This Conversation Is Going Nowhere | Hacker News Reader

This Conversation Is Going Nowhere

blog.digit.co

58 pointsjkestelyn27 comments

I've thought for some time now that the real killer app for natural language assistants—especially voice activated assistants like Alexa, perhaps slightly less suited for text-based chatbots—lies in short, episodic, choose-your-own-adventure-style games. The more I think about it, the more I'm not really sure why this isn't a thing yet, and why we haven't seen a resurgence in the popularity of interactive fiction—albeit this time for the masses, and not necessarily limited to those inclined to go load up a copy of Zork.

And content creation is more accessible than ever. I mean, have you seen an example of what the source code for an Inform7 game looks like?

https://en.wikipedia.org/wiki/Inform#Inform_7_programming_la...

To reiterate, it almost seems crazy that there isn't a massive catalog of $0.99 serials that you can play through in 45-minute chapters while cooking, driving down the road with the family or on the way to work, etc.

There are lots of these. In fact, Amazon even has a tool for creating your own.

As for regular CYOA interactive fiction, there are lots, And lots of tools for making them. Even start-ups promoting them. But so far there is no long-tail. And outside one or two successes (like "80 days" on iOS), players don't seem to be flocking to it.

It's easy to say 'it will be huge', but it it doesn't seem to be even the most popular form of a Alexa skill. Of course it only needs one big success to make it look inevitable, but so far, colour me sceptical.

Wow.

This is the tool for Alexa. Fantastic.

https://developer.amazon.com/blogs/post/TxEQV5K754YS77/Annou...

I tried it out and it wasn't very usable. You can never have two options reconvene at the same node. And there's no built-in option for adding chance or variables to decisions. I decided to start building my own framework for these options.

That seems sort of like a bad idea. I'm imagining trying to solve the coin-weighing puzzle of Spellbreaker or the endgame of Spider and Web while driving, and even with a voice assistant, it ends up like this:

  Lake (sitting inside the car)
     You are floating on the surface of a beautiful lake.
  Someone has foolishly driven an automobile into it.

  > CAR, CALL 911
  The voice assistant cheerfully burbles, "You shout
  '911', but the interrogator is unimpressed by your
  outburst."  The floor upholstery is becoming
  distressingly damp.

  > CAR, QUIT GAME
  The voice assistant cheerfully burbles, "Autosaving...
  Exiting GrueFrotz..."  Cold water has risen to your
  ankles.

  > CAR, CALL 911
  The voice assistant cheerfully burbles, "Contacting
  Emergency Services!"  Lake water creeps up to your
  calves.

I'm not trying to imply that any given game as it currently exists would be a suitable candidate for the format. In-game puzzles would obviously benefit greatly from consideration by the authors of the input method and player environment. Dense text descriptions and really tricky minigames ordinarily expected to be presented visually don't make a perfectly smooth transition for casual, voice-enabled play, the same way that traditional WIMP desktop apps don't make for good use on mobile touch devices.

1. What does this have to do with Spellbreaker or Spider and Web?

2. This could very easily be handled with overrides at the OS level for specific commands.

3. If this person perished while sitting in a sinking car, trying to arrange a rescue over the phone, they deserve a Darwin award.

1. They have some intricate puzzles that could be distracting to someone who is driving.

2. Yes, it could, but this hypothetical car manufacturer didn't think of that. They probably don't have a "drive it into a lake" test case, either.

3. I wrote another "transcript" that ended up with the person escaping, but seemed like too much text and not enough delivering the point.

This is such a great idea.

Apparently some exist:

https://www.polygon.com/features/2017/2/27/14643756/text-adv...

Oh man. I've had 0 interest in any of the voice activated assistants so far. But this would be the killer app that would get me on board.

Here is another company doing this:

https://www.earplay.com/

I've often thought the same. Zork is the path to a conversational bot that people find it worthwhile to talk to.

This is exactly a project I've wanted to make for a few months now - I have some ideas outlined and notes written for a game like this. I just don't have time to build it, especially when it'll in all likelihood make me far less than minimum wage.

"The technology to make AI-powered assistants truly useful is still far out of reach, and people aren't rushing to close that gap by adapting their behavior."

Exactly; if you build a bot with the intention of relying 100% on NLP, you're asking for trouble. But I fail to see how that fact leads to chatbots being useless when there are plenty of tools available for guiding necessary optimizations that can lead to a much better experience.

When websites were bad we turned to tools like Google Analytics to make website development data driven -- we didn't stop building them.

To be fair, regular people use websites. Regular people do not use chatbots in significant numbers

A killer app (well, in this consumerism-driven world) would be a TV ad that ends with "Order now! Just tell your voice assistant 'buy me a pair of the pair of Jimmy Choo's Megan Fox wore in her ad!'".

And it would either have your shoe size saved in your profile, or if you're not the right gender, it will prompt "For who is this pair? Your wife has size 38. Your mistress Mandy has size 36."...

I don't disagree but how is that relevant?

Because even when websites were "bad," the web was still a successful platform with wide adoption. Google Analytics came about because the ecosystem was growing and thriving, not to save something that had failed to gain traction.

So analytics are only useful for successful techs with wide adoption? I think many of those techs got that way because of analytics.

Chatbots should be used to augment humans, not replace them.

For example, one customer service rep can hold discussions with hundreds of people with the help of AI, rather than just 10-20 without.

AI is an augment, not a replacement.

As a digit user, this makes me kinda sad - I actually like the chat interface for basic stuff like checking my balances. Opening an app, logging in, and navigating a UI that may or may not changed since I last used it is far more friction than typing ‘balance’ into iMessage...

I loved Digit, but then they changed their model to charge me more for the same things and the value proposition shifted.

I use Trim now, which actually is more feature rich.

But man I loved Digit's simplicity.

Dear people who have product blogs like this: Please put a global header on your blog that says "X is a Y that does Z." In this case, "Digit is a service that helps you automatically save money based on your income and spending." That way, we get some much-needed context when we stumble across your blog articles talking about what you've learned by building and selling your product.

Very much this.

For quite some time, the original of this was my top-rated HN comment: https://www.reddit.com/r/dredmorbius/comments/27d5xr/please_...

And to add to this, if a product is only available in the U.S it should be marked clearly so. I hate having to look either in the <small> text of your website or the fifth page of your sign up process to find out.

Virtually every HN article would benefit from these descriptors as well.

Especially Show HN. Perhaps I shouldn't be too harsh as I've never personally released a product before, but over half of Show HN's leave me asking the question "what is the actual product" and even more "what is the utility of this product."

Reply on news.ycombinator.com