back to index

Why Bolt.new Won and Most DevTools AI Pivots Failed - Victoria Melnikova


Chapters

0:0 Intro
0:40 About Evil Martians
1:18 What doesnt work
1:55 Boiling the ocean
2:24 Building the tooling
2:40 The biggest killer
3:0 Step 1 Identify your unfair advantage
3:43 Step 2 Identify how AI can amplify your competitive advantage
4:26 Step 3 Create a category

Whisper Transcript | Transcript Only Page

00:00:00.120 | December 2023, StackBlitz board meeting.
00:00:03.520 | Eric Simons walks into the room knowing that his company is dying.
00:00:08.400 | Seven years of building this revolutionary web container technology, millions raised,
00:00:13.280 | but the revenue is practically zero and the growth is practically flat.
00:00:18.520 | The board gives him an ultimatum.
00:00:20.800 | You show us real progress by the end of 2024 or we shut this down.
00:00:25.380 | Just two months later, Bolt.new is showcasing explosive growth.
00:00:29.780 | 20 million ARR in just two months and today, 18 months later, we expect to see another update
00:00:36.500 | crossing over 100 million ARR for Bolt.new.
00:00:40.740 | This overnight success didn't really happen overnight.
00:00:43.880 | It took seven years in the making.
00:00:45.840 | And today, I'm going to tell you exactly how that happened so you can replicate it for your
00:00:50.780 | business.
00:00:51.780 | I'm Victoria from Evil Martians, the go-to agency for developer tools.
00:00:55.620 | We've been StackBlitz technical partners since 2021, back when nobody believed in browser-based
00:01:01.540 | technology, web containers, and investors thought that this was the dead end.
00:01:06.580 | We work with 40-plus early-stage developer tools startups each year and everybody is asking the
00:01:14.100 | same question: how do I bring AI into my product?
00:01:19.380 | First, let's talk about what doesn't work.
00:01:22.420 | The most common mistake I see is let's sprinkle some AI into our UI.
00:01:28.100 | And the problem with that is people are very tired of chatbot interfaces at this point.
00:01:33.540 | It's very hard to convince yourself to try something new to test new features because
00:01:39.220 | oftentimes the results are underwhelming.
00:01:42.420 | So if you are just slapping on a chatbot, you are taking a risk.
00:01:46.500 | Your customers might just lose interest and never try it again.
00:01:50.020 | And the result will be very sad as well.
00:01:52.500 | A quietly deprecated feature.
00:01:55.220 | The second common trap I see is boiling the ocean.
00:01:58.580 | When you are trying to AI power everything, you're competing with the big guys like OpenAI and Anthropic.
00:02:05.540 | And the reality is you just don't have that big of a team and that big of an investment behind you.
00:02:10.740 | Instead, try to think about what you can leverage.
00:02:14.420 | What's your unique domain?
00:02:16.340 | What's your unique expertise?
00:02:18.100 | What's your unique technology?
00:02:19.540 | If we bet on that, the chances that we succeed are much higher.
00:02:23.380 | Then there are companies that are waiting for the AI to be perfect.
00:02:27.140 | And the reality is we have to build the tooling as we go.
00:02:31.300 | It's up to us.
00:02:32.180 | If we get scrappy, if we get creative with our solutions, we will stay ahead of the game.
00:02:38.500 | But the biggest killer?
00:02:39.620 | Not talking to users.
00:02:42.420 | I've seen many companies reinvent the wheel.
00:02:45.860 | Invest a lot of money, a lot of resources into building beautiful,
00:02:50.260 | intricate interfaces that users don't need.
00:02:53.860 | And the reality is no market, no product.
00:02:57.060 | Talk to your users first.
00:02:59.620 | While everyone else bolted AI into their existing workflows,
00:03:05.460 | StackBlitz did something different.
00:03:07.540 | I want to take you through three simple steps that you can apply to your business
00:03:12.820 | and try to replicate StackBlitz's success with Bolt.new.
00:03:16.580 | Step 1.
00:03:17.860 | Identify your unfair advantage.
00:03:19.780 | Not product features, but your unique competitive edge.
00:03:23.140 | Let's take StackBlitz.
00:03:24.660 | They're the only company that can run a full development environment in the browser.
00:03:29.780 | They don't need containers.
00:03:31.460 | They don't need servers.
00:03:32.740 | They don't need local setup.
00:03:33.940 | What's your unique advantage?
00:03:36.500 | It could be something like domain expertise or data distribution or infrastructure.
00:03:42.340 | You name it.
00:03:43.220 | Step 2.
00:03:43.780 | Figure out how AI can amplify your competitive advantage.
00:03:48.580 | So instead of thinking, how do I bring AI into my product?
00:03:52.180 | Think, what becomes possible when AI meets my unique capability?
00:03:56.900 | For StackBlitz, it was pretty straightforward.
00:03:59.220 | Every other pipe coding tool had the same issue.
00:04:02.100 | The process is just too complicated for an average user.
00:04:05.460 | So instead, you describe your app with words.
00:04:08.580 | You see it right there in the browser.
00:04:10.740 | And you can share the URL with the public.
00:04:12.900 | Easy.
00:04:14.980 | It only became possible because they had this unique competitive advantage.
00:04:19.700 | So now it's your turn.
00:04:20.820 | What becomes possible when AI meets your unfair advantage?
00:04:26.500 | Step 3.
00:04:28.020 | Last but not the least.
00:04:29.460 | Create a category.
00:04:31.300 | Instead of thinking in existing workflows, think about how your capability can change the flow
00:04:39.700 | altogether.
00:04:40.340 | If we take a look at StackBlitz, this is called inventing a category.
00:04:45.860 | They reinvented how people vibe code.
00:04:48.820 | Why this moment matters?
00:04:50.340 | The reality is StackBlitz's secret wasn't unique AI.
00:04:54.340 | Everybody has access to Anthropic.
00:04:56.340 | It was their unique competitive advantage.
00:05:00.180 | And once they figured it out, they reinvented the user flow and they bet everything on it.
00:05:07.140 | They went all in.
00:05:08.100 | So I want to encourage you to take a look at your product.
00:05:13.540 | Be honest.
00:05:14.340 | Sometimes it's painful.
00:05:16.020 | Take a look at your product and think.
00:05:18.820 | What's your unique mode?
00:05:20.740 | What's your competitive advantage?
00:05:23.940 | How can AI amplify that advantage?
00:05:26.820 | Can you reinvent a workflow?
00:05:29.780 | Can you create a category out of this process?
00:05:32.260 | If you find answers to all of those questions, you're on the right track.
00:05:37.220 | Good luck.
00:05:39.380 | Thank you.