🤖 Reasoning about AI risk usually turns on "agency." We have no coherent, shared understanding of what "agency" means, so most AI risk evaluations are confused and misleading. From Better without AI... twitter.com/i/web/status/1… pic.twitter.com/3EAAw43mZF
Joscha Bach defines intelligence pic.twitter.com/lAiBYGOTib
How Transformers Work This is probably one of the most beautiful visualizations of how today's LLMs work. ig.ft.com/generative-ai/ twitter.com/i/web/status/1… pic.twitter.com/fryvlCJXPH
Goldman Sachs just released their AI report on economic growth. They predict 300 million jobs will be lost to automation. Here are the key takeaways everyone must know: pic.twitter.com/5J6aoqOSB4
Once people learned the messages were co-created by a machine, it didn’t work. Simulated empathy feels weird, empty.
Bye bye black boxes ? Have you seen this stuff ??? Mind blowing 🎉🎉🎉 #ai #MachineLearning #datafestafrica2022 #DataScience pic.twitter.com/IAq0lFtDqg
- AI is lots of rules behind it all
All companies will be AI companies in the next 5 years, just like all companies are digital companies right now.
"Prompt Explorer" - a GPT-3 powered google sheet that lets you explore the "narrative neighbourhood" of any prompt: eg "abandoned futuristic colony" will output detailed MJ/SD ready prompts for: 1. what's above (drone view)? 🚁 2. what's inside? 🔦 3. what's nearby? 🧭 pic.twitter.com/GMwnIxXQBm
AI-enabled coding is heating up! The new @Tabnine_ is *modular* - aka trained from ground up for Python/JS/Java/etc, not retrofitting general text models Means you can train custom models *on your own code* for full-line completions and code snippets *IN YOUR OWN CODE STYLE*🤯 pic.twitter.com/xg5Zcp7wdh