Luka

A.I. recommending restaurants in SF in a chat interface

Discussion

most upvoted
M
Eugenia Kuyda — Luka
Luka’s conversational technology can be applied to other verticals (restaurants are the first step). Consumer AI will be huge, and chat as a UX is intuitive. Luka is the first startup that is building practical consumer AI based on a chat interface now. We're waiting for an update to be approved by the AppStore this week - Luka will get rid of Facebook authorisation (phew) and will be able to reserve tables.
Stewart Rogers — Journalist/Analyst/Speaker - VentureBeat
@ekuyda Any plans for an Android version?
Gillian Morris — CEO, Hitlist
@ekuyda This concept is huge in Asia - WeChat and others already allow consumer AI, and Path's TalkTo functionality also does this somewhat (@tomlimongello could say more on that, I think). And of course Siri was supposed to do all this, though via voice rather than chat. (@libovness's piece linked below has a good analysis of why chat is more promising than voice, at least for now: http://whoo.ps/2015/02/23/future...)

But Siri has never been very good, and I'm excited to see another entrant in the space, especially one in my favorite industry, travel. It's worth considering the limits of what AI will be able to parse, though. There's a great discussion of the limits of chat-based AI in Brian Christian's book, The Most Human Human: http://www.amazon.com/The-Most-H.... Essentially, we can probably get to 90% of use cases, but it takes a book to even start explaining why cracking the final 10% will be so hard.

On another note, there are distinct similarities to what @xdotai does for email scheduling. I love things like this that don't require another app download or making you learn a new UI.
Jonathan Howard — Cofounder, Emissary
@ekuyda I really wanted this to work, but it went the same way Chat AI always seems to: awkward uncanny valley conversation and/or outright failure to get the thing I want.

@rrhoover suggested I share this funny exchange I had with Luka on my first try :) Trying to keep an open mind, so let me know I'm "doing it wrong" somehow!


M
Eugenia Kuyda — Luka
@tomlimongello @libovness @xdotai @gillianim Thanks, Gillian! WeChat is definitely the best example of what a messenger can be - not only for consumer AI but also for offering so many services in a chat interface (it's the biggest mobile wallet in the world for now).
The Most Human Human is perhaps my favourite book on the subject, gives a very clear perspective on what it really means to be 'human' in a conversation.
I believe there are limits at this particular point of history of what AI can do in regards of conducting a human-like conversation, but I strongly believe in assisted AI, where the most difficult cases would be solved by humans (the last 10%), while technology becomes better and better to finally take humans out of the equation.
Next week we're launching as a text number - so no need to download the app anymore :)
M
Eugenia Kuyda — Luka
@rrhoover @staringispolite @rrhoover @staringispolite Hi Jonathan and thanks a lot for your comment - these things make Luka better. We're aiming at a 90% quality meaning that for every 100 replies Luka will have adequate and relevant responses in at least 90 cases. However there are still 10% of mistakes (and right now for Luka it's a little more, around 16%, but it's getting better every day), and those could make for a bad experience for the user. We fixed the issue that we found in your conversation (the place you were looking for is called Red Door Coffee, not Red Door Cafe, which is another place in Bush street - we're learning how to handle all those different ways that people spell restaurants and cafes, and we're getting better with all those typos and disambiguations). The more you talk to it, the better it becomes, and hopefully we can make it so that no users get the uncanny valley conversation like you unfortunately had. Thanks a lot for trying the app out, I would be very happy to see it working for you!

Jonathan Howard — Cofounder, Emissary
@rrhoover @ekuyda Thanks for the reply! Coffee -> Cafe was actually a voice-to-text error I didn't catch until I took the Yelp screenshot above. Weird that it pretended to know the incorrect Red Door Cafe first, then progressively forget more and more context from the conversation. For what it's worth, this uncanny valley conversation was what made me feel like quitting, not so much the result of finding/not finding any one query.

I think I would've kept going if Luka had replied with any of the following instead:
1) Did you mean Red Door Cafe in Pac Heights?
2) Did you mean Red Door Coffee in SoMa?
3) "I didn't find that listing in Soma"
4) "I can't find that - did you mean one of these? [UI list to choose from]")

Also note that Yelp handles the incorrect spelling just fine. If the AI were just searching Yelp and guessing the top result that's in SoMa, it would've worked.

Keep on fighting the good fight!
ilan kasan  — Product and Business Leadership
@ekuyda nice product and great UX. Take a look at www.commersational.com which enables merchants to chat eable their ecommerce catalog and transact via text
Ryan Hoover — Product Hunt
I just gave it a test run (made with Tailor):



Very impressive. I see you have integration with Yelp (and probably other 3rd party services) but how are you deciding which restaurants to recommend and keeping your database up-to-date?
M
Eugenia Kuyda — Luka
@rrhoover Thanks! We collect all data automatically from more than 15 sources - Luka parses data from user reviews (Yelp, Foursquare, TripAdvisor etc) and professional reviews and blogs (from SF Chronicle to Eater) and guidebooks (Michelin, Zagat). We plug all that data into our data model, which is pretty elaborate (we have more than 300 data points for each restaurant, plus menus and photos) and also use some relevant quotes to show our users.
The recommender system is based on a knowledge-based recommender system. Luka's knowledge graph stores all the connections users may have with any possible entities (whether you only eat gluten-free food, or like only the places that are walking distance, or hate chinese) and matches that with all the data we collect about the restaurants (that are part of the graph too) - that makes recommendations hyper personalized and relevant.
Shaun Trennery — Founder X&Go
Love everything about this product. AI is not ready to take the controls alone thou. Here is an example of it going wrong. https://pbs.twimg.com/media/B9JU...
M
Eugenia Kuyda — Luka
@shauntrennery Thanks for your comment! That's an issue we've just fixed - by tomorrow yo won't see negative reviews when Luka recommends a place. Thanks for your feedback!
Irving Torres — Freelance Designer
@shauntrennery I completely agree. My experience thus far has been fun but every time I ask a clarifying question it recommends a new place instead. The grammar is also a little off. Very interesting to see where this goes though!
M
Eugenia Kuyda — Luka
@shauntrennery @irvingtorresyc Thanks, Irving! It gets better the more people talk to it - it's still very raw, but improving every day. The more you talk to Luka, the better the experience becomes for you... Thanks for your feedback - send me whatever you think works wrong at kuyda@luka.ai
Frederick Tubiermont — ceo adsy.me
@shauntrennery what about employing real people instead of just relying on AI? There are people out there who would be ready to become chat assistants I think ;-)
Shaun Trennery — Founder X&Go
@ftubiermont Agree. Prompt agents with one or more options.
M
Eugenia Kuyda — Luka
@ftubiermont @shauntrennery it's great when it's assisting AI in rare and hard cases, but we figured that users don't want to wait for real people's answers as it takes significantly longer (at least 10-15 seconds for an answer, up to a few minutes if you need to give a recommendation, quote a review and send the menu). We tried a Wizard of Oz experiment at the very beginning in Russia to prove the concept, but we figured it would be hard to replicate chat bot experience with humans, as strange as it sounds.
Ben Tossell — Community Lead, Product Hunt
Like I said on Twitter...I think this product is awesome. Such a natural way to discover recommendations. Would love to read about your story and how Luka came about!

Is there a way (or going to be a way) to review after you have been to a restaurant, so that Luka learns your likes/dislikes - i.e X needed better service
M
Eugenia Kuyda — Luka
@bentossell Thanks, Ben, it's really nice to hear that! There is a way to teach Luka - it usually would ask you how it went and what you liked at a place you went to, but that's something we're working a lot on now. You will soon be able to send Luka photos and chat about the restaurant while you're there too.
The main thing about Luka is that people share a lot in a conversation - so the recommendations eventually become hyper personalised. Yelp will never know stuff that people tell Luka - like whether they eat only gluten-free food or like places with hipsters.
Ben Tossell — Community Lead, Product Hunt
@ekuyda Awesome, got ya!

Really interested and keen to see how this develops :)
M
Eugenia Kuyda — Luka
@bentossell Thank you! We'll keep you posted :)
Jonathan Libov — Union Square Ventures
Looks very cool. Added it to a blog post that I've been drafting over the past few days on all the innovation that's happening in text as a medium: http://whoo.ps/2015/02/23/future...
M
Eugenia Kuyda — Luka
@libovness Thank you Jonathan! The blog post is super useful! QuickType experience is awesome. We are actually planning on adding some new cool keyboards to Luka - so that sometimes you get a GUI in a keyboard box (works for picking a time slot for a reservation, or a price range for instance).
Your idea about integrating services that could respond directly to your input is awesome. I believe that's how AI might work - as part of your messaging thread (I think Emu tried a similar approach).
I really believe in texting more than in making Luka voice-operated. Texting as you mentioned allows for much more different types of interactions - for us that means we can add group chatting soon, for example, and many other features that would be impossible if that was all by voice. In the end of the day, voice is only necessary when you're driving, which is a pretty limited use case.
Ryan Hoover — Product Hunt
I didn't realize Luka was in the current YC batch (TC's announcement by @jordancrook). Looking forward to meeting you, @ekuyda and team, at the alumni demo day in a month.
M
Eugenia Kuyda — Luka
@rrhoover Thanks Ryan! Looking forward to meeting you then) and thanks for sharing your experience at diner, it was very useful!
Jonathon Triest — Ludlow Ventures
Love the initial product. Excited for the future verticals. Brave to move the whole team from Russia. Love it.
Brett deMarrais — Partner Ludlow Ventures
I recently read a piece by @alexiskold that I thought was on the nose. The future of search on mobile will be through text. The mechanism of texting a "friend" to find the best restuarant for you is much easier than googling it and scrolling through results.

Luka and apps like it will be game changers.
Corley — COO @ Product Hunt
Just downloaded and started asking for all sorts of recommendations.

The UI is clean and the tour screens are simple. It appears to learn overtime - which of course is a goal, but does not always happen. There was only one question that I asked that I did not like the answer. But that can happen with humans, too. :-)
M
Eugenia Kuyda — Luka
@corleyh Thanks, Corley! That's great to hear) It's getting better every day - more questions that Luka will be able to answer plus it starts recommending new stuff on its own - so that you don't have to read Eater all the time to stay on top of all the openings and don't miss any places you might like! :)
Josh Hubball — Founder @Levelframes
I had a great first experience with this app, love that it will learn and get better/more personalized over time. Favorite feature is asking for pictures of the venue -- seeing pics frequently tips my decision on where to go.
Lisa Q. Fetterman — CEO, Nomiku
Pretty slick app! Very impressed with how far A.I. has come.
Ethan Mantel — Director of Marketing, MoviePass
Seems much more intuitive than adjusting filters and running multiple searches through another recommendation app. Looking forward to an NYC rollout!
M
Eugenia Kuyda — Luka
@eman21 Thanks Ethan! We actually started working on NY first but switched to SF as we got accepted to Y Combinator. We'll be rolling out our New York version soon - we can add more cities almost automatically now. We'll keep you updated on the NY launch!
Frederick Tubiermont — ceo adsy.me
love the idea of using the chat form factor for new applications. Great stuff, can't wait to try it out!
M
Eugenia Kuyda — Luka
@ftubiermont Indeed, with Magic, Cloe and JoinAssist seems like chat if the new big thing) Proud to be part of the trend :)
M
Eugenia Kuyda — Luka
@ftubiermont it would be interesting to hear your first impressions of Cloe, if you get to use it!
Siôn — Software Developer
Met you guys at TC London - really impressed by just how accurate your AI algorithm is. I would really like to discuss more about the UX and who designed it?
M
Eugenia Kuyda — Luka
@sion_smith Thanks, Siôn! We were the first Russian startup at TC Disrupt ever, and that was our first public presentation) By UX do you mean design or the overall experience? Our designer @meatpupped designed everything in the app, from the logo to everything that has to do with UI...
Steve McGarry — Product Grower
Super slick UX. I am excited to see where this goes!
Vivek M George — SALT - trysalt.com
@ekuyda congrats on your launch! And nice design. The chat concept seems cool, but to be honest how is this really different than Foursquare? You have replaced pushing recommendations to a user, with them having ask your app for ideas. As a user I have to trust your recommendations are good and better than Foursquare/food blogs, etc. Why should a user trust Luka? Keep up the great start and best of luck!
M
Eugenia Kuyda — Luka
@vivekmgeorge Hi Vivek! Thanks! Compared to Foursquare it is much more personalised as users share more in a conversation than they would share with Foursquare. Luka also collects data from all user reviews (Foursquare, Yelp and TripAdvisor among others) and professional media/blogs, so Luka's opinion takes into consideration more data.
Luka also started talking to users on it's own sometimes - so you don't have to ask anything, Luka will learn a little about you and then will keep you updated on everything relevant for you - all the coolest places so you don't have to read any blogs to stay on top of it, restaurants you might like that opened in your area etc.
Vlad Nov — Co-founder @kickback
This is an awesome product. Your landing page is beautiful, can't wait to see you guys in other verticals.
M
Eugenia Kuyda — Luka
@mrvladnov Thanks Vlad! We're looking to try applying the tech to other verticals this year. Which verticals would you be interested in?
Paolo Perazzo — Founder, Companyons
The concept and the technology behind are both amazing (read with great interest @ekuyda comments). Will give it a try asap also as I'm always looking for new places in SF :)
M
Eugenia Kuyda — Luka
@sivola Thanks a lot! Let us know what you think!
Luka Bratos — iOS Developer
Brilliant name for an app or maybe I'm just a little bit biased :)
M
Eugenia Kuyda — Luka
@lukabratos It's my favourite name and also the name of my 10 year old friend who I thought would be glad if I named the app after him (he loves it:)
Luka Bratos — iOS Developer
@ekuyda me to! I wish you all the best and I'm looking forward if you'll also expand to London at some point.
M
Eugenia Kuyda — Luka
@lukabratos Definitely coming to London too - we first showed what we do there at TC Disrupt, so we really are looking forward to launch in London!
Eddie Wharton — Data Scientist, Dinner Lab
I think text is a great front end for an app like this and restaurant recommendations are a great first use case for this type of AI. Can you discuss the underlying tech? Does this use some type of active learning? Are you humans ever involved in edge cases?
M
Eugenia Kuyda — Luka
@eawharton Sure! It's all automatic - there is no human involvement whatsoever.

All the tech is proprietary (although we use NLTK open source library) - and it's a combination of machine learning and rule based approaches. We use machine learning to extract of linguistic properties of the user input and collect data from user and professional reviews - we use SVM to train classification models and other machine learning techniques for that. We also do use some heuristics and thesauri-driven entity extraction for natural language processing. Understanding natural language in a conversation is hard - there are many issues like ellipsis reconstruction ('I like sushi' - 'Me too' - how do you understand what 'me too' stands for) and pronominal anaphora resolution ('Do they have wifi there?' - you need to understand what 'there' stands for) that have to be solved - for that we are using deeper analysis techniques.

Our conversational manager is rule-based mostly and allows anyone even not knowing how to write a line of code to come up with new dialog logic. It's like building blocks that you can put together in any combination. Our dialog engine allows you to take into account any context, user preferences, restaurant details etc. It also allows us to tweak the tone of voice whenever we want and have multiple tones of voice for Luka.

This is a very high level overview - let me know if you're interested in anything about our technology!

and thanks for trying Luka out)
Roberto Scaccia — I build bikes. I do what I love
@ekuyda I love what you guys have built...Luka is "clever" and knows what to answer also to off topic questions ;) (ie "what is the meaning of life"). Question: do you keep user behavior into consideration? For instance, different users may use different language patterns, that can be modeled as a Markov Decision Process, do you support user-specific pattern identification or do you plan to do so?
M
Eugenia Kuyda — Luka
@robertoscaccia we are now taking into account general linguistc patterns of how people ask questions, make comments or follow-up in the conversation. However we're now building a more context-aware model of user linguistic behavior. We are experimenting with MDPs and other Markov models, and as soon as we get enough training data, Luka will get much smarter and more context oriented. Thanks for your question!
Roberto Scaccia — I build bikes. I do what I love
@ekuyda Thanks for your answer, it is really fascinating what you guys are doing. It reminds me the old fashioned Cleverbot (http://www.cleverbot.com/) on steroids...
M
Eugenia Kuyda — Luka
@robertoscaccia yeah, I used to play with Cleverbot, they even have an iPhone app now! It's fun because it stores all the answers by real people and then match them with other answers that were given in a similar situation. That's why it answers "what's your name" in a different way every time :)
Ativ — LIT
Does this remind anyone of Smarter Child on AIM? I had a similar idea a couple years back, cool to see someone executing. Congrats on the launch!
M
Eugenia Kuyda — Luka
@ativslife We do have hilarious conversations the remind me of Smarter Child too))) Thanks a lot!
Kishan Gupta — Co-founder, UXCam
Downloaded and used the app today. The results are accurate, add to that the superb onboarding and UX. In love with this app now !
M
Eugenia Kuyda — Luka
@kishan_guptas Thank you so much! We really appreciate it :)
Alain — The Startup Conference
How do I integrate this with colunchers.com ? That would be fun!
M
Eugenia Kuyda — Luka
@alain94040 That's a cool idea! :) Thanks for the tip. The idea is to integrate as many services as possible - from table reservations (in the update waiting AppStore approval) to food delivery and paying your bill through the app and potentially even more interesting stuff like Colunchers.
Spencer Schoeben — College dropout working on something new
Sorry but can we not call this AI.
M
Eugenia Kuyda — Luka
@netspencer Why? What is your definition of AI?
Commenting is limited to those invited by others in the community.
Login to continue.