Woody Song

Feels like voice is less about replacing typing, and more about shifting how we think while building. When I tried similar setups, the biggest difference wasn’t speed — it was how it changes the “loop”. Typing feels like issuing commands, but voice feels closer to thinking out loud. That said, I still struggle to see it working in longer or more structured tasks (like reviewing diffs or...

This is interesting. Feels like most agent tools focus on capability, but not visibility. Curious how you handle debugging in real workflows. When something goes wrong, can you actually trace what happened step by step, or is it still mostly inferred?

I think the trust boundary is less about the tool and more about the type of action. For me it usually looks like: - high trust: analysis, drafts, research - medium: suggested actions with approval - low: anything irreversible (money, messages, system changes) Feels like most issues happen when those boundaries aren’t clear. Once agents cross from “suggesting” to “acting” without friction,...

Interesting approach. Feels like the hard part here isn’t scraping, but turning that data into something actually useful. Curious what typical use cases look like in practice.

This is interesting. Feels like we’re starting to treat AI outputs more like artifacts that need to be tracked and reviewed, not just ephemeral responses. Curious how you think about evaluation over time. For example, do you see this becoming something like a feedback loop where past sessions actually improve future workflows?

I think it comes down to two things: risk and ownership. Equity usually goes to people who take on long-term risk and build something that compounds over time. The tricky part is that roles like personal branding or growth can absolutely drive revenue, but they’re often seen as something you can swap out more easily. In reality, a great operator in those roles is just as hard to replace. Feels...

I actually ran into something similar recently. What’s frustrating is not just the limits, but how invisible they are. You don’t really know what triggered it until it’s too late. A local tracker makes sense, but I think the real value would be showing patterns, not just counts. For example: - how fast you’re doing actions - repeating similar actions in short time - sudden spikes vs normal...

This feels like a shift from “design tools” to “design systems that generate themselves.” What I’m curious about is the boundary between generation and control. At some point, especially in production, teams don’t just need good output - they need predictability and constraints. If Stitch keeps evolving layouts and structure across iterations, how do you prevent drift from the original system...

This direction feels inevitable. Once agents start improving their own workflows, it stops being just a tool and becomes more like a system you’re managing. The part I keep thinking about is control. If the system keeps evolving its own setup, how do you keep things predictable in production? Especially for real workflows, stability often matters more than raw capability.

Moving agents from cloud sandboxing to direct local execution feels like a major shift. Most real workflows live across files, terminals, and apps — so bridging that gap unlocks a lot of practical use cases. Curious how you handle safety boundaries when executing local commands. Is there a permission system for different levels of actions, or does every command require manual approval?

Interesting question. From what I've seen, it's less about founders not wanting families and more about timing and environment. Early-stage startups tend to consume a huge amount of mental bandwidth, not just time. Even when you're technically "off work", you're still thinking about the product, users, or the next experiment. I also know founders who built families and companies in parallel —...

Instant AI demos for SaaS is a really interesting direction. A lot of B2B products struggle with the friction between discovery and actually seeing the product. Curious how interactive the demo experience is compared to a real product environment.

Currently building AI developer tools focused on improving AI workflows. Lately I've been experimenting a lot with: • Python / FastAPI • LLM integrations • automation for developer productivity Curious what stacks other builders here are using lately.

One approach that worked well for many developer tools is: Free → core usage so people can understand the value quickly Paid → advanced workflows, automation, or scale If the free tier lets people experience the real value, conversion usually happens naturally once they rely on the tool.

For me it usually comes down to two things: 1. A very clear tagline that immediately explains the outcome 2. Something that feels like a real tool, not just a demo Developer tools especially stand out when the product solves a small but very real workflow problem.

Love seeing developer environment setups shared like this. Is this mainly optimized for Claude Code workflows or does it also work well with Cursor / other AI coding tools?

This feels like the natural evolution of maps. Is the AI answering queries using real-time map data or more like a conversational layer on top of existing search? Super curious how navigation + AI interaction will blend here.

Turning a phone into a brain performance coach is a cool idea. Are the exercises based on neuroscience research or more behavioral training? Would love to know how the progress tracking works.

This looks really interesting. Are the repeatable instructions more like automation workflows or closer to scripting for the computer? Curious how flexible it is for developers.