community
Agentic AI for Software Engineering
Community Insights
More Insights
The last mile in AI takes a lot of effort to get it right
New skills engineering teams need while adopting AI
From AI skeptic to an advocate
Don’t expect AI to do magic on large, abstract tasks
Leaders must actively re-engage skeptics to overcome early hiccups with AI
The problem isn't just generating code - it's knowing how to fix it when things go wrong
AI adoption beyond developer productivity
“Instead of calling it developer productivity, I like to call it developer quality of life.”
“Instead of calling it developer productivity, I like to call it developer quality of life.”
Insider insights
Josh Grose
I heard a profound statement from an engineer during a demo this week.
They said, “when I see a demo, the first think I ask myself is, do I think this product works?”
The second question they ask themselves is, “but do I REALLY think this product works?”
That’s the state of AI and demos today. And no wonder, yesterday I watched a hype demo of a launch of a competing product. In a produced demo, the slack message time stamps had the agent responding prior to the prompt being given. That truly is magic 🪄 ✨ 😉
Prototypes are easier than ever. To be successful selling AI, you need both a product that actually delivers, and the knowledge and ability to explain why it works, and the savvy storytelling to help customers see the possible.
Spiros Xanthos
Systems of Record are not just dead, they are dead weight, and we are moving from Systems of Record to Systems of Knowledge.
The data is valuable but what's even more valuable is tribal knowledge scattered across them: decisions, workflows, tribal know-how. Much of it never makes it into a system at all, it lives in minds, meetings, and Slack threads. And it walks out the door when experienced people leave.
Systems of Record require tremendous effort to maintain and use. And their biggest moat is how hard it is to migrate from one to another. In the era of agents they have become the bottleneck: rigid UIs, narrow and slow APIs, workflows optimized for humans.
Humans don't need data they need answers, and agents don't need a UI. They extract existing knowledge from systems and learn from every interaction becoming the System of Knowledge that drives the business.
Mayank Agarwal
The culture and momentum you set in the first couple of years becomes the blueprint for everything that follows.
Remote setups are great for getting things done individually. But in the earliest stages, startups thrive on clarity, speed, and shared direction.
What we've gained from being together in-office: • Ideas spark more naturally through serendipitous chats • Feedback happens in minutes, not on calendars • Teams develop a shared mental model faster • Architecture and strategy align more quickly • Everyone stays grounded in the same vision
In our experience, the compound effect of team energy and real-time collaboration beats the efficiency of isolated workflows.
