Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: Thoughter (aytwit.com)
312 points by dougk16 52 days ago | hide | past | web | favorite | 156 comments



Very cool base idea and could become something interesting.

Initial thoughts:

1. Even assuming both person A and person B use the service within a week of each other, the odds that they both use the same hastag are astronomically low. Person A says '#wannaGoOut', but person B says '#IWantToGoOut'. There are way too many possible permutations of a single thought.

2. As another commenter said, the design is clean but feels like it belongs in 1995. It will not get off the ground looking like this. The basic layout is fine, but the colors and fonts need to go.

3. Without an app, I think you'll have a hard time gaining critical mass. If someone has a thought and all it takes to get it out onto Thoughter is to (i) open the app, (ii) type, (iii) and send, then it will be way more likely to be used.

4. This feels a bit like Craigslist Missed Connections. Not saying that's a good - or bad - thing, it just does.

5. What if I see a cute girl at a coffee store, we hit it off, but I don't have her email address - or even her name? What then?


Thanks for your thoughts.

> 1. Even assuming both person A and person B use the service within a week of each other, the odds that they both use the same hastag are astronomically low. Person A says '#wannaGoOut', but person B says '#IWantToGoOut'. There are way too many possible permutations of a single thought.

My hope is that hashtags would become standardized, and hints (https://aytwit.com/thoughter#hints) will also help with this.

> 3. Without an app, I think you'll have a hard time gaining critical mass. If someone has a thought and all it takes to get it out onto Thoughter is to (i) open the app, (ii) type, (iii) and send, then it will be way more likely to be used.

Totally agree. An app would also allow the service to avoid email, and allow encryption on the client-side without the server seeing any plain text at all. Definitely a goal.

> 5. What if I see a cute girl at a coffee store, we hit it off, but I don't have her email address - or even her name? What then?

Good question! I don't have an answer unfortunately. Guess you gotta do things the old-fashioned way. :)


Sell this to Snapchat


This really does fit perfectly with Snapchat’s MO of plausible deniability in messaging; and could be exactly what they need to gain an edge against Instagram.

Seconding talking to Snap.


How would I go about that? I'm just a clueless engineer. What little I know of how the world works, is that they would just steal the idea if it were really attractive to them, and maybe give me a free T-shirt. Am I too jaded?


A Snap engineer could theoretically see this post and steal the idea. It’s usually useless to go around trying to guard an idea. However, if you develop the service enough so that it would be inefficient to rebuild it completely from scratch, then they might pay you for the concept and the technology, which is real IP and can’t be ripped off quite so easily.


I'm fairly certain you'd need a user base before Snap were even remotely interested, or at least prove the idea.

Build an app, get it out there, see whats up


The alternative is they do that and don’t give you the T-shirt.


You should consider a Progressive Web Application: requires SSL, and you can place this on the Apple and Google app stores. You just have to do a small and fair amount of JS witchcraft.


To get the most out of a native app, I would want to do client-side encryption/hashing (right now the server sees plaintext for a millisecond, just like most servers treat passwords), and use mobile notifications instead of email, for the communication mechanism. Not saying a PWA wouldn't fit the bill, but unfortunately it's a bit beyond just wrapping it in a WebView and packaging as an app. I suppose as an MVP maybe? Just go get on the stores?


You can do client side encryption with javacript. Not sure if a PWA can do native notifications though.


Would you be open to piggybacking on existing encrypted messengers, like Telegram and Signal? Maybe even WhatsApp (not sure how their API works). You could also build bots on Matrix.


Yes! This is in the future plans section: https://aytwit.com/about#technical_details__future_plans

I agree email is so five minutes ago, but it was the easiest path to get started.


1. The service will send an email to person A saying "Someone wants to #wannaGoOut". As long as person B doesn't make a typo when copying the hash tag, this problem is solved.

2. Yeah, it's a nod to certain eras of web design, which I like, but other people won't care so I agree.

3. It's email-based in order to make it work for recipients who don't have the app.

5. In 2019, it's probably best to support Facebook, Snapchat, and Instagram as recipient addresses... somehow. Although most of those aren't technically possible. Those platforms aren't really great for building services on top of.


> 1. The service will send an email to person A saying "Someone wants to #wannaGoOut". As long as person B doesn't make a typo when copying the hash tag, this problem is solved.

That's only one of four possible settings, and using it breaks half of the examples: it works great for #reunite, not so great for #breakup


My thoughts too on the breakup one, unless both partners have multiple relationships.


I built this service using Instagram before


Link?


> What if I see a cute girl at a coffee store, we hit it off, but I don't have her email address - or even her name? What then?

Something like Berkanan, a physical proximity-based messaging app that uses Bluetooth, may help.

https://berkanan.chat/


or try talking to her.


anxiety intensifies


Some people do not have enough confidence to risk a public rejection.


Then those people should work on building that confidence.


Psssh, what is this, the 2000s?


This sounds like an awesome idea! Too bad it's not available for Android. But that's how I found out about Firechat.


Machine learning could probably be used to fuzzy-match the contents (probably removing the need for the hashtag altogether). Of course, false-positives could cause problems.


Levenshtein distance, an n-gram comparison, or similar might be good enough instead of having to invite a machine learning solution.


Hello HN,

Thanks to everyone checking this out. I know it's kind of a weird project, but that's the idea! I've been kicking this project around since college in 2003 when I was afraid to ask a girl out even though I was pretty sure she liked me too. Instead of just asking her out I tried to come up with a way to fix my anxiety with software, and thus Thoughter was born. Don't worry I'm a bit wiser now. :)

You can read more about the back story at https://aytwit.com/blog/thoughter_origin_story and why it took so long for me to complete. Hint: I cared too much about privacy, before it was cool!


I actually built this specifically for the breaking up use case. It was called CleanBreak. I forget the numbers, but at least a few hundred people used it. I think it'd actually work if you started locally (e.g. get covered in a campus paper).

https://www.reddit.com/r/SideProject/comments/4287xp/i_built...


Nice! The comments on that thread are valuable, thanks for sharing. Do you happen to have an instance of it still running somewhere, or screenshots? I would be interested in some specifics of how you approached it since people seem pretty favorable in that thread.


just spun it back up on PythonAnywhere, http://mritchie712.pythonanywhere.com/, pretty sure most of it won't work, I didn't start the database, but you'll get the general idea.


Sweet thanks for that. I like it, very straightforward and to the point. I'm considering how I can "subdivide" Thoughter into little special purpose utilities like for breaking up, hooking up, etc. Maybe I just need to make hashtags more standardized, with a dropdown or something. Much appreciated.


As an introvert, I completely understand where this is coming from. What you are trying to create is a way for people to signal intent in an emotionally safe way. But I think this misses the mark (by substituting technical infrastructure for quintessentially human interaction; by encouraging "someone is" vs. "I am"). There is an art to "working toward" a discussion or disclosure (which many people are not good at, and many situations make difficult). To help someone find a way to say something (difficult or risky) to someone (who may not be receptive), and to allow the other party to reverse-signal gently (possibly by "affirmatively" ignoring) would be valuable, but as complex as the varied use cases. I think a system that helped people work through what they want to accomplish and possibilities for achieving that would be useful; the infrastructure for communication is already there, we just need to think about how to use it better (think guided workflow, decision tree or expert system). Having a mechanism to "buffer" between the sender and receiver might be helpful (ex., sending someone flowers or a card/letter vs. speaking to them) to overcome emotional resistance, but shouldn't substitute for thinking through <your state> + <situational state> + <other's state> => <appropriate action>.


Very thoughtful comment, thank you. I am going to write up a blog post on why you should not use Thoughter in the first place, or at least only as a "break the ice" mechanism, and I'm going to use a lot of the comments here like yours as fodder.


I think this (or at least, the use cases) is missing something that NLP also has a hard time with: individual words do not have positive / negative meaning.

Hypothetical example:

"Hey, you're a cool person! #wannaGoOut ?"

"No, I do not #wannaGoOut"

Now, in spite of the rejection, the person asking is revealed. The thin veneer of anonymity is utterly lost here.

(for those curious about my NLP reference, pretty much any remotely controversial topic, word or hashtag will be used with both positive and negative sentiments)


Even if you handle negatives it will be intentionally tricked by curious people.

Yeah I #wannaGoOut. Actually Not, just wanted to see who you are.


Just to fix possible confusion, there's no "reveal" step here. The recipient has to guess who sent the hint (if there's a hint). And with no hint, both parties have to spontaneously use the service with the same hashtag. Not saying there's absolutely no possibility of harshly rejecting people, but the point of the service is to give a little more breathing room to their pride/egos.


Cool, this is somewhat similar to an idea I had before: two persons each make a list, and the app informs both parties of common items of the list without disclosing the rest. That would solve a few of the same use cases you mention, and help decide e.g. what movie to watch, even if one of the parties (e.g. my SO) is somewhat ashamed of their preferences (e.g. romantic comedies), unless the other person shares them.

Not sure if this is interesting to anyone, but in any case, I'm unlikely to be going to make it, so feel free to run with it.


That's exactly the premise behind Mojo Upgrade and other couple sex practice questionnaires.


It may not be obvious, but you could use Thoughter for this right now! The UI/UX isn't totally there yet, but you could simply send off let's say 10 thoughts to your SO about something really intimate like sexual preferences, and see what your SO matches. Search for `#wellLetsLeaveItAtThat` on the page.

When I say UI/UX isn't there, I mean this needs an app, or at least sign in sessions, so you don't have to confirm your email every time and such.


Interesting concept! This could also be cool for missed connections.

In lieu of the missed connection's email address, you could provide a few specific details about the encounter (choosing from prewritten options):

Time (options provided in 3-hour blocks ie 9am-12pm), name of neighborhood, city, your ethnicity (or some other easy personally identifying feature).

If the other person also goes on the site to submit a missed connection within a few days of yours and fills out the same details, they'll be prompted to answer the detail about your personally identifying feature to receive your message and relay email to respond to.

(Yes this would only work well if the website became widely known as the place to go to for private missed connections.)

Probably a few loopholes with this, just thinking out loud.


Wow I had thought briefly of the "missed connection" case but never could work out the details. Your post makes me think there could be a way, if there were some standardized version of a "thought" that goes beyond the hashtag. Interesting. Thank you for jogging my brain a bit.


If you send a hint about a difficult conversation (#reunite, #breakup), what are the odds that the recipient won't be able to guess who sent it and the gist of the secret? Is this any better than just having the conversation directly?

If you don't send a hint, it seems extremely unlikely that the other person will use the service with the exact same hashtag, as another commenter pointed out.

So, where is the utility?


Great question. One answer is that it gives the recipient plausible deniability (I think I'm using that term right, sorta). They have the power to not respond even though they probably know who the sender is. And if they don't try to respond then hey they can never be 100% sure of who the sender is. And the sender can never be 100% sure their message was even received. So basically it gives both party's ego/pride some breathing room.

As far as the "no hint" case, well yes the service would need to be in the zeitgeist in order to have utility. Who knows maybe one of the big social networks will steal the idea. They can go for it! Just remember little ole me in the credits section somewhere, is all I ask. :)


Other attempts to get at some of the same issues:

Lovegety and other "proximity matchmaking devices"

https://en.wikipedia.org/wiki/Lovegety

Online questionnaires for couples that only yield matching answers:

https://mojoupgrade.com/

https://www.weshouldtryit.com/

https://www.letsexperiment.xyz/


Also: https://www.reciprocity.io/

Once in college I got an email from (I think) ecrush.com, saying "someone has a crush on you", and on the site you could enter email addresses of people you had crushes on and it would spam them that same thing, anonymous until there was a match. (Personally, I suspect it was bootstrapping by sending fake initial "someone has a crush on you" emails to scraped emails, but I didn't dig in too deep.)


Forgot to reply to this initially. Those links are much appreciated. Thoughter can be used for this kind of thing right now, although the UI/UX needs improvement to facilitate such "list matching" use cases.


This will get abused by spammers, even with email confirmations. Be prepared. I had a side project get killed because the utility of it couldn’t quite match the spam hassle.

Further, if I received an email from this site, I would assume it was a scam and delete it. I get dozens of emails from services pretending I’ve signed up to them or that they have real content from me.


> This will get abused by spammers, even with email confirmations.

Could you elaborate? I have vague worries since there's not a captcha or anything. But I can't think of an attack or abuse beyond a DoS-type thing, and I don't know why someone would do that beyond trolling. I know my naïveté probably sounds painful but I would appreciate more feedback here.

Would a captcha be enough? My concern there is that I'm being super paranoid about privacy (https://aytwit.com/about#privacy) and wouldn't want to incorporate Google's captcha if I don't have to.


You’re letting anyone with an email address send mail (through your service) to another address. It looks like the spammer wouldn’t be able to completely control the contents of the message, but they could still send unsolicited junk emails from your email provider or server, which might get your email or IP address blacklisted by spam filters and prevent legitimate users’ emails from being delivered.

I would implement heavy rate limiting and maybe come up with a basic captcha that would stop most spammers.

Oh, and by the way: if you want italics there can’t be a space between the asterisk and the phrase to be italicized (like * this * versus this).


I wouldn't worry about this unless/until it became a problem :)


You have a form on your site where someone can set a destination email address and a hashtag message. Am I missing something or can someone put target@domain.com and a private thought of a spam message or abuse or anything like that? Trolls will hit it and as soon as bots find it, you are going to get more spam than users by some stretch. If people report you, you will be blacklisted very quickly and your host will get stroppy. I'm speaking from experience, but in my case it was a tinyurl-type site.

Even if you require approval from the sending email before pushing the thought, someone can still use Mailinator/similar for that.

Personally, the combination of solution-in-search-of-a-problem and this sort of hassle would make this a non-starter for me. I can't think of a single time someone would realistically use any of the modes the site lists. If they used a hashtag, there's no way to handle knowing which email address someone might've used (personal, work, etc).

The first-date example had me checking for April 1.

"It's all up to Lisa now. If she really wants to get out of there she might check for a bunch of variations like #wannaGetOutOfHere, #letsBounce, #notFeelingThisPlace, #wannaGo, etc."

What are the chances that two people on a date use their phones long enough and are both aware of an obscure site to make more progress than just saying "Well, this is very formal. What do you think?"


There is at least a confirmation email step. Otherwise hehe your worries on spam do echo mine. So far I've just been ignoring that aspect and leaving it open. Honestly I don't know how any web service that ultimately sends transactional email really deals with it, beyond brute-forcing it down to an acceptable rate. Even Google's captchas can be defeated by bots nowadays AFAIR, and they're the best. My thoughts are that ultimately you have to hope that it doesn't happen to you. I would appreciate other people's opinions and experience here as well. Thanks for your comment.


Congrats on the idea!..

But some of the use cases seem a bit childish.

Maybe I'm too mature but it's much easier to just suck it up and be honest and open about a topic. This seems like it would enable a father to forgo talking about a difficult subject with his child because he would have a false sense of "well at least I tried" from using the app. Long story short I think the idea risks decaying a relationship more then strengthening it.

I'm probably wrong though! I'm wrong constantly.

Good luck!


I like this comment and the two replies. It shows the two sides of the coin. When I came up with this idea initially in 2003, I was the kind of person who would use this service. Nowadays I'm more in line with your thinking. But I don't see myself as "more mature" nowadays, just different. Especially since I flip flop on so many things throughout life, I wouldn't be surprised if at some point I turn back into the kind of person who would use the service. Neither kind of person is "wrong" or immature IMO, just different.


You're not wrong, you're so right. The author of this app described it as the technique you use in highschool to ask a girl out. Direct communication, although difficult, will get results.


being unable to relate to the fear some feel in these use cases, and how hard that can be, does not make you mature. it's another type of immaturity.


This is a pretty clever idea though the execution feels a bit off.

My guess is that this would appeal most to a younger audience but how many teenagers communicate via email?

I think it'd be a hit if it could somehow integrate with social media or even just SMS.


Yup email support was just the proof of concept. Definitely agree that it will need to go beyond. In fact if I make a native app then mobile notifications can be the communication medium itself.


They can even in browser. Get an account on Twilio and explore their SMS API.


Sorry, I just don't see the point.

If you send hints like "someone wants to #reunite" I'm sure the recipient will accept just to find out who the person is regardless of what their offline intentions are. Since your identity is revealed anyway I don't see how this saves you from the perceived embarrassment of just asking the person directly.

In the past I've often desired and tried asking for things subtly or in a round about way so as to hunt at what I want without the risk if rejection. In my experience it's not a great path to take.

If the fear of rejection outweighs the potential benefits and you can't bring yourself to ask (for a date, raise, reunion, etc.) then maybe you just don't want as much as you think you do.

When you want something deeply enough you'll ask for it without fear or embarrassment.


> If you send hints like "someone wants to #reunite" I'm sure the recipient will accept just to find out who the person is regardless of what their offline intentions are. Since your identity is revealed anyway I don't see how this saves you from the perceived embarrassment of just asking the person directly.

The way hashtag-based hints work is that the recipient would only see that someone had a thought with that hashtag, not who. I don't know if that addresses your point exactly, but the recipient in this case doesn't necessarily know who wants to #reunite, it's not revealed just by clicking or anything. If they suspect someone specifically then it's up to them to guess essentially.


Yeah but then you just write "#reunite" and the match is made and identity is revealed right?


The recipient of the hint #reunite has to guess the identity of the person who sent the hint. It is never simply revealed in this case. To add to the confusion ;) there is another type of hint that only reveals the identity and not the hashtag.


Is there anything to prevent guessing dozens of people until you guess the right one?


OP suggested that they might add a cost per guess.


There is zero chance this would work. Maybe if each email address started with 10 credits and chewed through them.


That sounds like both a good way to both monetize and rate limit spam guessing. I'm not seeing where OP suggested this idea, so I'm crediting you! Thanks. Or maybe I'm OP?



Ah lol I didn't figure anyone would read through that!


Nope! Beyond a lot of clicking :)


> If the fear of rejection outweighs the potential benefits and you can't bring yourself to ask (for a date, raise, reunion, etc.) then maybe you just don't want as much as you think you do.

> When you want something deeply enough you'll ask for it without fear or embarrassment.

It sounds to me like you are not an introvert and judge who is by "not wanting things enough". It's not as simple as you think.


As an introvert, I disagree. Being introverted does not mean you lack the ability to be assertive, it simply means you don’t draw your energy from social stimulation (quite the opposite).

There is wisdom in the phrase carpe diem.


I'm definitely not an extrovert. I think I've come to realize over time that I'm always making cost / benefit or risk / reward decisions with my actions.

If I want a raise, I have the potential for a reward and a potential risk of letting my boss know that I'm not happy with my current salary. I think that fear of asking is really that fear of rejection or fear of disconnection.

What I'm saying is that this tool won't always be there and we need to find ways on our own to minimize the fear / risk of asking for what we want in life.

Sure, you might get lucky a couple times with this app, but life will keep throwing you more awkward situations.


Based on what I understand from a cursory read of the site, you can't choose to "accept" just based on getting a hint. You have to actually guess the identity (email) of the person requesting to receive the full message...


I don't like this for the reason that it pushes people to use technology instead of hard conversations and personal contact, because it's so much easier.


I kind of agree with you, but the cat is way out of the bag on this one. See: Tinder et al.


I don't understand how Tinder is the example of that.

If I have time to meet a woman tonight and it's 4pm, what exactly is Tinder replacing? The hell of cold-approaching women at bars until one of them seems to like me? Mine my depleted social circle for a date? Wank off in my bedroom because I don't have any new prospects and be happy with that?

The whole point of Tinder is to find new face to face interactions.


Not fair.

A lot of people are not comfortable with confrontation in the first place. And it is entirely normal. It just how people are wired and expecting them to change themselves drastically is not fair on them. It should be a gradual process and a thing like thoughter is a great use for "getting started". End goal is of course to confront.


Good idea but it needs some work. I get the reason of having hashtags is to only reveal the attempt to connect when the intent matches, but with arbitrary syntax allowed you will get a lot of false negatives when your chances of people even matching their timeframes are low as it is.

Instead offer a few predefined categories of tags, and the user can select one or more of them. That also improves the UI as you can now offer autocomplete in that field and lower the barrier of entry (removing option paralysis).


Agreed in full. I was thinking a dropdown of common hashtags. It's sort of a chicken and egg problem so far though, in that I don't know what common hashtags would be. Maybe I'll just use those that are examples on the front page to start.


One obstacle is credibility: Since the dawn of the internet there have been companies trying to recruit users by pretending that someone winked at you, or similar. That would be my working assumption, getting a message from you. How do you get past this?


It may sound either harsh or naive, but seriously my only strategy right now is to just hope that people using the service don't feel that way. If I understand your point, I personally (or someone working for the service) would have to start trying to recruit people by (a) getting big lists of known emails (easy to do), and (b) sending them fake messages that (for example) someone is interested in dating them.

So I guess I just have to keep my morality in check? :)


Love the idea. I was trying this out. Turns out that the email goes to the spam folder of gmail :(


I noticed people having this problem. Not sure what to do about it. :( Going to look into it more.


Not sure if you can solve this at all. Because thoughter will be used for a reason that it isn't meeant to be used for.

So correct me if I am wrong but I think thoughter is most useful when you almost "know" that you have a thought resonating with somebody else and you want to confirm it. Its not for making wild hits at people. Right?


Correct, not for making wild hits. And yea getting past spam email anywhere near 100% of the time is basically impossible, but I did seem to be getting into people's junk folders more often than I expected. I'm using https://postmarkapp.com/ FWIW, so going to see if there's different language or format or paygrade or something that will improve my delivery rate.


This is not a great idea in my opinion, here's why:

One should never feel withdrawn from any conversations, or exploring their feelings. If you want to ask someone out, but have fear/anxiety, you have a "problem" you need to address bravely. Get to the bottom of your fears, reach a conclusion and act on it. Whatever happens, happens. Using thoughter just leaves you buried in your comfort zone, making you never questioning the nature of your own self. It's an easy way out or an excuse for not being accountable to your self. Leaving result of important situations in your life to external circumstances is not a good idea.


As weird as it sounds, I'm the author and I kind of agree with you. I feel like Steve Jobs selling iPads but not letting his kids use them lol. Not to say I think someone is wrong or immature to use the service. Indeed I came up with the idea in 2003 when I was the kind of person who would use it. Now I'm not so much, but I'm not better/wiser, just different. In fact I fully expect to flip flop back to my 2003 mindset at some point in life, and hopefully Thoughter will be there going strong!


So I actually had this idea years ago. People were joking that people can "sync" up so much on twitter that the next evolution is just thinking thoughts to each other.

Even bought the domain thottr.com for this back in 2011. Hilarious


Woah crazy! How much you selling that domain for? ;)

There's actually a patent from the mid-90's that basically describes the same idea that we both came up with: https://patents.google.com/patent/US5117358A/en. I found it after I had finished the first version and was crushed that someone had thought of my genius idea first.


Keep in mind that "thot" is nowadays being used as an acronym[1] that you might want to avoid being associated with.

[1] https://www.urbandictionary.com/define.php?term=Thot


stop killing my dealflow man! :)


Lol! I was joking about the domain, but I checked your profile and really dig what you're doing. Followed!


shoot me an email if you're serious :)


Doesn't this leak too much information if both parties aren't being honest? E.g. you could just submit spouse + every likely hashtag--#breakup, #haveKids, etc.--and then see what ends up matching.


I don't think so. You have to confirm your email, and the target has to enter your exact email + hashtag for it to be matched.


Right, but suppose I know or suspect that a specific person is trying to send me a hashtag, but I don't know what that hashtag is: let's use the example given in the posted website of a new romantic partner wanting to take things to "1st base", "2nd base", etc. Can't I just post all four bases and see what matches? That would reveal their preference without revealing mine.


Ah but one can argue you have revealed your preference, namely that you have no preference, that you'll take the highest base on offer! :)

And note that the other party can check if you sent those other thoughts out if they wanted, so it's socially risky for you to do so ("Ew what a creep, he blasted out all four bases!"). This stuff is fun to think about.


I really like this idea, and I think that exposing the underlying SQL schema (especially in literate programming style like that) is the next level in technical transparency.


Thanks! I spent a lot of time on it for just that reason, being as transparent as possible. Here's the link in case others are wondering what you're talking about: https://aytwit.com/sql


I think this is a really interesting product, but I also find it really weird. Maybe I’m just getting older and this isn’t for me. It wasn’t that long ago that people actually had to have the courage to have a conversation. It kind of breaks my heart to think that a father (or son) wants to connect again but only if the other person does as well. That’s just not how that relationship works.


Maybe this makes you fortunate for having kind relationships throughout your life, or less fortunate for not knowing the difficulty of reconciling a relationship.

One perfect use-case for this is for an adopted person. They don't know their birth mum and dad, and it takes intense bravery to face up to that and reach out and trust them to feel the same way as you. What if you get rejected? Wouldn't you try to make this interaction safer? What if you could even practice it knowing that the communication would be erased in a short time?

Yes, I'm talking about me. And others in my position.


I have been fortunate to have mostly kind relationships. I apologize for my insensitivity. I think you’re right about the use case you’ve presented and there are plenty more, I’m sure. My criticism of an example use case shouldn’t have presented as writing off the whole premise. Presenting it as a way to ask a girl out, or making up after a fight doesn’t present the power it could have.


I couldn't think of any more powerful examples than those. I might wish otherwise, but if Thoughter does become popular it seems like it would be, if I may be blunt, off the backs of teenagers and/or college kids looking to hook up, so Tinder or Snapchat-type interactions. Or maybe it should go in the opposite direction and be used for more frivolous situations? Open to any suggestions.


This has nothing to do with your age or the "kids these days" or social change. Just because you can't see yourself using something doesn't mean there's not lots of people who could use it, or that something is wrong with society.


> That’s just not how that relationship works.

You're suggesting this is an unrealistic example? The tool is far from elegant enough to facilitate such a reunion, but I can count a handful of such estranged father/son pairs for which the right mediation might help.


No, I don’t think it’s unrealistic. You’re probably right that it would work and could be helpful. We’ve all probably known, or have experienced first hand, a situation where you don’t talk to family for years. I think the only part of it that I am being critical of is that both parties need to engage, opaquely. People are stubborn. I’m no different. I just think it would be unfortunate if a person really wanted to make amends or right their wrongs with someone and never had an opportunity to do that. It’s okay for the other person to have no interest in that, but what if they want the same thing but are too angry to admit that? Maybe I’m thinking about it too much.

I do think some of the other examples are really good though, and I like the possibilities. I think I’m just sensitive to one part of it, and that sensitivity shouldn’t write it off fornwhat it could be.


If you have no use case for this product, consider the ways in which you have been fortunate.


Perhaps the father/son example is too dark, but it does offer a use-case other than the obvious asking-out-for-a-date example.


Cool premise, but how will the recipient know what they're receiving?

Also, layout is clean, but feels a bit antiquated. Bootstrap perhaps?


Best wishes to you.

If I am reading correctly, if person S sends person R a message M bearing hash tag H, if R sends their own message M2 to S bearing hash tag H2, then M will be revealed to R and M2 will be revealed to S, if and only if H equals H2. (S may choose some behaviors, such as revealing their identity, in order to prompt R into sending a message.)

As for your second paragraph: I like this site's design<1> a lot. I was able to skim it quickly, and absorb enough to answer your question and more. With respect, I think more work has gone into its design than "antiquated" would suggest.

For one example, on mobile the body paragraphs on the right of the screen can themselves be dragged to the left to read all the info, instead of my having to use the tiny fiddly scrollbar. I think that's a smart feature for a mobile site with a fair bit of text (such as on the SQL page.) Consider also that the site uses colors for text, background, and UI elements (body, nav, etc) that neither clash hideously nor strain the eyes with too-low contrast; plenty of old websites don't take that level of care (think a typical professor's academic homepage - now THOSE can get antiquated!)

Bootstrap will get you a "modern"-style site without too much effort, but I don't think it would add anything here.

<1> By design I am waving in the general direction of layout, color scheme, ease of use, etc.


Wow thanks for the compliments on the design. I indeed spent ridiculous amounts of time on the design and making it super simple and accessible, all the while knowing it would mostly get crapped on. :) I appreciate your appreciation!


You're welcome, Doug. I also enjoyed checking out your personal site in your HN profile - that crazy interface seems tailor made for a Mobile user to "explore"! Lots of fun.


> Cool premise, but how will the recipient know what they're receiving?

I think https://aytwit.com/thoughter#hints answers your question.


The hint, "Give them the exact hashtag that you use" seems rather easy to brute force. Maybe I am just young, but there are only a small number people who would send me hashtags like "#reunite" or "#wannaGoOut".

As others have mentioned, without this hint hashtags are rather sparse.


Note that hints can also simply be "who sent the thought" and not the hashtag or anything. Doesn't directly address your point, just another way to use the service. See https://aytwit.com/thoughter#hints for the different hint types.

Edit: realized from your comment you probably read that section already, apologies!


Interesting idea! Any stats on usage (e.g., how many people have actually connected without hints)?


This post has driven the only real usage since I finished this about a year ago, and I'm guessing most of the connections so far are people just testing it out.


I tried testing it, but it hasn't come through yet. HN hug of death?

Do the messages to the other person get a delay? Being able to set that delay, or potentially set a window during which the thought would be randomly sent would be cool.


Quite possibly! I never anticipated getting to the front page with anything so I don't know what to expect. Let me know if it eventually came through.


This sounds like a romantic idea for some possible contexts (though I feel like it's better to go out on a limb in most situations).


Why use email? Why not integrate with Facebook? Or use phone numbers? Or Snapchat usernames?

Using an email address for this seems archaic.


Integrating with Facebook or using phone numbers seems archaic – not everyone on the internet has either of those. Using Snapchat usernames seems even more myopic.


Fun concept in a game theoretic kind of way, but it doesn’t work in practice obviously. Telling someone to go to a site and push a thought about you is extremely awkward, the only context where it might work is within a dating site or similar, but people don’t seem to need that indirection there nowadays.


Perhaps https://aytwit.com/thoughter#hints addresses your game theoretic concerns?


Yeah in theory this is a good way to handle that concern, but I still think it will be awkward.

Also, it seems this service is only relevant at all in romantic contexts, and My gues is that most girls are not attracted to men that are to shy to talk to them directly. But maybe I’m old fashioned.


Not a comment on the idea. I like the UX execution on mobile. The menu on the left is always visible, in the background. The content is swiped left / right to reveal the menu.

It’s counterintuitive at first, but within a second you work it out. Works well on iPhone 6s safari.


Thanks, it's mostly gotten negative feedback, which I understand as far as the pure styling (color, font, etc.), but I did spend a lot of time on the "structural" aspects of the layout so they work well on all devices without any real per-screen-size logic. Thanks for noticing! It was a lot harder than it probably looks. I will probably spruce up the styling going forward.


A chat bot would probably be a neat interface: 2 people on Telegram/Facebook Messenger/Whatsapp/Instagram/Snapchat, so no app install needed. You can still collect the email address, but that could get around the initial hurdle of accessibility.


So it's a bit like tinder with arbitrary notions and not just I like you. Pretty cool idea actually. (I didn't mean the tinder comment negatively, I also think Tinder is a fantastic idea)


I live in a hole so I don't even really know what Tinder is. No offense is possible! Thanks for the compliment :)


The reason it won't work in current implementation is people are using too many identities on too many communication platforms today. Many not using email at all or not checking it often.


I don't really have anything to comment on except for the… name. Sorry but how am I suppose to pronounce this?

Thought - ter? Sounds like I'm saying Daughter.

Tough-ter? As in tuff - ter?

Terrible name imo.


I don't like the idea since it essentially kills hope. Hope that a relationship might work out when the time is right. Putting a timer on hope is a bad idea.


"You received a hint: someone wants to #breakUp with you"

The thought of that just cracked me up :D Seems like a really good idea though, I like it!


Lol in that case the initial sender could be more subtle and send a hint that they specifically are thinking something about the other person, but not what. Then it would be up to the other person to try to match the thought without direct knowledge of what the hashtag is.


I am very curious if OP can share some stats on number of times two people were thinking about each other at the same time.


Here's a real minimal stats page I created: https://aytwit.com/stats

Short answer: 22 matches so far from this HN post. I suspect most of these were people testing it out, but who knows maybe there were some real ones? If anybody's reading still, post if you got a legit match! I purposely don't store any thought/person-specific data. Just anonymized and aggregated metrics: https://aytwit.com/sql#event_table


My thought: You would need to be adopted by an already very popular company to have much hope of going anywhere.

This is a classic chicken-and-egg issue: No one will want to use it unless/until enough people are using it to have some hope of a reply.

Any plan for how to address that issue?


Thanks for your comment. I recognize your username and enjoy reading your comments.

https://aytwit.com/thoughter#hints sort of solves the adoption problem, in that you wouldn't necessarily need a critical mass of people for the service to be useful. Even just one patron could initiate a thought with another.


My bad. I failed to read all the way down.

I will suggest that maybe you need to move the "Hints" section or otherwise highlight it in some fashion. If they have the thought "Eh, no one even knows this service exists. Why bother?" before they get to the Hints section, this could be a barrier to adoption.


Definitely not your bad. I've been in the game long enough to know people don't read things! (No offense to you personally).

I tried linking to that "hint" anchor tag in the first paragraph but even that is expecting too much. I'll give a think about how I can highlight the whole hint/adoption thing more since it's a main point of feedback. Thanks!


Free suggested edits that you can use as you see fit (or not at all -- whatevs):

Sometimes you want to ask someone a question or tell them a secret and you suspect or hope they're thinking the same thing, but it's uncomfortable to bring up. This is what Thoughter is for!

And they don't even have to know Thoughter exists. You can send them a Hint. But more about that later.

Let's say it's the holidays and you and a loved one had a fight a while back. You aren't sure they would welcome hearing from you. You can send a Thought like "I miss you #reunite" that lasts one week. If your loved one sends the same sort of thing with a matching hashtag, you each get the other's email. If not, nothing happens.

This used to be done much more naturally back when people routinely lived in close-knit communities. People would ask a friend a question and that friend would ask their friend. Word traveled by the grapevine, which helped sort things out.

But in our increasingly mobile world, a lot of grapevines have basically died. This is your modern grapevine, now with Encryption!

Move: "The first project out of Aytwit's software research labs." to much later. Like close with it or something. It's important to you. It mostly isn't important to the process of getting adoption. Making it the first sentence amounts to "noise."


Woah great feedback, I will definitely use some of this. I'll ping you again privately when I make some changes, just so you know I'm not blowing smoke. ;)

The opening line "The first project out of Aytwit's software research labs." is making me cringe now lol.


That site says they are open to consulting, but I'm not sure if it's satire or not.


Good idea.

But this requires too many assumptions on participants.


font-family: 'AytwitHeaderFontEmbedded', 'AytwitHeaderFontFromUrl', Comic Sans MS, Comic sans-serif;

I would reconsider this decision.

Edit: To be clear, a font sets the tone for the content of a page and unless your project is an invitation to a clown orgy then Comic Sans sets the wrong tone.


Also, the colors in use combined with borders and rounded edges gives me a distinctive vibe of 90s kids' software for some reason.

Edit:

Compare the current version: https://i.imgur.com/W8stZVf.png

...to the same thing with a couple of quick styling tweaks I added in dev tools: https://i.imgur.com/RJhLz7B.png

Still offbeat, much less "child's keyboard color scheme" (https://www.amazon.com/dp/B007YW3CL6).


Thank you for taking the time to do that. I hope you don't mind, I may have to steal some of that when I eventually do some tweaks to make it more modern. :)


Feel free. The main things were just the color scheme, playing with spacing, and keeping the underlines on the nav links from standing out.


Where is the clown orgy? Asking for a friend.

Seriously though the font put me off at first glance, people who just visit it might not hang around long enough to become a user.


haha the comic sans made me laugh out loud, i have not seen that in 10 years


[flagged]


This comment breaks the Show HN guidelines (https://news.ycombinator.com/showhn.html) as well as the site guidelines (https://news.ycombinator.com/newsguidelines.html). Would you mind reviewing those?

It doesn't take much poking around to see that the project is in earnest.


Are the Show HN submissions systematically verified to be in earnest ?


No, but some are unsystematically verified to be in earnest.




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

Search: