted it as.


Think of it as a TLDR version of an article with just the bare facts. If you find it interesting then you can go on to read the entire article.

I agree, there are times when you want to read the full article but think of all the other times you try to skim through a page and hit the back button after 30 seconds.


There's a typo in the Andrew Warner quote. I think you mean "work" instead of "word". :) The concept looks really interesting. Do you think the target market is large enough to make your service profitable?


Fixed typo, thanks.

The initial concept will focus on HN articles because I read a lot of articles here and it's a great audience for feedback. Once I make sure the summaries are good, I'll expand to more news sources and build out the site.


When I start to become too busy to read Hacker News, I simply stop reading it. I removed it from my RSS feed long ago (who can actually read 1000+ submissions per day, and remain employed?), and only visit the site manually now.

A page or two later, and I'm usually done. Great things usually stay fairly sticky on the main page, and truly fantastic things will generally be mirrored elsewhere.

YMMV.


What if I delivered content that gave you something between stopping to read it and only getting the truly fantastic?


I might be interested, depending on the editorial quality of the summary. Though honestly, I might still come directly to Hacker News specifically for those stories that fly under the radar.

I've noticed that the goals of a startup-oriented aggregator do not always intersect my interests. I think cherry-picking out of the broader feed would only decrease the total number of items to read, but not necessarily improve the signal. The target audience (entrepreneurs, investors, startups) may benefit greatly; I don't know.

It's difficult to get a feel for how this works without some visible examples, though.


I'm working on the first batch of summaries now. If you signed up, you should get some examples shortly.


Curious how the summarization works - are you using something open-source or customized?


I'm having writers read and shorten the articles. They're taking it down to the bare facts with no personal editorial.


Any ideas on how you would run this operation at scale?


Thinking a team of editors around different time zones so they cover news throughout the day. Each editor can do about 40 articles a day.


Or perhaps human? I hear virtual assistants will do that kind of thing.


I was thinking something like http://tldr.it/ at first but human seems like a better solution.


What are you using for the summaries? Is it algorithmic or human? Perhaps mturk or similar?


Hiring editors to read and shorten the articles at first. Once I get a feel for how the summaries should be written, I'll experiment with something like mturk and other ideas so I can scale.


Who is your market and how much do you think they will pay?


My market is someone who wants an overview of news. I expect to operate on ad revenue.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact

Search: