Peter Wang

Peter Wang

Just a lazy guy building lazy tools
12 points

About

I am a full-stack software engineer with love of new techs, done projects in blockchain, social media, machine learning, business applications etc. I was made in Taiwan and I been out side of country for years, does that make me a "MIT but drop out"? My main focus now is building Agentainer, a platform that provides for future AI application infrastructure.

Badges

Tastemaker
Tastemaker
Gone streaking
Gone streaking
Gone streaking 5
Gone streaking 5

Forums

Dustin Heaps

6mo ago

What’s the last product that genuinely made you go: 'Whoa — this is smart'?

I m always on the lookout for those aha moments where a product nails the timing, design, or simplicity so well it just clicks.

What s the last product that made you pause and think:

Damn this is clever ?

Bonus points if it s something small but brilliant. Curious to see what s been impressing you lately

Yanjun Lin

6mo ago

Are Vertical AI Agents still worth building?

I recently came across some news:

Claude for Financial Services officially launched in mid-July. It s designed specifically for the finance industry, integrating data from platforms like PitchBook, Morningstar, Snowflake, S&P Global, and Databricks to support market research, due diligence, and investment decisions. During early testing, Claude Opus 4 hit 83% accuracy on complex Excel tasks.

It really makes me question how much room there is left for AI Agent startups ----LLMs are getting better at handling more and more tasks on their own.
For an AI Agent to have long-term value, it must be able to understand, remember, and adapt to a user's evolving preferences and context ---- something LLMs still struggle with due to their limited memory and continuity.

Peter Wang

6mo ago

How We Built a Solution Runs Long-Lived LLM Agents

Introduction

Most cloud platforms AWS, GCP, Azure are optimized for stateless web apps or short-lived serverless functions. But deploying long-lived, stateful LLM agents is another beast entirely. You need durability, resilience, and observability. When we tried to push our own multi-agent AI system to production, we hit walls with all the complex infrastructure work that not only took hours but unstable.

View more