The Bondi AI Collective returned for its third event to WOTSO Bondi Junction, bringing together an eclectic group of builders, creatives, technologists, and curious minds—all exploring the fast-evolving frontier of AI.
From property tech to party planning, finance to front-end code, this session was a deep dive into how people are actually using AI right now—and where it’s taking them next.
✨ Opening Prompt: What Are You Learning About AI?
We kicked things off with a simple but powerful nudge: What’s one thing you’re currently learning—or want to learn—about AI?
The answers were refreshingly diverse:
- Kunal is diving into “vibe coding”—a term gaining traction for the way non-coders are now building apps through conversational prompts.
- Ross is experimenting with no-code agentic tools like Flowise and Langflow, and looking to go deeper with frameworks like crew.ai.
- Stefan shared how he’s stitching together voice pipelines and planning to integrate OpenAI’s code interpreter for more advanced automation.
- Phil, who runs an automation agency in Bondi, is building across industries—from legal to luxury lifestyle—and trialling WhatsApp-based voice agents for surveys and lead gen.
Others spoke about their journeys moving into AI from finance, content, software, or even waste management—each with a unique angle and ambition.
🛠️ Tools in Focus: Lovable, Bolt, Manus & More
A highlight of the event was Kunal Gupta‘s very insightful practical showcase of AI-native app builders like Lovable, Bolt, and Manus—tools that allow anyone (yes, even those who “can’t code”) to create working apps in minutes.
One standout example was a property management tool built by Kunal during a long-haul flight using Lovable. What started as a simple spreadsheet headache turned into a fully functioning dashboard—customised to his family’s real estate assets across multiple countries. The kicker? He did it in a few hours, with patchy in-flight WiFi.
Another example: a social media post analysis tool built for a friend’s agency after a casual dinner conversation. A two-line prompt turned into a working prototype with a landing page and upload functionality—generated in a single shot.
The power here wasn’t just speed—it was about access. These stories showed how AI tools are enabling ideas to come to life before anyone writes a single line of code or hires a dev team.
💬 Healthy Debate: Is This “Real” Software?
Naturally, not everyone was convinced.
Some developers in the room raised valid concerns about quality, reliability, and whether AI-generated code can be trusted in production. Others emphasised the importance of rich product requirement documents—treating prompts as the new spec sheets.
But there was broad consensus around this idea: AI is becoming the junior dev you can talk to, and when paired with a thoughtful human, it’s getting scarily good.
Even skeptics admitted surprise at how accurate tools like Claude 3.0 are becoming, especially when guided well.
🧠 From SaaS to “Software of One”
A recurring theme throughout the night was a quiet revolution underway in software development. Instead of one-size-fits-all SaaS tools, we’re now seeing the rise of what some called “Software of One”—hyper-custom apps built by individuals for their exact needs.
Think: personal ERPs, custom CRMs, tools for tracking real estate, managing creative projects, or planning destination weddings. Built fast, owned fully, and modified as needed.
One participant put it this way: “This is the worst these tools will ever be—and they’re already game-changing.”
🧩 What’s Next? Agents, Standards & a Voice-First Future
The conversation naturally expanded to the broader implications:
- The shift from front-end UI to voice-first, agent-driven computing.
- A need for new standards to help AI agents talk to each other.
- The blurred line between “good enough” code and production-grade infrastructure.
We also touched on the looming macro questions: How does this all hold up in a recession? What’s the countertrend? Who’s being left out?
One attendee asked a simple but powerful question: Is this AI boom strong enough to survive a Trump-era downturn? The room didn’t have a single answer, but the optimism was tangible—this tech, many felt, is too useful to stall.
🔚 Final Reflections
As we closed, everyone offered a brief thought, takeaway, or provocation.
Some spoke to the democratising power of low-code tools. Others shared concerns about data privacy, technical debt, and over-trusting AI. One saw it as a way to prototype and raise capital, fast. Another saw it as a mirror—forcing us to ask not just what can we build, but what should we build?
There was a general sense that this moment in tech feels different. Faster. More accessible. More uncertain.
But mostly, more human. Because ultimately, it’s not about the tools—it’s about the people using them.
