Quick case study of Adobe Firefly (“creative generative AI models with an initial focus on image and text effect generation”): With picture #1 below, I wanted to emphasize on the model in the center, thus clean up the distractions on the left side, and also move the laptop on the right upwards. The “Inpainting” function in Affinity Photo.app wo... Read more 08 Nov 2023 - 2 minute read
I have rigged up a quick web application for easier experimentation with the Claude-2 Foundation Model. Major features (see screenshot): (Text) File upload Formatted output (Markdown) Copy-to-clipboard saving of access credentials (in the browser’s local store) Notably... Read more 04 Nov 2023 - less than 1 minute read
Maybe Saturday Shenanigans, maybe not: “Large Language Models as General Pattern Machines”: Large language models (LLMs) are trained to absorb the myriad of patterns that are woven into the structure of language. […] A key observation of our work—and perhaps contrary to the predominant intuition—is that an LLM’s ability to represent, manipul... Read more 28 Oct 2023 - 1 minute read
Updates on Document-Processing: long context models Claude-2 100K and GPT-4-32K have landed on my desk. Huge thanks 🙏🏻 to the benefactors! My first use-case: convert a project overview I had foolishly written in Word to plain text Markdown, to grace my Bitbucket source code repository. I saved the Word document to the “Word XML document (*.... Read more 28 Oct 2023 - 2 minute read
A video is trending on social media that makes some thing of “impersonation attacks”. My take: just „clickbait“, and not a real threat: inter-frame variance too high, caused by different hallucination artifacts introduced into each individual frame. See screenshots. very low-frame rate, obviously. Scaling it up may introduce even more artif... Read more 19 Oct 2023 - 1 minute read