A year ago this would've been a PhD thesis.
@whats_ai
-
Eliminating Polling in Agent Loops Improves Development
By
–
Moving away from polling in the agent loop is huge. Finally a way to stop babysitting dev servers haha, will play around with it!
-
Claude Code and Coworking Usage Scaling at Towards AI
By
–
That cap moved fast! We push Claude Code and cowork usage to the max at Towards AI and it's still hard to burn $250k without just paying for retries haha
-
Humans Are More Complex Than Fully Controlled Bits
By
–
Humans are complicated. Much more than bits we fully control
-
Memory Consolidation and Consciousness Digitization: Scientific Skepticism
By
–
We barely understand how memory consolidation works, let alone digitizing consciousness. The confidence on this is… something haha.
-
AI Agent Building Obsidian Second Brain Graph Quality
By
–
An AI agent building your Obsidian second brain is exactly the kind of use case that makes agents practical and not just hype. Would love to see the graph quality after a few months of use.
-
Open Source Community Adopts Muse Spark for Image-to-Code Generation
By
–
Super cool to see the open source community putting Muse Spark to work this fast. The image to code with proper asset extraction is next level.
-
Builders Focus on Creation Over Hype Cycles
By
–
Spot on. The people who actually use the models daily are too busy building to post "IT'S SO OVER" every two weeks haha.
-
Visual Understanding in AI: Asset Extraction and Code Generation
By
–
The fact that it extracted assets from the screenshot and used them correctly is the impressive part. That's not just code gen, that's visual understanding.
-
Claude Powers Daily Coding Automation and Content Pipeline Workflows
By
–
In fairness the products are actually great haha. I use Claude daily for coding, automating workflows, and running a lot of my content pipeline at this point.