AGI might be quietly taking shape in Minecraft, of all things. JARVIS-1 navigates Minecraft's challenges with multimodal input, utilizing pre-trained models for interpreting visual cues and instructions, all while enhancing its capabilities through a growing multimodal memory.
@aibreakfast
-
Desktop Whisper API tool enables efficient long-form speech-to-text
By
–
Someone made an easy Whsiper API for desktop (Mac) that is the best speech to text generator I’ve ever used. Doesn’t stop recording during pauses, you can just rant at your desktop for an hour and get the full text from it, use ChatGPT on it from there. It’s called
-
US and China to Pledge Ban on AI in Autonomous Weapons
By
–
Joe Biden and Chinese President Xi Jinping are expected to pledge a ban on the use of Al in autonomous weapons today.
— AI Breakfast (@AiBreakfast) 11 novembre 2023
Probably for the best. pic.twitter.com/GJvnDpZ967Joe Biden and Chinese President Xi Jinping are expected to pledge a ban on the use of Al in autonomous weapons today. Probably for the best.
-
Next iteration of ChatGPT announced
By
–
4.5, 5, call it what you want. The next iteration of ChatGPT. Don’t get hung up on the wrong details.
-
ChatGPT Vision recreates a dropdown menu in Tailwind CSS
By
–
ChatGPT + Vision API perfectly recreates dropdown menu in Tailwind CSS from a screenshot:
— AI Breakfast (@AiBreakfast) 10 novembre 2023
pic.twitter.com/gU6IKA4iyYChatGPT + Vision API perfectly recreates a dropdown menu in Tailwind CSS from a screenshot:
-
GPT-5 ‘Gobi’: video and multimodality in 2024
By
–
GPT-5, codenamed 'Gobi,' rumored for early 2024 release, will handle video interpretation as part of its multimodality. Altman says 'What we launched [at Dev Day] is going to look very quaint relative to what we're busy creating for you now.' '[GPT-5] will work for most things.
-
Rumor about a new Amazon LLM named Olympus
By
–
Rumor that Amazon is developing a new LLM, codenamed Olympus. They have one giant data set that others don’t: What you buy.
-
Birth of Agents with Custom GPTs
By
–
Plugins being replaced with “GPTs” – customizable agents trained on expanded knowledge, instructions, and actions. A natural language version of Zapier with the adaptability of a LLM. The birth of Agents.
-
GPT-4 bientôt avec fenêtre de contexte de 128k tokens
By
–
Wow – 128k token context window for GPT-4 about to be announced. Your prompt can be the length of an entire book.
