- Innovation Profs Newsletter
- Posts
- Innovation Profs - 11/26/2024
Innovation Profs - 11/26/2024
Your weekly guide to generative AI tools and news
HOLID-AI is back
Join us for HOLID-AI 2, a free event on Dec. 6, to learn about recent updates in the world of generative AI. We will discuss ChatGPT with Canvas, computer use with Claude, the potential plateau of LLM abilities, and the latest movement in the chatbot leaderboard.
On the multimedia side, we will discuss updates to the Midjourney editor, new Ideogram features and Meta's new AI video tool. Did we mention the event is free? What are you waiting for? Sign up today.
Generative AI News
Amazon doubles down on AI startup Anthropic with $4bn investment
Amazon’s investment in Anthropic, the company behind the LLM Claude, has reached a total of $8 billion, as Amazon is literally doubling down on their previous $4 billion investment in Anthropic. As part of the deal, Anthropic will use Amazon Web Services for their cloud services and will work with Amazon on the development of their Trainium chips, which will be used to train future Claude models.
New York Times Says OpenAI Erased Potential Lawsuit Evidence
A new development in the ongoing legal battle between the New York Times and OpenAI over OpenAI’s potential copyright violation in using NYT articles to train their LLMs: OpenAI engineers inadvertently erased data that the NYT was using as evidence in their lawsuit, as a result of what OpenAI’s legal counsel referred to as a “glitch.” However, it sounds like much of that data has been recovered, but the NYT alleges that the data is missing the original file names and folder structure, which makes the data largely unusable for legal purposes. When I (Porter) asked ChatGPT to comment this development, it responded, “It feels like a stark reminder of how fragile trust and accountability are in the intersection of groundbreaking technology and the legal system.”
Anthropic proposes a new way to connect data to AI chatbots
This week Anthropic introduced the Model Context Protocol (MCP), a standard that allows models to “draw data from sources like business tools and software to complete tasks, as well as from content repositories and app development environments.” MCP is a single protocol that provides two-way connections between data sources and chatbots as an alternative to building custom connections for each data source. According to Anthropic: “As the ecosystem matures, AI systems will maintain context as they move between different tools and data sets, replacing today’s fragmented integrations with a more sustainable architecture.”
Microsoft’s 10 new AI agents strengthen its enterprise automation lead
Microsoft announced the launch of ten new AI agents at Ignite 2024 that are now available for enterprise use, with applications in areas including financial services, supply chain management, and manufacturing. Among these agents are Financial Meeting Prep, Financial Insights Agent, Factory Operations Agent, Factory Safety Agent, Retail Industry Data Model, and Personalized Shopping Agent. This move puts Microsoft at the forefront of Agentic AI.
Quick Hits
Tool of the week: Adobe Enhance Speech
Adobe last week released version 2 of it’s audio editing tool Enhance Speech. Enhance Speech uses artificial intelligence to turn low-quality audio (badly muffled, background noise, etc.) and convert it to studio-quality recording. Try it out on audio files as part of Adobe’s podcast tools. Premium users can use the tool with video files.
AI-generated image of the week
We hopped on the trend where you ask ChatGPT to make an image based on everything it knows about you. Here are the results (Porter on the left, Snider on the right).
Prompt: make a picture based on everything you know about me
Get starting with Generative AI
New to generative AI? Here are some places to start…
What we found
Luma AI Dream Machine platform got an update this week. Dream Machine can create images and videos. The update if faster and better at text. It can generate consistent characters and maintain them across images and videos.
Try it out for free here.