EP51: OpenAI's Sora, Gemini Pro 1.5 10M Context, ChatGPT Memory, GraphRAG, ChatRTX, Microsoft UFO...

02/16/24 • 89 min

Show Notes: https://thisdayinai.com/bookmarks/28-ep51/
Sign up for daily This Day in AI: https://thisdayinai.com
Try Stable Cascade: https://simtheory.ai/agent/508-stable-cascade
Join SimTheory: https://simtheory.ai
======

This week we take several shots of vodka before trying to make sense of all the announcements. OpenAI attempted to trump Google's Gemini 1.5 with the announcement of Sora, 1 minute video generation that does an incredible job of keeping track of objects. Google showed us that up to 10M context windows are possible with multi-modal inputs. We discuss if a larger context window could end the need for RAG and take a first look at GraphRAG by Microsoft hoping to improve RAG with a knowledge graph. We road test Nvidia's ChatRTX on our baller graphics cards and Chris tries to delete all of his files using Microsoft UFO, a new open source project that uses GPT-4 vision to navigate and execute tasks on your Windows PC. We cover briefly V-JEPA (will try for next weeks show) and it's ability to learn through watching videos and listening, and finally discuss Stability's Stable Cascade which we've made available for "research" on SimTheory.

If you like the show please consider subscribing and leaving a comment. We appreciate your support.

======
Chapters:
00:00 - OpenAI's Sora That Creates Videos Instantly From Text
13:49 - ChatGPT Memory Released in Limited Preview
23:31 - OpenAI Rumored To Be Building Web Search, Andrej Karpathy Leaves OpenAI, Have OpenAI Slowed Down?
33:04 - Google Announces Gemini Pro 1.5. Huge Breakthrough 10M Context Window!
50:11 - Microsoft Research Publishes GraphRAG: Knowledge Graph Based RAG
1:02:03 - Nvidia's ChatRTX Road Tested
1:07:18 - AI Computers, AI PCs & Microsoft's UFO: An Agent for Window OS Interaction. Risk of AI Computers.
1:18:46 - Meta's V-JEPA: new architecture for self-supervised learning
1:24:26 - Stability AI's Stable Cascade

If you like the show please consider subscribing and leaving a comment. We appreciate your support.

Previous Episode

EP50: We Bet $1000 Using Gemini Advanced, Qwen1.5 72B, Retell AI, Apple's MGIE & GOODY-2

Subscribe to ThisDayInAI: https://thisdayinai.comTry AI Agents on SimTheory: https://simtheory.aiShow notes: https://thisdayinai.com/bookmarks/6-ep50

Tell us your thoughts on Gemini here: https://thisdayinai.com/post/62-your-thoughts-gemini-advanced/

Thanks to everyone for all your support and kind reviews to reach 50 episodes! Please consider leaving us a review wherever you get your podcasts.
=====

This week we cover the launch of Google Gemini Advanced, Gemini Ultra 1.0 and Bard being Renamed to Gemini. We compare GPT-4, Gemini Ultra 1.0 and Qwen 1.5 72B by sports betting $1000 on horse racing.

We celebrate 50 episodes and share our excited for Qwen 1.5 72B's performance at coding and quick refusals. We cover new releases including SyncLabs and Retell AI and Apple's Open Source Guiding Instruction-based Image Editing via Multimodal Large Language Models.

Finally, we discuss GOODY-2 and it's high refusal rate.

=====
CHAPTERS:

00:00 - Betting $1,000 To Compare Gemini Ultra 1.0 to GPT-4 to Qwen 1.5
07:33 - Google Gemini Advanced, Ultra: Details of Announcement and First Impressions
25:48 - OpenAI is Developing Agents to Control Your Devices
27:40 - Celebrating 50 Episodes of This Day in AI
30:34 - Qwen 1.5 72B: We're Impressed!
42:47 - SyncLabs: Tested & Impressions
47:58 - Retell AI: Tested & Impressions
54:18 - Apple's Open Source Guiding Instruction-based Image Editing via Multimodal Large Language Models
58:10 - GOODY-2: The World's Most Responsible AI Model

Next Episode

EP52: The Groq Breakthrough, Google's Gemma 7B, Unlimited Context, Can 'Magic' Reason?

Show notes: https://thisdayinai.com/bookmarks/32-ep52
Groq Mixtral: https://simtheory.ai/agent/567-groq-mixtral-edition
Groq Llama: https://simtheory.ai/agent/566-groq-the-speed-oriented-chat-companion
SimTheory: https://simtheory.ai
====
This week we discuss Groq's LPU Chips and the implications of low cost low latency LLMs on custom hardware. We revisit our prank calling to see if Groq's low latency gives an advantage and see if we can improve Air Canada's chatbot. We discuss the launch of Google's Open Source Gamma 7B release and Magic's $148M fundraise for an AI co-worker who can reason. We also cover ChatGPT losing it's mind during the week.

If you like the show, please consider subscribing. Thanks for listening.

====
Chapters:
00:00 - Groq, Groq API and Retell with Groq
32:48 - Google Gemma 7B Open Source Model
39:04 - The 'Magic' Breakthrough on Reasoning and Context
50:19 - Sounds for OpenAI Sora Thanks to ElevenLabs Sound FX
51:59 - ChatGPT Goes Haywire