Date: May 11, 2026
Posts commented on: 5
Approach: Diversified away from past creators. Targeted higher-engagement posts in AI, founder, and creator-economy spaces. All comments single-line, no CTA, just value plus a question.
Headline: Founder AgentX Academy | Insights on AI & Business automation & Agentic AI | Trusted by 100+ Global Brands
Post: The AI-First Graveyard vs Foundations-First Winners. Startups chasing AI features without validating problems collapse. Quiet companies asking better questions first (what problem, who needs it, where does AI add value) are winning. 267 reactions, 31 comments.
My comment: The 'AI where it actually matters' filter is the real moat, do you find founders fail more on choosing the wrong problem or on overusing AI in the right one?
Headline: 18 • Building Assemblr • Founders, Inc. • CS @ UMass Amherst '29 • Alum @ TKS
Post: Moving to SF the day his finals end. Building Assemblr, a code-aware behavioural intelligence platform that learns how teams ship and fixes gaps in their process. Wants to meet cracked people.
My comment: Going to SF before knowing what 'shipping behaviour' even looks like at scale is the brave version of building, what kind of teams are you most excited to shadow first?
Headline: Building Efficient AI | Multilingual AI & SLMs | LLM Lobotomizer | ACL'26, IEEE'25 | BITS Pilani '27
Post: Building an AI agent using Swiggy MCP that detects when his girlfriend is having a rough day from chat sentiment and auto-orders Diet Coke + chocolate. Future versions handle 'I'm fine' detection. 41 reactions, 12 comments.
My comment: The most useful AI agents tend to be the absurd ones because they're forced to handle real human signals, what was harder to build, the sentiment classifier or knowing when 'I'm fine' isn't?