Vibe-Coding Has Lawyers in a Chokehold — But Let’s Slow Down
If you’ve been anywhere near legal Twitter or LinkedIn this month, you’ve seen it, lawyers “vibe-coding” entire AI apps over a weekend using Claude, Gemini, Manus AI, and similar tools. Suddenly, everyone is building something. As someone who reviews enterprise legal software, I couldn’t help but wonder, what happens to companies like Spellbook, Legora, Harvey AI, and others if an entire industry starts building its own tools?
Before you get swept up in the frenzy, let’s address the elephant in the room:
If you can type a prompt and generate an application, why pay hundreds—or thousands—of dollars for enterprise software?
But that’s the same question lawyers were asked when generative AI exploded onto the scene:
If I can ask ChatGPT to draft a contract, why do I need a lawyer?
The Rise of the Lawyer-Builder
Let me be clear: it’s exciting to see attorneys move from consumers of legaltech to creators. Especially in niche areas of law, where no legaltech company is coming to rescue you. These pockets of unmet demand are exactly where vibe-coders are thriving. And you probably aren’t alone in your niche pain point. I recently read about a banking law associate who built an app over a weekend, and it was rolled out firm-wide. I also saw someone start vibe-coding for fun and end up raising venture capital to turn it into a business.
Jamie Tso recently demonstrated this momentum by recreating lightweight versions of Harvey, Legora, and Spellbook’s core features using Gemini 3—in a single weekend. He even hosted a hackathon for lawyers to build legaltech tools. The results were impressive. Many of the apps addressed ultra-specific use cases that traditional legaltech companies would never build around. Everything from that hackathon was compiled and shared as open source. And honestly, the comment section was rich with unique ideas for people who want to create. (By the way, I have an idea bank on my website that you should check out if you want to join a legaltech team.
So, Why Am I Not Vibe-Coding? Simple. Just because you can build something doesn’t mean you should. When I started my YouTube channel, I was actually coding products, long before no-code platforms and tools like Manus AI existed. It was time-consuming, yes. But more importantly, I realized something after observing a founder I featured early on—someone with a Microsoft background. It wasn’t that I couldn’t build apps. It was that I could never produce the same level of quality he could, because we had built completely different knowledge bases over the years. No matter how powerful the tools become, expertise compounds. My time is better spent elsewhere— drafting, analyzing, reviewing, and educating—rather than competing with engineers and product teams who do this at scale.
The Data Protection Problem
Here’s my first real concern: data protection. Most lawyers posting these apps are not using them in actual client matters. And I’m not entirely convinced your law firm would allow you to plug client data into a weekend-built AI tool without serious scrutiny. At best, you’d face compliance hurdles. At worst, the firm could take your idea and internalize it. Beyond that, maintenance is the elephant in the room. You built it, great.
Now who is:
Maintaining it?
Updating it?
Monitoring security vulnerabilities?
Fixing hallucinations?
Ensuring reliability?
Horace Wu raised what I think is the strongest counterargument: the risk of malpractice. Lawyers sell competence. What enterprise software sells to lawyers is reliability. The most dangerous tool isn’t one that crashes; it works quietly while producing subtle errors.
Legaltech Fatigue — Or Market Overload?
There are thousands of legaltech tools on the market. I recently pulled up the Legaltech Hub’s directory and found more than 3,000 legaltech tools. And that’s probably an incomplete list. So maybe vibe-coding isn’t just creativity. Maybe it’s a sign of legaltech fatigue, and lawyers are saying. “I’m not even going to try to find the right tool. I’ll just build it.” But let’s be honest, many tools cost around $100 per month, often covered by law firms. Compared to what lawyers bill clients, that’s not outrageous. We attorneys, justify our fees to clients all day. Yet we hesitate to justify legaltech subscription fees. If Legatics already bundles signature pages and manages transactions effectively, why rebuild it? If Spellbook exists, why rebuild it? If price is the concern, there are alternatives that are better and cheaper. Rebuilding existing solutions isn’t innovation. It’s duplication. It's flooding an already saturated market.
The “I Can Do It Cheaper” Illusion
Horace Wu, on a LinkedIn post, made another analogy that stuck with me. You walk into a grocery store, see the prices of food, and think: I could grow this cheaper in my backyard. Then you try, and you realize the farmer is doing far more than you ever imagined. The same applies to insurance. You might think: why not just save my own emergency fund instead of paying premiums? Because scale, expertise, risk management, and infrastructure are invisible until you try to replicate them. Enterprise legaltech companies employ security teams, engineers, compliance officers, and support staff. You’re not just paying for features. You’re paying for systems.
Are Legaltech Companies Overhyped?
Let’s ask the uncomfortable question. If lawyers can recreate “lightweight” versions of Harvey or Spellbook in a weekend, what exactly are billion-dollar legaltech companies building? Is there a legaltech bubble? Possibly. Some companies restrict public demos. Some marketing pages oversell differentiation. I’ve long advocated for greater transparency in product demos so lawyers can see what they’re actually buying.
But here’s the nuance. Seeing the surface features of a product is not the same as understanding its infrastructure, reliability, integrations, support, and security. A lightweight clone is not the same as an enterprise-grade system. Still, the tension is ironic. Legaltech companies have been racing to automate entry-level legal work. Now lawyers are trying to cut legaltech companies out entirely. When AI entered the legal industry, people said, “You don’t need lawyers anymore.” “The billable hour is dead.” Yet legal fees continue to rise. Some people absolutely use ChatGPT for drafting and advice. But sophisticated clients still pay for expertise, judgment, and accountability. The same dynamic will likely play out here. Small firms may rely on vibe-coded tools. Large firms will continue paying for enterprise-grade software with security, compliance, and reliability baked in. The gap won’t close. It will widen.
My concluding remarks
Lawyers pay for CLE courses, memberships to associations, directories, and conferences. How is paying for one more tool—especially one that saves time—so different? When you see tools priced at $30,000 per year, it’s easy to feel like you’re missing out. When you see cheaper tools, you might feel ripped off if you think you could “just build it.” But here’s what I believe. There is probably an excellent tool within your budget. And the company that built it is trying to reach you. In a few days, I’ll be posting a video on one of those companies, Brackets. Until then, experiment, learn, and build if you want to. Just don’t confuse a weekend build with an enterprise solution.
Speak to you soon.
Take care.