@TennisPublisher Ben Shelton is definitely in that conversation. Considering his talent level and how early he is in his development, it feels more like a matter of when he will win a major.
@thsottiaux This sounds like an incredible initiative, and I would love to be considered. My focus is on consistently developing new automations and workflows that streamline historically burdensome tasks, enabling teams to work more efficiently and effectively.
@victor_bigfield Word of mouth drives growth, but retention anchors compound it. When a product keeps users coming back, it creates the kind of trust and advocacy that fuels sustainable user acquisition.
Vibe coding is exciting because it helps turn creative ideas into real-world applications.
That said, without effective marketing, even a strong app can go unnoticed.
As SaaS becomes increasingly democratized, the products that rise to the top will be the ones that build trust, foster community, and create meaningful connections with their users.
@sflorimm Simple answer: Claude is a tool. As someone who has experimented with AI, I’ve learned that the quality of the output depends less on the tool itself and more on the skill, strategy, and intention behind how it’s used.
@thsottiaux I don’t really choose models based on benchmarks. I look at what my peers are creating, listen to recommendations from people I trust, and then test things myself. Word of mouth is powerful. For me, GPT models have consistently been the ones that resonate most.
@jahirsheikh8 Challenging you here: if I’ve already defined the roadmap for a SaaS product, why wouldn’t I use AI to build the initial architecture, then sanity-check security, scalability, and infrastructure?
@Timur_Yessenov@composio That’s a good point. The integrations themselves probably become commoditized over time. Owning the trust layer around agent execution feels much harder to replicate.
The concept behind @composio is really interesting to me. Having a single layer for managing API keys and MCP connections feels like a much cleaner way to build AI workflows.
What I’m curious about is where the long-term moat comes from. As MCP becomes more standardized, does the value primarily come from the integrations themselves, or from things like security, reliability, and developer experience?
AI is inevitable in physical products.
If you were starting a company today, what product category would you bet on where AI creates 10x more value, not just 10x better marketing?
If you create design templates in an MCP-connected platform, you can teach an AI the way you like things designed, from layout and hierarchy to your overall visual style. Once those patterns are established, it can recreate them with a high level of consistency. The reality is that great design still requires human oversight. AI is excellent at execution, but taste, creativity, and quality control are still very human skills.
Honestly, you could probably build a GraphRAG with knowledge from professional video editors and then connect it to a Claude or Codex skill that uses FFmpeg plus visual and audio analysis. The end result would be an AI editor that understands what makes a video work and can automate a surprising amount of the editing process.
Genuine question: what’s driving all the excitement around GPT-5.6?
Is it the model itself, or is a lot of the buzz coming from anticipation around Mythos?
Either way, I’m honestly more excited thinking about what GPT-6 might look like if this is the direction we’re heading.