November 2023 - AI Tidbits Monthly Roundup

Open-source LLMs outperforming GPT-4 and the first 200k tokens LLM, Stability AI’s and Pika Labs’ revolutionary text-to-video models, and prompting GPT-4 to outperform Med-PaLM 2 on medical tasks

Welcome to a subscriber-only edition 🔒 of AI Tidbits, where I curate the firehose of AI research papers and tools so you won’t have to.

This is the monthly curated round-up, so if you're pressed for time and can only catch one AI Tidbits edition, this is the one to read—featuring the absolute must-knows.

Upgrade to premium to access monthly AI round-ups and deep dives into key topics like OpenAI's DevDay and autonomous agents.

---

Welcome to the November edition of AI Tidbits, where we unravel the latest and greatest in AI. November was filled with groundbreaking announcements from industry leaders and exciting progress in open-source AI.

This month, OpenAI announced GPTs, Assistants, and GPT-4 Turbo, followed by Anthropic's new release of Claude 2.1 and Perplexity's novel language models having access to online information. On the open source front, Yi open-sourced a 200k context window model, and Berkeley unveiled Starling 7B which has achieved performance on par with GPT-4, thanks to RLHF.

In the arenas of generative image and video, there has been a surge of remarkable advancements. Pika unveiled a groundbreaking text-to-video model, pushing the boundaries of visual content creation. Stability AI introduced Stable Diffusion XL, revolutionizing the generation of high-quality images at remarkable speeds, and launched Stable Video Diffusion, marking their foray into video generation.

These and many more exciting updates across novel promoting frameworks, autonomous agents, multimodal AI, and open-source repositories are part of this month’s roundup.

Let's dive in!

---

Overview

• Industry announcements (12 entries)

• Large Language Models

  • Open-source (5 entries)

  • Research (8 entries)

  • ✨ Special feature: Prompting techniques (6 entries) - Next Sunday, a new AI Tidbits Deep Dive will be sent out, detailing research-backed advanced prompting strategies to boost the performance of LLMs

• Autonomous Agents (5 entries)

• Image and Video (9 entries)

• Audio (3 entries)

• Multimodal (5 entries)

• Open-source (13 entries)

• Cool Tools (4 entries)

Recent AI Tidbits Deep Dives

OpenAI DevDay - a pivotal moment for AI

The rise of autonomous agents

Revolutionizing document processing with multimodal GPT

Industry announcements

1. OpenAI announces a host of product releases in its inaugural dev conference: GPT-4 Turbo with a 128k-context window, a new text-to-speech model, and an autonomous agents API

2. Anthropic announces Claude 2.1 - a new version of its chatbot with a 200k context window, function calling support, and fewer hallucinations

3. Perplexity unveils two new language models, pplx-7b-online and pplx-70b-online, providing up-to-date and accurate, accessible through a first-of-its-kind public API

4. xAI announces Grok-1 - a language model outperforming GPT-3.5 and Llama 2 having real-time direct access to tweets

5. Amazon introduces Amazon Q - its new chatbot for the enterprise, allowing companies to synthesize content and answer questions based on the company's data

6. Pika Labs launches Pika 1.0 - a substantially improved model for video generation and editing

7. Inflection releases Inflection-2 - a new version of its language model

8. Humane introduces the AI Pin - a $699 OpenAI-powered wearable device for everyday use

9. Adept announces Adept Experiments - a new program for users to explore Adept's models, along with ACT-2 - Adept's new multimodal model for UI understanding and action-taking

10. Together AI releases Together Inference Engine - the world's fastest, achieving unparalleled speeds of up to 171 tokens per second

11. Figma introduces the new FigJam AI to automate mundane design tasks

12. Runway ML has updated its Gen-2 video generator to improve video quality and consistency, and to support higher-resolution video generation from images

    

    Pika 1.0

Large Language Models (LLMs)

Open-source

1. 01.ai openly releases Yi-34B - the first Chinese model to top Hugging Face's LLM Leaderboard supporting a 200k context window