Summarized using AI

Startups on Rails in Past, Present and Future

Irina Nazarova • July 08, 2025 • Philadelphia, PA • Talk

Overview

The talk, delivered by Irina Nazarova at RailsConf 2025, explores the enduring appeal and current momentum of Ruby on Rails (Rails) for startups, examining its historical trajectory, current success stories, and what is needed for its future growth.


Main Topic

The session investigates the ways in which Rails has helped startups grow from inception to IPO and its resurgence among new founders, using concrete examples from both well-known and recent companies.


Key Points

  • Recent Startup Successes on Rails:

    • Chime Financial completed an IPO, exemplifying Rails' capacity to scale from an idea to a public company.
    • Figma, less known for its use of Ruby, also filed for IPO, using Sinatra and ActiveRecord in its backend.
    • Bolt.ne achieved $40 million ARR within five months of launch and ran the world’s largest hackathon, all products served by Rails and ActiveRecord.
    • Whop (an e-commerce platform targeting Gen Z) processed $1.4 billion in transactions within a short period after its 2021 founding on Rails and Next.js.
    • Uscreen, a platform for creators, raised $150 million, demonstrating continued investor faith in Rails-powered startups.
  • Rails' Historic Trajectory:

    • Review of the Gartner Hype Cycle, illustrating how Rails experienced early hype, a period of migration away by some companies, and ongoing improvement as the community invested in tooling and the ecosystem.
    • The foundational role of Rails in the rise of major startups (e.g., Shopify, Twitter, Heroku, GitHub, Stripe, Intercom, GitLab, Coinbase, Instacart), many emerging from early Y Combinator (YC) batches.
    • The cyclical pattern of Rails adoption, hype, and renewal underscored by strategic investments in open source and infrastructure during visibility slumps.
  • Current Rails Community and Adoption:

    • Strong present-day momentum with an increase in meetups (800 in the last 12 months) and many new startups across fintech, AI, e-commerce, healthcare, and other sectors choosing Rails for its productivity, conventions, and rapid iteration capabilities.
    • Founders value Rails for conventions over configuration, ease of building business logic, community, expressiveness, and efficient prototyping.
    • Renewed advocacy from YC partners and successful entrepreneurs, encouraging new founders to leverage Rails for full-stack development.
  • Challenges and the Path Forward:

    • Noted gaps: need for a more comprehensive frontend story, scalability for larger teams, improved hiring/training resources, enhanced advocacy, and especially robust tooling for AI/ML-related workflows.
    • Despite recent SDK releases for Ruby in the AI ecosystem, there remains a need for better support for AI workflows, code generation, observability, and integration with Python-based data pipelines.
    • Emphasis on the community's responsibility to build missing pieces, promote Rails successes, and ensure relevance, especially in the age of AI-powered startups.

Conclusion & Takeaways

  • Rails' unique combination of developer happiness, productivity, and a mature ecosystem continues to make it a compelling choice for startups aiming for rapid growth and large-scale success.
  • The recent wave of high-profile Rails-powered IPOs and startup success stories signals a renaissance, driven by collective community investment and new technological advancements (e.g., Hotwire, Inertia, Kamal).
  • Ongoing challenges—especially around AI tooling and advocacy—represent opportunities for the community to further strengthen Rails’ role in powering tomorrow’s most innovative startups.
  • Networking, advocacy and knowledge sharing, as showcased at events like RailsConf and local Ruby meetups, are key to sustaining this momentum.

Startups on Rails in Past, Present and Future
Irina Nazarova • Philadelphia, PA • Talk

Date: July 08, 2025
Published: July 23, 2025
Announced: unknown

In 2025 Ruby on Rails continues to power successful startups: during the last 12 months one of them became the leading AI-powered app builder, another raised $150M, and another filed for IPO. But let’s put things into perspective and explore how the essence of Rails’ appeal to founders since the beginning and into the future.

We'll journey through Rails history with pivotal stories from Intercom, Gusto, and Y Combinator - companies that leveraged Rails' productivity to achieve remarkable growth. These legendary examples demonstrate which aspects of Rails and its ecosystem were key for bringing products to market and sustaining explosive growth.

Then we'll celebrate recent Rails success stories—such as Bolt.new, Whop, Chime and Uscreen—examining the tremendous success these startups have posted recently. We'll explore the key factors behind Rails' resurgence among founders who are choosing the framework in recent years, including Hotwire, Inertia, Kamal and more.

Finally, we'll take these learnings and look into the future by identifying what we can all do to ensure Rails continues being the choice for tomorrow's entrepreneurs. We'll discuss tooling, open source, educational resources, and community voices that could catalyze even more startup success stories on Rails.

RailsConf 2025

00:00:19.680 everybody? Awesome. Yeah, let's let's go into it. Just a sec. Yeah, we'll start
00:00:29.199 with the news. First of all, this company Chime Financial, a financial app built on
00:00:37.040 Ruben Rails. Um, co-founded by Ryan King from San Francisco and Noel right here
00:00:44.160 is who's also speaking at this conference later. Uh, is leading the I love Ruby team at Chime. So, I'm
00:00:50.719 assuming you guys all know this company. They completed the IPO. Woo!
00:00:57.680 Congrats. And this is important because this is the latest fulfillment of the promise of
00:01:05.280 Rails that it's going to take you from idea all the way to IPO. And we see this
00:01:11.280 happening right now, right in June. Now, another Ruby company, Figma, which
00:01:18.799 not everybody knows that, is using Ruby on the back end. Sinatra, active record
00:01:23.920 in addition to a lot of web assembly, rust and go. Um they filed for I for
00:01:29.759 IPO. So we might have two Ruby IPOs this year. Can we have three? Um this company
00:01:38.159 Bolt.ne. a very popular new web coding application
00:01:44.880 also built on Ruby and Rails back end uh and a lot of uh web assembly um and
00:01:51.520 remix uh co-ounded by Eric and Albert from San Francisco
00:01:57.040 uh they posted an incredible $40 million ARR from after five months from
00:02:03.920 launching Bolt the their new product which is pretty historical honestly.
00:02:10.080 Uh they also ran the world's largest hackathon uh last month with about a
00:02:16.239 million projects built and on bold but every single one of them was served by
00:02:24.239 Rails and Active Record. Keep it in mind. Uh this company uh also new a
00:02:31.680 relatively new Rails startup from 2021 co-founded by Jack Shy from New York. um
00:02:39.200 built on rails and Nex.js. This is a Gen Z e-commerce platform
00:02:45.519 they posted. So this is a special chart chart I got from Jack. This is their
00:02:52.239 growth of their transaction volume and the platform is now serving $1.4 billion
00:02:59.040 transactions. And again, this is a new startup that just they just founded in 2021 on Ruben
00:03:06.239 Rails. this company uh I'm sure somebody knows uh Nick who
00:03:13.440 co-ounded it. Nick is from Maryland. This is a product for creators
00:03:19.760 and in February they raised $150 million which I think is obscene. I don't know
00:03:27.599 why you need this amount of money. Uh again built on rails. Um, this guy
00:03:32.879 Carmine Paulina, he built Ruby Lalam and Ruby Lalam went to the on the to the
00:03:38.720 front page of Hacker News back in March. Incredible conversations and like a
00:03:45.440 pretty big deal. something else. Um, a partner at YC, Tom uh Blumfield is
00:03:52.879 recommending Rails publicly uh in a response to Andre Karpathy from OpenAI
00:03:58.560 saying that hey Ruby on Rails is actually great for uh a great option for being this kind of full stack uh
00:04:05.280 framework for VIP coding. So let's let's use Rails there. He so
00:04:11.680 again why commoditative partner is publicly pitching rails and he also
00:04:17.519 built this VIP coding uh playbook for for Y comative founders and he's he's
00:04:25.040 actually saying look I am vip coding a rails app and I have more success with
00:04:30.639 that than some other folks who built in JavaScript stack. Why? Well you got to
00:04:36.080 watch this. Um, so where are we? Let's put this this is these are all the
00:04:42.479 recent news, but where are we? So we see startups on Rails succeeding, growing.
00:04:49.280 We see IPOs, which is great. We see Ruby opensource getting popular, getting on
00:04:54.960 the front page, getting in front of people. And we also noticed that investors again
00:05:02.080 publicly support Ruby on Rails which was my dream for some reason. Um but let's
00:05:08.880 put this into a bigger perspective. So apparently there's this chart that is
00:05:15.759 a great illustration of what I felt was happening in the Rails community in the
00:05:21.520 Rails ecosystem and it's called Gartner hype cycle. It's just exist out out there. So apparently
00:05:30.560 so what it what it depicts is the visibility of any great new technology.
00:05:36.800 So when we have this new technology people learn about it the expectations
00:05:43.280 skyrocket immediately. It's just our human nature. We always expect
00:05:49.520 everything at once. It's kind of natural. But the problem is
00:05:55.440 um at the peak of those expectations the technology itself is not yet ready. It's
00:06:01.680 not yet ready for all the production use cases where it is kind of trying to be applied. And it's just the reality of
00:06:08.479 things because technology is not an idea. Technology is decades of work of
00:06:16.639 multiple groups of people put into it. It's not just a single idea. Never.
00:06:22.000 So, but we as human society, we expect it to be
00:06:28.080 something more of an idea, a sparkle, and we expect everything to happen at
00:06:33.520 once. And this is at the peak of those expectations, people realize that, oh, well, actually,
00:06:41.360 it's not ready yet for my use case. It's not as fast maybe or performant or
00:06:47.520 something like that. and and then they drop it and and the problem is we see those kind of stories
00:06:54.960 of people migrating off of the technology that they've just picked and
00:07:00.960 yeah then the visibility drops and at the lower point we see people asking
00:07:08.000 you know what they ask is Ruby dead
00:07:13.280 yeah is it dead yet that's what they ask but this is exactly the test. The test
00:07:20.319 of is it is it going to be a real thing? Is it going to be a real story? And the
00:07:27.840 test is if people continue building investing their time and their lives,
00:07:33.039 the time of their lives and investing as organizations into building out tooling,
00:07:38.479 building out the core, building out the whole ecosystem around this technology to actually support all of those
00:07:44.400 production use cases that that are needed. And if this is happening then pragmatic value of the technology is
00:07:51.680 growing all the time. It's it continues to grow and at some point people begin
00:07:57.520 to think or it's actually great. We dropped it but
00:08:02.560 hey um what if we if we go back to it now?
00:08:08.319 It looks amazing. It looks better than anything we have right now. And why? Not because it stayed the way it was, but
00:08:15.599 because it continued getting getting better and better. This is the the reality. And this upward trend is called
00:08:23.280 they call it at Gartner hype cycle they call it slope of enlightenment.
00:08:28.720 Calling it a cool term. Um I don't know. So this is where I believe we are today.
00:08:36.320 So we had this kind of hype in the past and today we've uh we are on this
00:08:43.120 upwards uh trend. So where are we going to this is what we're
00:08:49.360 going to talk today uh about. But first I'm Arena uh from Evil Martians uh
00:08:56.880 co-founder of any cable and I run SF Ruby meetup San Francisco uh meetup and we'll talk a little about that at the
00:09:04.160 very end. Uh but at Devil Martians, we work with startups. Um and we write a
00:09:11.040 lot of code. I hope you maybe recognize you love some of our open source. Just a
00:09:18.000 little portion of it is here. Um and we also talk to users and that's what I
00:09:23.839 want to do right now. Please take your phones uh because I want to talk to you.
00:09:31.920 Yeah, let's do it. So, please scan this uh QR code.
00:09:39.120 Nice. So, as a little warm-up, uh as a little warm-up, just type your
00:09:45.600 name. Just type your name. Yeah. Please no bad words. It's not, you know,
00:09:54.959 perfect. Okay. Joy is winning. Nice. Ben, hey, even
00:10:01.680 um Mikey, Josh, Lucas. Yeah, great to see all of you. Okay, keep the comment
00:10:07.040 for a sec and then we'll move on to the next one. Uh yeah, I kind of I kind of love this
00:10:13.839 moment. So, this is actually another Ruben Rails startup that is running this
00:10:19.200 uh it's called Poll Everywhere. It was co-ounded by Brad Gestler from San Francisco. It is one of the first it
00:10:26.800 went through white com one of the first patches and yeah and the real time is
00:10:32.959 served by any cable prop that's why I'm so proud um yeah so uh and by the way
00:10:38.880 Brad is doing a bunch of new projects the latest is beautiful ruby.com check this out so again uh if you didn't have
00:10:47.279 the QR code this is your chance because we'll do we'll do another one uh right
00:10:53.440 now. Okay, everybody scanned.
00:10:59.519 Uh, please. Okay, let's go. So, when did you first learn Ruby? Put a
00:11:06.480 the year the year 2017 2020, 2015,
00:11:14.959 2021. We We're seeing a lot of new new folks here. That's great. That's great.
00:11:22.560 20 2012. Okay. Okay. I feel like Okay. There's a there's a mix. There's a mix.
00:11:29.120 Um. Yeah. Nice. All right. Love it. Anybody?
00:11:35.279 Anybody? 2025. No. Nice. Good job. All right. So, this is
00:11:44.160 this is great. Um, now can This is a harder question. Why what do you love about Ruby and Rails? What do you love
00:11:51.120 about it? Why why are you doing this? And why are you here? Right?
00:11:56.959 So, it's you're not only using it, you're also going to events, conferences, but yeah, community,
00:12:04.720 simplicity. Yeah, simplicity is great. Expressiveness,
00:12:10.079 productivity, ease,
00:12:15.360 conventions, joy. Yeah, I love this. Elegance.
00:12:22.079 Ah, perfect. RSpec
00:12:29.519 alliteration. Nice. Um, AO,
00:12:36.639 I know who did that. Who did that? All right. Um, okay. I love this. Okay.
00:12:42.480 Community and simplicity. Okay. We're seeing those two big things, right? Um,
00:12:47.519 so this is this is pretty awesome and I mean th this is why we're here but
00:12:55.839 um now we move move on. Um, pleasure to meet you and
00:13:03.440 we're going to talk more about startups today. Uh, because I believe the the
00:13:08.800 core of this community is in building startups, building new businesses,
00:13:14.000 building ambitious projects. uh you will see the questions you just answered were the questions that I asked startup
00:13:21.279 founders for the second year uh in a row and a bit more but first
00:13:28.720 uh these are the Ruby and Rails not all Rails uh but all Ruby startups that uh
00:13:36.320 of course are not all the obviously not all the stories out there but those are important stories that I
00:13:43.440 we all should know about it's Dylan from Figma. It's Adam from Hiro. It's Edward from Gastra. It's Apurva from Instacart.
00:13:50.079 Morton from Zenesk. Toby from Shopify. Tom from GitHub. John from Maki. Ryan
00:13:56.000 from Chime. Kieran from Intercom. Nate from Deximity. John from Stripe. Brian
00:14:01.440 from Coinbase. Deitro from GitLab. And finally Gary Tan who's not the founder
00:14:06.880 of YC but he built a an internal social network in YC that became like a
00:14:13.680 superpower of YC founder on rails. So this is internal YC tooling that is
00:14:19.920 continuous being built on rails. Um and I want to go uh a little back in
00:14:26.959 time so we look at those stories in the kind of historical perspective. It's going to be quick. I can promise. um
00:14:35.600 20 minutes and we're back. So let's let's look at this. Apparently in 2005
00:14:43.440 uh not only Rails one was introduced but also why combinator was founded was kind
00:14:48.800 of maybe it has something to do with that and also the same year obviously 15-minute blog from David goes viral.
00:14:57.440 Next year already Shopify, MRAI and Twitter are all founded on Ruby on Rails
00:15:03.360 the year after and there's some references to open source projects
00:15:08.639 just arbitrarily for you to uh just kind of remember those years right to kind of
00:15:15.040 think about the corresponding years. So um next year Heroku Zenesk uh are
00:15:21.120 founded on rails and also like as we learned um just an hour ago power home
00:15:26.399 started there uh huge Ruby monolith um then 2008 we have GitHub Airbnb
00:15:34.480 co-ounded and apparently already in 2009 Twitter starts starts migration off of
00:15:39.519 rails you remember the quick cycle think that's what happened there um So 2010
00:15:47.920 stripe deximity and what happens is Hioku Hioku the company that was
00:15:53.440 building uh a the cloud for rails apps originally. It was acquired by
00:16:00.079 Salesforce uh very quickly and it was one of the first uh it was Hero itself
00:16:06.240 was in one of the first batches of YC and the story goes that the money we got from this deal financed Y combinator for
00:16:13.360 many years after yeah but not only that 90% of those uh
00:16:18.720 of those batches in YC were on rails and this means that this was the the the
00:16:25.440 height the high of the hype cycle Because again, if 90% chooses Rails, it
00:16:32.079 doesn't mean they do it pragmatically. They analyze. No, it's just the default right now. They pick it because it's
00:16:38.240 there. You got to be aware of why people are making those decisions, right? Uh so
00:16:44.800 what happens next? We have new companies. Um GitLab Intercom 2011 and
00:16:51.040 2012 is the year astrologologists proclaimed you have to start a company on Rails. um just there's no way around
00:16:58.720 it. Chime, Figma or Ruby like Figma, Chime, Figma, Gasta, Instacart, Coinbase
00:17:04.000 are all founded that year and Mira is acquired by Gh by Cisco.
00:17:10.079 Then in the next few three years, we have consecutive IPOs. Twitter that is
00:17:15.919 migrating off of Rails, but then we have Zenesk and Shopify. By the way, your screen that you saw, uh, it's they
00:17:21.760 started as a Shopify store the year Shopify went and IPOed. Yeah, maybe
00:17:27.520 that's why. So then the next five years, okay, yeah, we
00:17:33.440 saw Steisov is founded, but there's kind of nothing else.
00:17:39.120 There's really no uh big Ruby stories founded. Aore founded. Oh, nice.
00:17:46.720 Okay, there are new what I'm trying to say is we don't see big news from Ruby
00:17:52.960 and Rails companies. Uh we see something else. I think when uh these are the
00:17:59.600 years when we see this kind of deflated visibility of Ruby and Rails but at the
00:18:05.440 same time companies and also people not just companies people invest in building out the ecosystem. So this is when we
00:18:12.080 see falcon um any cable test prof and also white jit work starts then so
00:18:20.080 we see people you see what's going on right people
00:18:25.200 continue building it out this is actually the main determinance of what
00:18:30.559 is happening now this is the work that started then and um now suddenly we see
00:18:38.080 a bunch of IPOs gitlab coinbased daximity um Instacart, Chime, this is this is
00:18:46.000 this is all happening in the last few years and I we see a growing adoption a growing adoption and growing visibility
00:18:52.640 of Rails but I hope you you follow the and again we see a lot more new startups
00:18:59.120 a lot more new projects but this this this is what we're going to talk about
00:19:04.240 in our present part you see we're back
00:19:09.679 now we're back and we're going to talk about the present, what's happening right now who is building on rails new
00:19:16.000 projects today. And I spoke with a bunch of people like Brian um from New York
00:19:21.120 building six-fold AI AI workflow platform for insurance underwriting.
00:19:27.120 Damian uh from Warso building vendor e-commerce solution they also maintains pre-commerce. Now Ian from Detroit
00:19:34.799 building good bill uh hospital bills negotiation software. Ask from Denmark
00:19:40.559 building open pay uh build bill splitting app. Uh Jake from San Francisco building sparity health
00:19:47.200 preventive health platform. Scott from New York uh building sub layer DX for AI
00:19:53.200 automations and also is running artificial Ruby meetup. Eric uh from
00:19:58.480 Virginia open source Cupid a search uh quality testing platform.
00:20:03.919 Avin from Dallas uh building craft food delivery from private chefs powered by
00:20:10.960 AI for automation. Nidita built uh from Bengaluru building trem line uh release
00:20:17.919 management for mobile apps. Bark from London building fertility opens kind of
00:20:24.160 like operating system for fertility clinics customer. Um Ken uh from Sydney
00:20:30.159 building clipflow software for creators. Victor Compass from Brazil building uh
00:20:35.760 V360 in voice management system. Jacob from Arkansas building ordinal AI assistant for local governments. Uh
00:20:43.760 Glenn U Luther AI from San Francisco building AI agents that validate
00:20:49.440 compliance of marketing campaigns. Very hard. Uh Kelly from Seattle building
00:20:55.200 scholarly software and e-learning platform for faculty. Uh Erin Snyder uh from uh North Carolina
00:21:03.520 building core pilot an AA powered ERP for aviation
00:21:09.520 aftermarket. Uh Jesus Ryan from from
00:21:14.960 Atlanta building terminal the e inc uh display that you can program. Uh Thiago
00:21:21.200 from Brazil building uh digital bank. Uh Carmine from Berlin building u chat with
00:21:28.320 work AI assistant. Uh Joaw from Brazil building Cashi an invoice financing app.
00:21:34.880 Trent from Alabama building equip a platform for organizations helping
00:21:39.919 people with autism. And Muhammad from Morocco building test skills and
00:21:45.200 alternative to testing gerilla. This is not all. I tried to squeeze them into
00:21:50.240 one slide but those conversations kept coming. We have Colleen who built Simple
00:21:55.520 File Upload from San Diego and Matthew from New York who built uh Rescale sorry
00:22:02.000 Rescale Rescale company um that helps
00:22:07.679 work with food manufacturing vendors as you can see all sorts of all sorts of
00:22:12.720 new startups. I also spoke with some of the older guys um you screen uh Mark
00:22:18.880 from Battleist Ryan from Chime and Eddie from Gusto. So, a bunch of people, a bunch of feedback and what I asked them
00:22:26.480 were a few questions. What one is why Rails and why Ruby? And I also chatted
00:22:33.520 with Dylan from Figma and he said, "I love Ruby."
00:22:38.640 That's that's fair. Um, so, uh, Thiago said, "Look, uh, it's this story, by the
00:22:45.360 way, similar to what you've just listened to the one you've just listened to from John. Uh, yeah. when you're
00:22:51.360 switching from Java, you're immediately noticing, yeah, I'm I'm just going much faster.
00:22:57.360 Uh Barkin uh said, "Look, convention over configuration allows us to build to focus on the business logic, just focus
00:23:03.840 on our business and our products." Uh Ryan from China said, "Look, Rails uh
00:23:11.120 really helped us iterate when we were searching for the right products, when
00:23:16.400 we were trying out lots of things. uh uh Rails as the grownup in the room,
00:23:22.640 a lot of work put into it. Uh Rails developers uh being more entrepreneurial and
00:23:29.600 product and businessoriented as as the another aspect of it. Uh
00:23:35.280 Kelly points out that Ruby is over represented among successful startups. So you will not see it over represented
00:23:42.080 among new startups. But if you look at the successful ones and the ones that are actually growing, you'll have you'll
00:23:49.360 you'll have a different picture. And finally, Glenn said this look few years
00:23:55.280 in people will look back and see and say all those years all those successful
00:24:00.960 companies were actually building on Ruben Rails. This is what's going to happen.
00:24:06.400 And he's a like San Francisco Y combinator founder. Exactly the the the type.
00:24:12.880 But let's also look into the future. And I think our look into the future is based on what's missing.
00:24:19.840 What people point out as problems as the the missing pieces
00:24:25.679 and like I hope I convinced you the last time uh I think this is up to us. That's
00:24:31.520 why I think we should listen because the core of Rails, the core of Ruby have a
00:24:38.000 bunch of work already and those things is something that the community and the
00:24:43.120 ecosystems are um really working on. This is coming
00:24:49.120 from us. That's what I'm trying to say. So there were a bunch of points. I will not get into all of them. Uh still the
00:24:57.200 comprehensive front end story is missing. uh still like the path to performance
00:25:03.360 is a is still a question. Um yes for larger company uh for larger companies
00:25:09.440 big problem is hiring and and training engineers and um by the and finally the
00:25:15.760 there is one point about advocacy which is something that we kind of miss
00:25:21.520 from the hype um cycle is the time when all the SDKs were in Ruby and if you
00:25:28.400 think about this this is the reason why we should be loud and we should talk about successes on Rails that happen
00:25:34.960 today because we want those SDKs. You want people that build whatever click
00:25:40.320 custom portal any any company they should build Ruby uh Ruby SDK if they
00:25:46.000 don't want uh only new apps but they also want successful companies among
00:25:51.840 their clients right and future future successes but I hope you can guess what
00:25:58.080 was the number one topic uh we discussed in the missing part with all those
00:26:03.520 founders based on what they build right based bas on the descriptions, the short
00:26:09.120 bits. Um, what do you think it is?
00:26:15.440 Yes, it's AI. It's AI. Seriously. Um, so
00:26:23.200 everybody I spoke with, everybody's using AI. Let's Let's be honest. Um,
00:26:29.520 half of these startups would say, "Look, we're pretty early. We're fine storing
00:26:35.440 prompts in code. It's fine. Yeah, we don't need any a lot of special tooling.
00:26:41.840 We're fine. But then the other half actually said, look, it's been two or
00:26:47.120 three years already and there's there's not enough tooling in in the Ruby and AI
00:26:52.960 ecosystem. They're missing bits and pieces and we we want to build in Ruby.
00:26:59.120 We want to build our product and business logic in Ruby, but we're struggling.
00:27:04.480 And what exactly they're struggling with? By the way, we have Ruby LLM, we have Lchain, Ruby, we have more. We
00:27:09.600 finally have Antropic and OpenAI Ruby SDKs. They released them in April.
00:27:16.799 But this is this is not enough. This is just just kind of scratching the surface. What are we talking about? Like
00:27:23.279 first of all, of course, code generation. Uh code generation is something everybody's using, right? We
00:27:28.400 use assistants so you of all sorts. Uh the the few questions are is the lack of
00:27:35.039 static typing is going to slow us down and this is this is the question we should all consider. Um what like the
00:27:43.760 question why is why is this the question? Because the assumption is if we have types and we notice that kind of
00:27:50.960 like TypeScript uh code generation in Typescript works pretty well. Um this
00:27:57.120 might be one of the reasons. Of course the of course the assumption is that we
00:28:02.720 have very short feedback loop uh when we have types. The answer is well when we
00:28:09.520 have a test suit we also have a pretty short feedback loop but again it it
00:28:14.640 remains to be figured out I will say. Um also do we have enough code to train on
00:28:21.600 remains remains to be answered. Finally, the vibe coding story. The story of
00:28:27.120 people building applications in those all of those bold lovable replet etc. Pe
00:28:33.600 people build new applications they don't have rails there in in the stack that
00:28:40.000 um that is out there. Usually they build NodeJS and all sorts of kind of JS um
00:28:47.200 they can use all sorts of JS frameworks. But there's an answer to that. come to our talk on Wednesday. On Thursday, by
00:28:53.120 the way, on Thursday. Yeah. Um, we might hint on something. We might have
00:28:58.159 something actually. Uh, but there is also a an interesting project from OB.
00:29:03.440 Check this out. Cloud on rails. And there is an interesting project from Jezeim Tidewave.
00:29:09.600 Uh, there's missing MCP tooling. I will not go too deep into this. Uh, but yeah,
00:29:15.039 we need a Ruby SDK at least. But we have fast MCP now. Check this out. Uh now if
00:29:22.559 you think about all those startups that use AI, they they now have something
00:29:28.159 they call usually they call it data pipelines or they have kind of data engineering components in in their stack
00:29:35.360 and that part should remain in Python because this is
00:29:40.720 what data engineers know, right? But
00:29:45.840 um we want all the product engineering to be to be on rails. All the product
00:29:51.360 all the business logic we want to keep it on rails and how exactly to separate the two um I would say it's not yet
00:29:59.919 clear but also not only how to separate them but also to give to bring all the missing tool into rails and also to
00:30:06.799 ensure a connectivity between let's say our data pipelines and our product stack. Uh sometimes we need streaming,
00:30:14.159 sometimes we need um background jobs, we need different types of things here. Um
00:30:20.399 all the agent tooling um again we we need a lot. We need new abstractions uh
00:30:25.840 in Rails. We need new infrastructure because um agentic workflows that are
00:30:32.159 what are they? They are kind of undeterministic longunning workflows similar to background jobs but they
00:30:40.080 require because of the undeterministic nature of them we need more observability more control more
00:30:46.159 resumability more context. So it it is it requires uh additional um and maybe
00:30:52.559 new infrastructure. uh observability becomes central when we are building non-deterministic workflows
00:30:59.919 because we don't know what happening by default. Um and finally we need quality
00:31:05.520 evaluations and we need other new things like browser agents um which are not
00:31:11.840 available in rails yet. So we have here active agent uh acidic job leavar it's
00:31:18.000 just a few examples there's it's actually a booming open-source field
00:31:23.039 check this out um there are a couple books uh the book by Obi Fernandez it
00:31:29.120 lays the grounds pretty well uh there's also a new chapter in the upcoming edition of layered design book by my
00:31:35.840 colleague VVA and there's also a uh newsletter uh Rubyai uh
00:31:43.600 beehive beehive.com. Sorry, it's called Robber Ruby actually. So, check check out this newsletter. It
00:31:50.080 collects all the links. Yeah, I I I know I should wrap up. Just just a couple more slides. Uh you know how many
00:31:58.320 meetups we had in the last 12 months in the Ruby community? 800.
00:32:10.960 we're doing great. Um, when we look into the past, I would say it was tons of fun. Interesting.
00:32:18.640 Not Not everybody gets this joke. Okay. Um, the future is up to us, up to people
00:32:24.559 in this room for real. And the present, I want to say, has never been better.
00:32:30.399 Never been better. And yeah, this is my honest opinion. I think when you ask
00:32:37.039 when people ask you, "Oh, Israel's great for a new startup," I will say it was never better than today.
00:32:48.240 Yep. And with that, so that you have something to look forward to, we're
00:32:53.919 running the San Francisco Ruby Conference. This is where you're going to hear from those startups. You're
00:32:59.120 going to connect with them. Uh it's in November 19th, 20th. You're going to you're going to have amazing experience.
00:33:05.679 We are supported by a bunch of sponsors. We welcome additional sponsors and these
00:33:12.240 people already are confirmed speakers and the CFP is open right now till this
00:33:17.919 Sunday with the idea that you all will submit a proposal. U yeah thank you.
Explore all talks recorded at RailsConf 2025
Manu Janardhanan
Christopher "Aji" Slater
Hartley McGuire
Yasuo Honda
Ben Sheldon
Chad Fowler
John Athayde
Mike Perham
+77