Are We All Developers Now?

Summary

Paul Ford and Rich Ziade tackle the question of whether non-technical people will actually use tools like Claude Code to build software. Sparked by a recent article from Wharton professor Ethan Mollick about his experience with Claude Code, they discuss the distinction between AI as a consumer “genius butler” (which was an immediate home run with near-zero barriers) and AI as a software development tool (which has much higher barriers to entry despite the promise of democratization).

They draw a sharp line between two narratives about AI and software: the first where classically trained engineers become 10x more productive using tools like Cursor and Copilot, and the second where non-programmers can build their own software. Paul describes attending an Every (Dan Shipper’s media company) meetup around Claude Code and finding that even in a room full of excited tech people, “nobody knows anything” yet. Rich argues that even if you lower the barrier by 5%, millions more people can come over the wall.

The episode lands on a pragmatic middle ground: while tools like Claude Code are genuinely changing what is possible, the flower shop owner in Chicago is never going to use Claude Code directly. Instead, platforms like Airtable, Salesforce, and others will integrate AI to let users customize solutions within their ecosystems. The real opportunity is not a billion new developers, but perhaps five or ten million more people who can ship software — and that alone would be transformative.

Highlights

”The Niche Anthropic Has Found”

yt-dlp --download-sections "*2:39-3:40" "https://www.youtube.com/watch?v=LO01DBQEUiY" --force-keyframes-at-cuts --merge-output-format mp4 -o "LO01DBQEUiY-2m39s.mp4"

“The niche Anthropic has found is around building software with AI. I wouldn’t even call it vibe coding anymore. I would say it’s a new kind of development environment that uses LLMs based on prompts.” — Paul Ford, 2:39

”Nobody Knows Anything”

yt-dlp --download-sections "*5:55-6:40" "https://www.youtube.com/watch?v=LO01DBQEUiY" --force-keyframes-at-cuts --merge-output-format mp4 -o "LO01DBQEUiY-5m55s.mp4"

“I went to an Every meetup around Claude Code… really smart people. Dan is just building apps all day long. But what I noticed reading the room is he and I are aliens. Most people are really trying to get their heads around it. And what did you find? Nobody knows anything.” — Paul Ford, 5:55

”The Consumer Space Was an Absolute Home Run”

yt-dlp --download-sections "*10:12-11:10" "https://www.youtube.com/watch?v=LO01DBQEUiY" --force-keyframes-at-cuts --merge-output-format mp4 -o "LO01DBQEUiY-10m12s.mp4"

“The consumer space was an absolute home run. Partly because of the tech. Because the answers are great most of the time. But also, the usability was probably one of the most basic entry points you could ever put in front of someone… Now you shift gears and you want to conquer other worlds with this thing.” — Rich Ziade, 10:12

”I Own Seven Flower Shops in Chicago”

yt-dlp --download-sections "*15:20-16:20" "https://www.youtube.com/watch?v=LO01DBQEUiY" --force-keyframes-at-cuts --merge-output-format mp4 -o "LO01DBQEUiY-15m20s.mp4"

“I have a chain of flower shops in Chicago. I own seven of them. I’m losing track of who’s got what. And I need software for it. I’m not going to hire some fancy firm. Accenture won’t return your call. You can’t afford SAP. Can AI solve this for me?” — Rich Ziade, 15:20

”When Everybody Can Do Something, It Doesn’t Matter Anymore”

yt-dlp --download-sections "*23:50-24:50" "https://www.youtube.com/watch?v=LO01DBQEUiY" --force-keyframes-at-cuts --merge-output-format mp4 -o "LO01DBQEUiY-23m50s.mp4"

“The minute everybody can do something, it doesn’t matter anymore. There’s no moat left. What’s going to matter is discretion, utility, security, trust.” — Paul Ford, 23:50

Key Points

• Rapid Inflection Points (0:46) - Every couple of months a new inflection point shows up, keeping software builders off balance
• Ethan Mollick’s Claude Code Article (1:35) - Wharton professor’s recent experience with Claude Code shows a non-programmer feeling comfortable in the driver’s seat
• Two Kinds of AI Programming (3:49) - One-shot app building for non-programmers vs. programmer-directed acceleration of existing codebases
• Every Meetup Scene (4:53) - Paul attended Dan Shipper’s Claude Code meetup and found mostly engineers from orgs trying to figure it out
• Enterprise Tool Adoption Barriers (6:13) - People in giant organizations cannot use these tools yet due to policy restrictions
• Engineers Rarely Think About Users (6:25) - Rich asks if Paul has met an engineer who pauses to share power with others; Paul says he knows all of them personally
• Consumer AI Was a Home Run (10:12) - A billion people use AI chatbots because the barrier was near zero and it felt like chatting with someone
• Two Narratives (10:57) - Engineers get 10x more productive (real), or non-engineers build their own software (still struggling)
• Vibe Coding First Wave Failed (12:02) - First wave of vibe coding looked shiny but immediately revealed software complexity; a non-programmer cannot get a custom-tailored CRM
• Mollick’s New Path (12:33) - Claude Code may represent a different path that could actually work for non-programmers
• Blue Sky Sentiment Shifting (13:53) - Even on AI-hostile Blue Sky, sentiment is shifting as people admit coding tools genuinely help
• The Flower Shop Scenario (15:20) - A Chicago flower shop owner with seven stores needs supply management software but will never use Claude Code
• Platform Wars (17:30) - Big platforms like Claude will try to capture all SaaS value vs. small players creating specific value
• Solo Unicorn Fantasy (14:27) - The Valley fantasy of one person building a billion-dollar app with AI is nonsense; building is only one part
• Not a Billion, But Ten Million (25:49) - The real opportunity is five or ten million more people who can ship software, not a billion

Mentions

Companies

• Anthropic (2:19) - Maker of Claude; found their niche in software building with AI
• Aboard (0:08) - Hosts’ company; ships complicated software powered by AI
• Every (4:53) - Dan Shipper’s media company; hosted Claude Code meetup
• Replit (12:28) - Vibe coding tool, insanely well-funded
• Cursor (11:09) - IDE integration for AI coding; making engineers more productive
• Salesforce (17:07) - Will try to lock in users: “We already have all your data. Don’t take it out.”
• Airtable (16:42) - Consumer-grade database tool that may evolve to offer AI customization
• Microsoft (17:20) - Notorious for copying competitors; used Teams to undermine Slack
• Midjourney (10:37) - Referenced as example of low-barrier entry via Discord

Products & Technologies

• Claude Code (2:08) - Anthropic’s coding product; central focus of the episode
• Copilot (11:09) - Microsoft’s co-pilot metaphor for AI as assistant to expert engineers
• Ember.js (14:57) - JavaScript library Rich’s old agency used before React took over
• React (14:57) - JavaScript framework that replaced Ember

People

• Paul Ford (0:00) - Host, President of Aboard
• Rich Ziade (0:02) - Host, CEO of Aboard
• Ethan Mollick (1:35) - Wharton professor and business explainer of AI transformation; wrote influential Claude Code article
• Dan Shipper (4:53) - Runs Every; building apps with Claude Code all day long

Surprising Quotes

“We’ve been building software for over 2,000 years.” — Rich Ziade, 0:54

“If you’re a programmer it’s going to look like it programs really quickly. If you’re a product manager you’re going to look like you’ve now got a team of engineers working for you. If you’re a neophyte it’s going to look like you can now build apps all by yourself.” — Paul Ford, 4:25

“The person who’s never built anything and doesn’t understand engineering can’t have a full-blown CRM that’s custom-tailored. They just can’t.” — Rich Ziade, 12:13

“99.9% of business owners, no matter how good Claude gets, are never going to want to come near Claude Code today. Look at how they just adopt Seamless. Look at how they adopt Squarespace. It’s not their job.” — Rich Ziade, 25:07

“You have two choices. You can say a lot of value is about to get destroyed or you can say I wonder what it’s going to be like with five or ten million more people who can ship software.” — Rich Ziade, 26:23

Transcript

0:00 Hi, I’m Paul Ford. And I’m Rich Ziade. And this is the Aboard podcast. It’s the podcast about how AI is changing the world of software. It is. And we’re gonna play that theme song, but boy is it. Oh my goodness.

0:31 AI changing the world of software. We say that every week. It’s trying. Well, it certainly keeps changing our world of software. We’re trying to build and deploy a product. We’ve got lots of people we’re working with. Seems like every couple of months a new inflection point shows up. We have to sit down and be like, what does this mean for us as software builders who’ve been building software for over 2,000 years?

0:57 When you go to market, you assume that the market goes at about 10 miles an hour. The whole point is you’re going to go 15 miles an hour, a little faster and surprise everybody. And this isn’t an us thing. The level of flat-footedness that’s happening across the industry is wild because the innovation that’s happening — also massive endlessly funded companies are playing around.

1:29 Let’s take a step back. There’s an article by Ethan Mollick who is a professor at Wharton and is kind of the business explainer of AI transformation. What’s great about him is that he’s a tinkerer to no end, but he’s a layman tinkerer who deeply empathizes with his audience which is not a technical audience. He’s doing good and everybody yells at him.

2:03 His most recent is just like, okay, Claude Code, which if you’ve been listening to this podcast, you’ll know that Claude Code is a big deal. Claude Code is a product out of Anthropic. Anthropic has a product called Claude. That’s their LLM. That’s their ChatGPT. It’s like a slightly nerdier ChatGPT, but it’ll do everything. But the niche Anthropic has found is around building software with AI. I wouldn’t even call it vibe coding anymore. I would say it’s a new kind of development environment that uses LLMs based on prompts.

3:10 Let me describe the value. The value is you can describe software you want based on your level of expertise around what software is and what it should do. And you can say, “Hey, make me a game that’s like a clicker game.” And it’ll do it. You’ll get a thing. You can keep talking to it about the thing. You can keep typing in what you like and don’t like, and it’ll keep taking swings.

3:49 Now, there’s another kind which is what I’ve been doing, which is instead of trying to get a one-shot app, give it some existing code and say, “Hey, I want you to use this and do this new thing with it.” And then kind of get in there with a screwdriver. It’s me as a programmer saying, “Can I accelerate what’s in my brain?” Managing a team. I’m essentially a product manager at that point.

4:25 There’s a whole range and I think wherever you’re coming from is going to look really interesting because if you’re a programmer it’s going to look like it programs really quickly. If you’re a product manager you’re going to look like you’ve now got a team of engineers working for you. If you’re a neophyte it’s going to look like you can now build apps all by yourself.

4:53 I went to an Every meetup around Claude Code and it was really useful. Dan Shipper runs Every. I really wanted to check out the scene. Really smart people. Dan is just building apps all day long. He wants access for everybody. But what I noticed reading the room is he and I are aliens, which you’ve pointed out about me many times. Most people are really trying to get their heads around it. And what did you find? Nobody knows anything.

6:00 There are people like Dan who are just like great, this unlocks things that are in my head. But a lot of people are just like, I’m not quite sure where this is going to land. I talked to one person who’s really struggling because they’re in a giant organization and they can’t use tools like this yet. I have so much autonomy. Most people don’t have that.

6:25 Have you ever met an engineer that pauses and is in awe of the power in front of them and says to themselves, “Gosh, I wish I could share this with the rest of the world and empower others?” Very few do. I think I personally know all of them.

7:15 Even if I say, “Here, I took the value. I put it in a bag labeled potato chips and I want to just serve it to you while you lie on a couch,” people are like, “I don’t know.” I want to draw a very strong distinction between the AI as genius butler that seems to answer everything — a billion people are using it — and that’s because its value was immediately recognizable and the barrier to use it was near zero. It felt like you’re chatting with someone.

8:15 It’s an absolute no-brainer to use these tools. But it’s pretty good at making a list of state colleges that have good biology programs. It’s incredibly good at that. That 16 hours of your life just got given back to you. Those use cases get lost because there’s so much drama and everybody involved is so whackadoodle.

9:01 People are using it to shop. People are getting medical advice. The truth is it’s essentially a much more elegant way to mine the internet. It’s just easier than sifting through search results with ads flying in your face. That is a massive resounding success in every way. Now you shift that and what humans tend to do is like wow it did it for that. That means that’s the end of disease. That means you can just stare into a box and see software materialize.

10:12 The consumer space was an absolute home run. Partly because of the tech. Also the usability was probably one of the most basic entry points you could ever put in front of someone. Now you shift gears and you want to conquer other worlds. We can talk about that through two narratives. One is engineers are going to do way better. Classically trained people who know how to make software are going to be 10 times more productive. There’s Cursor that integrates with your IDE. There’s Copilot.

11:37 But I am the pilot. To use Microsoft’s parlance, you’re a co-pilot. I’m the pilot. I’m an engineer. Then there’s this second narrative about changing the world of software which is like there was this promise of democratizing building software so that it’s not just in the hands of engineers and the truth is we stumbled in the first go around. Everyone did vibe coding. All it did was look real shiny in the beginning and then it immediately revealed the complexities of software. The person who’s never built anything can’t have a full-blown CRM that’s custom-tailored. They just can’t.

12:33 Anthropic is wading into new waters. This is what gets to the Mollick piece. He’s sort of a very curious professor of business. And he’s essentially saying, “I grabbed the wheel here for a minute and I’m not a programmer, and I feel like I was moving forward and I understand how the steering works.” I’m not an expert now. I can’t make full-blown enterprise software yet, but I’m starting to feel comfortable in the driver’s seat.

13:30 It’s interesting because he’s looking at these tools and now he’s a tool builder. And I think that’s a really confusing moment. Software is funny as a category because photography is a good example. People patrol forms of communication and they say this is what’s real and this is what isn’t. Code is ultimately interpreted by a computer and it either runs or doesn’t. If it runs in the computer, it’s real.

14:27 The Valley always has this fantasy of the solo unicorn because they can only think about this stuff in terms of how it translates. One guy will be sitting in a room and he will make an app and it will be worth a billion dollars. That’s nonsense. You can go build your thing. Then what? Can you get the LLM to market it?

15:20 I have a chain of flower shops in Chicago. I own seven of them. I’m losing track of suppliers. I need software for it. I’m not going to hire some fancy firm. Accenture won’t return my call. You can’t afford SAP. Can AI solve this for me? I want you to focus on running your flower shop. We are going to get to a point where you’ll be able to type something in the box and it’ll ship software. But it’s got to get hosted somewhere. It’s got to be secure. It should integrate with your accounting software.

17:00 What you’re really going to be doing in the future is you’re going to pick a core platform like “my friendly ERP for small flower shops” and you’re going to go there and it’s going to say what are the things that you need and you’re going to type them into the box and it’s going to customize for you. That’s where your Airtables are going to land. Coming from the other direction, Salesforce is gonna be like, “We already have all your data. Don’t take it out.”

18:30 There’s two ways that value can be created. One is the AI LLM companies capture all the value. Claude is like, “What are you even talking about? I got a baseline here for you called Flowershop.biz.” But you don’t know how my flower shops work. “I’m going to ask you some questions because I’m Claude and I’m going public at $350 billion valuation and I need to capture all of SaaS.” The other extreme is anyone can have any software they want and the big platforms will end up enabling that.

20:30 Back to Ethan. He seems to be pointing at a future where a much broader population beyond engineers are building things they need. That’s coming. But the picture you painted was all about hurdles. You’re neglecting how hard it is to learn how to use a spreadsheet the first time. The barrier is higher. There is a barrier. That’s all you need.

22:00 The engineers are like, “Now I can engineer 10 times faster.” The product managers are like, “I don’t need engineers anymore.” The designers are like, “I don’t need product.” Everybody is just essentially saying, “I’m going to grab all that value just for myself and my discipline.” And the users don’t change their habits. People still think in dashboards. Software has the shape it has for a very specific reason.

23:30 Figure out where the value is and then drive it outwards. What instead I see is everybody wants to capture it. The minute everybody can do something, it doesn’t matter anymore. There’s no moat left. What’s going to matter is discretion, utility, security, trust. Better product.

25:07 99.9% of business owners, no matter how good Claude gets, are never going to want to come near Claude Code today. Look at how they just adopt Seamless. Look at how they adopt Squarespace. It’s not their job. We assume a natural curiosity. But when you lower the barrier 5%, many millions can come over the wall.

26:23 You have two choices. You can say a lot of value is about to get destroyed or you can say I wonder what it’s going to be like with five or ten million more people who can ship software. I think that’s the answer. Not a billion. We shouldn’t try to replicate what just happened. Software skills are actually only going to be more and more valuable and accessible.

28:28 We are Aboard. We are in the middle of this. We ship complicated software powered by AI. Our job is to take the value of this bizarre inflection and drive it to your organization. But you be ready. Hit us up hello@aboard.com. Give us five stars. 2026 going to be wacky, man.