home.social

Matt Williams

Matt Williams is a full time content creator on Youtube. Before that he was the first Evangelist for Ollama and Datadog, Emcee for DevOps Days Seattle, and a former organizer of DevOps Days Boston and Serverless Days Boston. Ages ago, he was also at OpenText, Placester, Yottaa, and lots of other places that don't exist.

Posts
292
Followers
102
Following
30
Joined 2022-11-02 · View on fosstodon.org →
  1. Join @technovangelist and @ryan_maclean for the weekly stream about all things tech. We start in just 15 minutes here: youtube.com/live/aJyRMUaK2n8

  2. Join @technovangelist and @ryan_maclean for the weekly stream about all things tech, almost every Tuesday at 2 US Pacific time. Find the next one here: youtube.com/live/aJyRMUaK2n8

  3. Join @technovangelist and @ryan_maclean for the weekly stream about all things tech, almost every Tuesday at 2 US Pacific time. Find the next one here: youtube.com/live/aJyRMUaK2n8

  4. Model selection in the new UI could use some improvements (filtering!). But hey, it's v1! Looking forward to updates. youtu.be/prrWESXl7wg

  5. New to AI? The UI might be the perfect, easy-to-use entry point. It's a fantastic start! 🎉 youtu.be/prrWESXl7wg

  6. k-quants are genius! 🧠 Multiple mailrooms for numbers = smarter organization & even MORE efficient AI. youtu.be/K75j8MkwgJ0

  7. Okay, I'm officially lost in the Bon Entendeur abyss. It's triggering *serious* Life in Mono vibes. Anyone else feel this way?

  8. @tailscale lets you securely access services from *any* device – iPhone, iPad, android, laptop! 📱💻 youtu.be/byr434VBeFk

  9. Ever wish your workflows could interact with your local Obsidian notes? 🤔 It's possible! I'm showing you how to bridge the gap. youtu.be/G_aep31hYsk

  10. No more guessing games! 🕵️‍♂️ 's new 'think' feature cleanly separates the model's internal thinking from the content. Easy to enable - just 'think': true in your API request. youtu.be/yBD598s5g8c

  11. Flash Attention + context quantization = massive memory savings! 💾 Unlock even more power from your AI. youtu.be/K75j8MkwgJ0

  12. Base64 encoding: a key to AI image processing! 🔑 Understanding this can unlock a new level of creativity and customization. Learn how I'm leveraging it! 📚 youtu.be/8n_tpLn6Xbo

  13. Stop exposing your local services to the internet! Use @tailscale to create a secure private cloud infrastructure. ☁️ youtu.be/byr434VBeFk

  14. Quick start: Download an q4 model, enable Flash Attention, & test! 🔥 Unleash the power of AI. youtu.be/K75j8MkwgJ0

  15. Unlocking the power of local scripts in your workflows! 💥 Learn how to use Webhook to make it happen. youtu.be/G_aep31hYsk

  16. Ready to create your own AI documentary? 🎬 The possibilities are endless! What creative AI projects are you dreaming up? Share your ideas in the comments! 👇 youtu.be/8n_tpLn6Xbo

  17. Memory-hungry AI models stressing your system? 🤯 Quantization is the clever trick to making them fit! youtu.be/K75j8MkwgJ0

  18. Cloud + Local GPU = Best of both worlds! 🤩 No need to spend $$ on cloud GPUs. Learn how with @tailscale. youtu.be/byr434VBeFk

  19. Don't just use the highest settings! 🚫 Find what *works* for YOU. Perfecting your AI setup is about finding the right balance. youtu.be/K75j8MkwgJ0

  20. Bringing it all to life with ElevenLabs! 🗣️ AI-generated voices add a whole new dimension to my project. The future of storytelling is here! ✨ youtu.be/8n_tpLn6Xbo

  21. My UI says "Error 404: UI not found" 💀 It WORKS, but my eyes are suffering 😩

  22. Sunshine ☀️ & a brand new workflow ⚙️. Today's a day for getting things done, not just sitting on the patio. Who's with me?!

  23. Want to leverage (Model Context Protocol) without the hassle? Learn how to set up a powerful server on for with @tailscale and

    Save $200+ over 2 years with this method 🔧
    youtu.be/OmWJPJ1CR7M