Create README.md

README.md

@@ -0,0 +1,31 @@
+# GitHub Universe 2024
+
+I met with Andrea Griffith at GitHub Universe 2024. I'm sharing the conversation with her here.
+
+## Videos
+
+- Original recording on Dropbox (30 GB): https://www.dropbox.com/scl/fo/rzozc1bvo5smcz5lqwess/AApMQuZ8Jfs0MaxE93Bcyow?e=1&preview=02+GH24+Zeke.mov&rlkey=7vn7ty83ddheyd5bvw2jfxq8v&st=w7zpea40&dl=0
+- Demo of Flux fine-tuning on Google Drive (240 MB): https://drive.google.com/file/d/1pjVWan3SeR8Fkw45_nY_jLTcqLUTwp1a/view?usp=sharing
+
+## Timestamps
+
+02:01 - https://octogatosconf.com (GitHub's employee resource group for Spanish-speaking employees)
+03:24 - https://github.com/andreasjansson (Replicate co-founder)
+04:57 - https://cog.run (a tool for packaging and running AI models as containers)
+05:15 - https://replicate.com/orpatashnik/styleclip (one of the very first models on Replicate!)
+06:00 - https://en.wikipedia.org/wiki/DeepDream (AI can generate images but they're weird)
+08:11 - https://replicate.com/stability-ai (stable diffusion came out)
+09:42 - https://replicate.com/dribnet (first paying customer!)
+11:48 - https://replicate.com/docs/reference/open-source (open source at Replicate)
+15:22 - https://replicate.com/collections/text-to-video (video generation models)
+16:05 - https://replicate.com/docs/get-started/fine-tune-with-flux (guide to fine-tuning an image model)
+16:16 - https://replicate.com/black-forest-labs (Flux image generation models)
+16:38 - https://replicate.com/zeke/ziki-flux (Flux fine-tuned to generated images of me)
+17:22 - https://replicate.com/docs/reference/http#models.create (API to create models)
+17:38 - https://www.roomgpt.io/ (AI furniture staging)
+19:12 - https://github.com/zeke/flux-fine-tune-action (Store your image model training data in GitHub and use Actions to train new versions of your Flux model)
+19:18 - https://replicate.com/blog/using-synthetic-data-to-improve-flux-finetunes (using synthetic data to improve Flux finetunes)
+21:01 - https://replicate.com/explore (explore new and popular models on Replicate)
+22:20 - https://replicate.com/daanelson/imagebind?input=nodejs (AI playground with code snippets for each prediction)
+23:59 - https://replicate.com/meta/meta-llama-3-8b-instruct (Llama models are cheaper than OpenAI and Anthropic)
+29:30 - https://simonwillison.net/2024/Oct/17/video-scraping/ (extracting structured data from videos with AI)