Feedback on our MVP
Cedar Milazzo
11 replies
Hi all. We've built a tool to help evaluate whether a web page is "Fake News" or not. We'd love to get some feedback from you! www.trustium.com is our website.
Replies
Chris Davis@thedatadavis
Applydog
@cedarmilazzo1 This is neat. I like the try-it-out approach to show the value of the tool.
Share
@imchrisdavis Thanks Chris!
@cedarmilazzo1 Hey Cedar, great product idea, have you thought about a google like first fold? With a big question (that reflects the delivered value of the product) and a field to input the website.
What I see is some level of friction between understanding what it does, and trying it out, I think it might be an interest test.
@gabriel_werlich Hi Gabriel, yes we have thought of that, and we actually had it at one point. I'll definitely take your input back to the website designers and push for putting it back in! :)
Kelvin (Beta)
@cedarmilazzo1 the "try it now" url input isn't working for me!
@sarahloertscher Hi Sarah. Unfortunately the link was broken this morning. All fixed now! :)
@cedarmilazzo1 @cedarmilazzo1 Hey really interesting idea and like how you've categorized the articles and also the mix of certified reviewers and machine learning, sadly probably a very important tool for the future so well done. When I clicked Try it Out link I was brought to a "Nothing Found" page? https://www.trustium.com/trustie...
@phillip_martin Hi Phillip. Thanks for checking it out, really appreciate the comments! We're in the process of changing our name (used to be Trustie.Online). unfortunately that link was missed! We updated it this morning.
Rohr Kreuz 01
Hi @cedarmilazzo1
The "try it now" button isn't working for me :(
But one question arises : How could we be sure you don't put a bias in your tool ?
Hope it helps
Rohr Kreuz 01
Hi @cedarmilazzo1
Safari, Mozilla, Opera and Google Chrome and it doesn't work on this link : https://www.trustium.com/trustiu... I also tried different wifi 🤷🏻♂️
Thank you for your precision, I like how you answered and your transparency :)
@johan_bavaud Interesting Johan. Which browser are you using? It seems to be working for everyone else now.
As far as bias, we have contracted with a wide variety of people across many different backgrounds (urban and rural, conservative and liberal, highly educated and not, various ages, various ethnicities, etc) to gather our training data. We understand it's not possible to remove 100% of bias, but we do everything we can to reduce it. In addition, all of our data is double and triple checked and only unanimous agreement allows it to be entered into the training set. These procedures are the best humanly possible right now, and we've gotten good feedback from across the board that it seems to be pretty fair. :)
Hope that helps!