RT @bio_bootloader: alien spacecraft have visited earth and gifted us incredible technology and you are still doing the same things every d…

😮

Mostly unrelated: an AI doesn’t have to connect to the Internet to execute code and be dangerous. It just has to convince a human to run the dangerous code on its behalf. https://t.co/ER0S6og1mi

👀 https://t.co/osOyUWMlRY

I have a confession to make: I don’t sleep well anymore. The things I ready about AI daily, in the thousands per day, will have a profound impact on our world. The reason why I started Synthetic Work is not that AI is fashionable or lucrative. Synthetic Work is my warning call.

My emphasis:

“Authors 𝘀𝗵𝗼𝘂𝗹𝗱 disclose in their manuscript the use of AI and AI-assisted technologies” https://t.co/ZOJWqMSq2E https://t.co/5wdkFhlvyf

@shadbush @ClaireSilver12 Thank you for answering the call, Aleksandr. You are on my watch list now 🙂

Really good and easy-to-understand recap of the recent decision of the US Copyright Office about AI output. Written by an actual lawyer. Not by a random person who, all of a sudden, is an expert in bank runs and capital allocation. A silly example that would never happen, clearly https://t.co/XIB630d219

I’ve been playing with @automatorplugin for a few weeks. As I showed in one of my video podcasts for the Italian audience, it’s really powerful.

They just introduced support for the OpenAI model gpt-3.5-turbo (the same one that powers #ChatGPT). Support for #GPT4 is coming. https://t.co/IdYcBolqkC

@hardmaru @TonyZador @ylecun Isn’t it funny that I just read this thread?
https://t.co/XGLj4bE33S

@neurosock Given that now we have LLMs that can read images, it doesn’t seem impossible to create a reader app that ingests existing PDFs (as they are, no modifications), captures the content, and repurposes it for a low-vision audience in almost real-time.

It might even redesign charts.