Clever idea. I like the concept, and think it could be a great product.
But your site gives me a bad vibe:
1) I feel like the term, "Socratic method" is fairly well understood. Am I wrong? Because trying to rephrase it feels like you are dumbing down your marketing for an audience that... well, isn't me. Which is fine if I'm an outlier. I just thought it was a common term.
2) Either let people try it, or make them create an account. Having a page that looks like it is going to let you try it, lets you enter a subject, and then prompts you to login feels worse than just forcing a login right away. (Which is also a problem, but at least doesn't feel like you are trying to trick us.)
Yes! Completely agree with your second point.
Users shouldn't have to think and type a query only to find out the "Create Draft" takes you to a Google login page. Poor UX design choice.
1) I will take note of that. I wasn't sure how common the term was because I never heard of it, but I'm realizing it's actually a lot more common than I thought
2) I absolutely agree with you; I should've given more thought on that. In short-term, I will just redirect users to login-page right away. In long-term, I'm planning on making a 'guest-mode' so users can try it out before having to sign-in
I'm definitely seeing the trend here about the topic prompt!
Just starting to realize that this indeed is a bad UX choice. I was always logged in when I was developing so didn't think about this issue.
Will fix ASAP!
Your product seems nice. I have nothing to take away from it.
However, consider what codingdave have told you about the Socratic questions. The ones in the video are no doubt helpful for someone trying to find their way on a text. But they are not Socratic questions.
This is so cool. I'm a Hyperwrite subscriber and would absolutely consider switching to Scraft.ai if it works as well as I'm imagining. Although GPT-3 can be a powerful ally when writing emails, proposals, documents, etc., on tight deadlines, for me personally, I find it's dreadfully easy to get "stuck" in vicious feedback loops as a result of the language model's tendency to expand rather than react. A great editor will push you to clarify sentences that are vague, improve arguments that are muddled, and (where I'm thinking, Scraft.ai might really shine) is a master of dialectic, poking and prodding at the logical substance of each draft and helping you to realize more clearly at each step of this iterative process what, exactly, it is that you want to say.
I look forward to giving this a shot! It's funny, I'll sometimes spark up a conversation with Socrates in the GPT-3 sandbox if I'm ever looking for a spirited adversary to pick apart an argument before I make it. Explaining one's thoughts to an imaginary interlocutor not only makes writing more fun -- joyful, even -- but in my experience, helps me produce a higher quality product vs. simply asking "mediocre Westworld fascimile of myself, fine-tuned on a corpus of mundane office emails I wrote" to finish my sentences for me.
When it generates sentences like (from the video) 'according to the research in 2012' and 'MIT and Harvard developed Tissue-Bio', is that (intended to be) accurate? Or is that to be taken as suggested boilerplate for real facts to be replaced there?
As mentioned in the info popup, I cannot guarantee how factual/true the AI-generated contents are. They are only meant to be used for inspirations so I would take caution just copying and pasting to the draft like I did in the video.
Probably need to mention that that part of the video is actually not advised for plagiarism issue.
I think this is an amazing tool and implementation, particularly if it is geared toward helping the author understand the issue efficiently.
I would recommend a better voice-over for the video, it is by far the limiting factor of the presentation.
Finally, I would recommend adding more plagiarism controls to the tool. People are going to be so tempted to just paste in what the AI summarizes without understanding or even reading it in depth.
My first thought here was "tools like this could be a really cool to help improve the quality of my student's papers." My immediate second thought was "this trend in tooling is going to lead to tons of AI churned summary copypasta in all of the future paper's I read."
that's actually one of my biggest concerns since I do not yet have a way to hard-prevent them from just copying and pasting summaries.
I'll have to do more research to figure this out, but it's definitely on top of my mind
I used my iphone for recording, so that's probably why.
I agree to your second point. I added caution in the info tooltip but I have yet to figure out an active way to prevent users from just copying and pasting the content.
The demo video was cool. I wish the www/internet existed when I was in high school and this tool would have been a killer app for writing student essays / reports.
I like the idea of this if I am writing an article - I gave this (pretty) webpage only about 17 seconds of time to search for if it had anything for fiction writers - it doesn't appear to, which is fine, but might be worth nothing this is for non-fiction.
I literally released it couple of days ago so I haven't thought about it yet. The costs are minimal atm, so I'm just trying to reach students who can make good use of it for free
This is my first web project so I'm still learning css. TailwindUI just seemed like the shortcut to me, but I feel what you're saying. I do see them a lot too
I just chose google since I thought it was easiest way for someone to sign in without having to verify email, re-enter password, etc.
However, I never thought about the privacy concerns many people have regarding google signins. Thank you for the insight! I will definitely consider a workaround to this.
Good luck when google decides you infringed some TOS because you uploaded a youtube video with music playing on the background or some BS like that and just locks you out of all the services you registered with that feature
It took me less than 15 seconds to create an account on hackernews using a temp-mail (any free service would suffice) to respond to this comment (less than writing the comment itself)
try creating a new google account:
- whoops, you need a phone number
- whoops, you need to go through a welcome guide (it's getting progressively harder to skip it)
- etc. etc.
All that for trying some use random service some random posted on some random news aggregator website without tying it to your 'main online identity'
............
Same thoughts.
Wouldn't prefer using Google to try out a site.
I know a lot of developers who will close the page as soon as they're blocked with a Google/Facebook login.
The service definitely sounded interesting.
Hope to try it when they add an email registration option.
1) One of GPT-3 usage guideline was not to make the app available for public without an account. I am required to send the userID for every GPT request by their guidelines, so I had to force-signin for now. I am considering to make a guest_mode in the near future tho.
2) Seeing the trend here, I'll make a email signup asap
But your site gives me a bad vibe:
1) I feel like the term, "Socratic method" is fairly well understood. Am I wrong? Because trying to rephrase it feels like you are dumbing down your marketing for an audience that... well, isn't me. Which is fine if I'm an outlier. I just thought it was a common term.
2) Either let people try it, or make them create an account. Having a page that looks like it is going to let you try it, lets you enter a subject, and then prompts you to login feels worse than just forcing a login right away. (Which is also a problem, but at least doesn't feel like you are trying to trick us.)