Mike Watson @ Product Party

Dibs!

I burnt my mouth just looking at it. That’s a lot of food lava packed in that shell.

I couldn’t find a record for that but I did find the world’s largest chocolate nut bar was 5,943 lb. So maybe that’s our initial target.

Is that a giant version of these?

Who needs kitchens when…AI.

Are there any fellow vibe coders that have started to “buddy up” on projects with other vibe coders?

I’m wondering if there’s a community of people interested in tag teaming solving problems, a way to find them, examples of how their projects have succeeded faster/been better than solo.

Those plates are cool!

Shipped something this weekend. An AI chat assistant on my portfolio that actually sounds like me, not like every other ChatGPT wrapper.

38 files of my career, my writing, my opinions.

If you've got 2 minutes, go ask it something and tell me if it passes the vibe check.

mikewatson.us/portfolio/

Bookmark ChatGPT - Pin AI Responses - Chrome Web Store Bookmark ChatGPT, Claude & Gemini responses. Pin AI messages, add labels, and jump back to saved responses instantly. 100% local.

Open source project now available on Chrome to install as browser extension.

Bookmark ChatGPT, Claude & Gemini responses. Pin AI messages, add labels, and jump back to saved responses instantly. 100% local.

chromewebstore.google.com/detail/bookm...

Upgraded Kurt Warner.

Damn. Okay. I will buy a box then.

Is this the way? I need to reload on salt.

GitHub - productparty/aichatanchor: A browser extension that lets you bookmark specific responses in AI chat conversations and jump back to them instantly. A browser extension that lets you bookmark specific responses in AI chat conversations and jump back to them instantly. - productparty/aichatanchor

If you use Claude, ChatGPT, or Gemini, you've probably lost a good answer somewhere in the scroll.

I made a custom-built open source browser extension for Chrome and Edge that lets you pin AI responses and jump back to them later.

Check it out:

github.com/productparty...

Want a deep dive on production-grade image handling? I'll be dropping a detailed breakdown on https://mikewatson.us/blog next week. Ping me if you're interested.

Pro tip: Store multiple image sizes/variants during upload. Reduces CDN calls, improves performance, gives you instant thumbnails/previews without reprocessing.

Technical pattern: Layer image sources from most specific → least specific. Personal upload → user avatar → generated avatar → generic silhouette. Zero downtime UX.

Rule #1: Never let missing images break your core user flow. Design multiple fallback strategies - default placeholders, generative AI fills, graceful degradation.

Want a quick audit of your image attachment resilience? DM me. Always happy to give founders a rapid product strategy review.

Real-world approach: Design your image pipeline like a reliability engineer. Every potential failure point needs a deliberate, user-friendly recovery path.

Pro tip: Use probabilistic loading with graceful degradation. If primary image fails, cascade through alternatives without disrupting core functionality.

Fallback image strategy: Always have 3 layers. 1) Ideal branded image 2) Generic contextual placeholder 3) Minimal UI indicator. Don't let missing images break user flow.

Want a framework for turning AI capabilities into actual solutions? DM me. I've spent 14+ years turning complex tech into simple wins.

Tactical AI product test: Can you explain the value in a 10-word sentence a 5th grader understands? No? Back to the drawing board.

Real product secret: Users don't care about YOUR breakthrough. They care about THEIR problem disappearing with minimal effort.

User adoption framework: Pain × Urgency × Easy Fix = Product Potential. Most AI products nail tech, miss human motivation entirely.

Most AI product teams get bottlenecked by capability, not user value. Solving what can be done vs solving what MUST be done. 🧵 https://productparty.us

Building smarter knowledge management isn't about more data. It's about creating intelligent connections. Interested in how? Happy to dive deeper. DM me.

Framework: Document relevance = Semantic similarity + User context + Recent interactions. Traditional keyword search is dead. Intelligent routing is the future.

Radical solution: AI-powered contextual routing. Not just search - intelligent navigation that understands relationships between documents and user intent.

The real problem: Knowledge isn't static. It's dynamic. Current tools treat docs like dead PDFs instead of living, interconnected systems. Context is king.