The Voicebot Podcast

Google dominated the generative AI news earlier in 2024. The unexpected introduction of Gemini 1.5 is covered in depth, and guest host Allen Firstenberg discusses his first-hand experience testing the model. We also cover the Gemma open-source models and Google’s latest PR misstep related to its image generation. Also on tap are discussions around Mistral, NVIDIA, Adobe, Pindrop, and a few more.

• Google Goes Big on Context with Gemini 1.5 and Dips Into Open-Source with Gemma
• Google Unveils Gemini 1.5 Pro LLM With Staggering 1M-Token Context Window
• Google Unveils Open-Sources Gemma Generative AI Models as Lightweight Gemini Alternative
• Trust is the Hidden Force Shaping the Generative AI Market and Google Keeps Squandering It

• Mistral Introduces Two New High-Performance LLMs, and Neither Are Open-Source
• Pindrop Launches Real-Time Audio Deepfake Detection Tool Pindrop Pulse
• Adobe Brings Generative AI Assistant to Help Extract Information from 3 Trillion PDFs
• Microsoft Introduces a New Red Teaming Toolkit for Generative AI Security and Safety Testing
• Stability AI Upgrades Synthetic Media Engine to Stable Diffusion 3

• NVIDIA is Officially the Giant of Generative AI
• Abridge Raises $150M for Automating Medical Documentation and Plans to Create Generative AI Foundation Models
• Generative AI Image Editing Startup Photoroom Raises $43M

• Elon Musk sues OpenAI and CEO Sam Altman over contract breach
• The Surprising Tactics of Deepfakes Targeting Contact Centers Today
• Klarna Says the Quiet Part Out Loud - Generative AI Can Replace a Lot of Contact Center Staff

The Generative AI News (GAIN) rundown for April 27, 2023, is here. Another week of breaking news has piled up, and we have a breakdown of the top stories and what they mean for the industry. The developments include news from ChatGPT, HuggingFace, Google, Nvidia, Sensory, Hour One, D-ID, deepfake musicians, and more.

Your hosts today are Bret Kinsella and Voicebot.ai's Eric Schwartz. The top stories in generative AI land this week include:

• Plugging in a new vision: Greg Brockman from OpenAI demonstrated some new ChatGPT plugin features; several are jaw-dropping.
• The “super app” virtual assistant we were promised: Brockman’s demo and the discussion about the product philosophy offer an insight into where ChatGPT is headed. Move over, Alexa. Get out of the way, Siri. ChatGPT may be the virtual assistant we have always wanted.
• ChatGPT is anything but incognito: While everything ChatGPT seems to play out in the public eye, OpenAI recognized that not every user wanted every one of their chat conversations saved in perpetuity and used for future model training. Incognito (i.e., private chatting) is now available, and a “business mode” is coming soon.

• Open source competition for ChatGPT: Hugging Face stepped up and provided a ChatGPT alternative built on open source models and data. It’s a smaller AI model than ChatGPT and is pretty good.

• Drake, The Weeknd, Bad Bunny, and Rihanna go viral: Viral hits from big stars are common. Deepfake viral hits mimicking the voice, style, and likeness of big stars may also become common. ghostwrider777 strikes again!
• Joe Rogan comments run deep: New deepfakes mimicking Joe Rogan’s podcast have the comedian and commentator talking about a “ slippery” slope.
• Grimes jumps on board: The musical artist says she will split royalties 50/50 with anyone deepfaking her voice. She has no label and no binding legal constraints giving her more flexibility than most musicians.

• Prompt-to-video: Hour One introduced a new text-to-video solution that enables full video generation for presentations from a single prompt.
• Canva gets digital people: D-ID introduced a new Canva app that enables you to add generative videos to any project.
• Chatbots are suddenly popular: Character AI landed $150M in funding at an obscene valuation. Virtual Elon Musk, Mark Zuckerberg, and 2.7 million other chatting avatars with personalities have driven 100M user visits in just two months.

• Bard learns to code: Google is slowly catching up with the generative AI leaders. It’s ChatGPT competitor—or, is it a Bing Chat competitor— can now code. This is not a true competitor to GitHub Copilot yet.
• Sec-PaLM gets into security: Google also rolled out a new cybersecurity solution with the parsimonious name of Google Cloud Security AI Workbench. It is based on a fine-tuned version of the PaLM large language model (LLM).

• ChatGPT gets an edge: Sensory rolled out a new hybrid on-device and cloud solution that can enable the use of ChatGPT and similar services on devices.
• Nvidia on rails: NeMo, Nvidia’s LLM, now has a new feature...

The Generative AI News (GAIN) rundown for April 20, 2023, was recorded live at the Model Mania conference, which focused on enterprise generative AI solutions.

News this week has more on Elon Musk and some surprising news from Stability AI. We also talk about a deepfake of Drake and The Weeknd that went viral, Adobe Firefly, Atlassian, ChatGPT in government legal actions, Universal Music lawsuits, and more.

Bret Kinsella hosted this week with his Voicebot.ai colleague Eric Schwartz. The top stories in generative AI land this week include:

• Big Data LLM: Stability AI introduced a new large language model trained on 1.5 trillion data tokens. It’s open-source and comes in a variety of model parameter sizes.
• Stable Diffusion for the Enterprise: The new XL model from Stability AI offers better photorealism, more coherent text, and is positioned for enterprise use. Oh, and the company’s valuation may have risen from $1B to $4B in less than six months.

• Generative AI for designers and video makers: Adobe Firefly will make it easier for designers to incorporate generative AI into their workflow. The new services for video production will take that to a new level in Premiere and After Effects.

• In-Context Search and Answers: The creator of Jira, Confluence, and Trello has added generative AI features for summarization, text generation, and question-answering from your productivity software data.

• What is Elon up to now: Musk created a new company in Nevada last month called X.ai. He says he wants to create a third option beyond OpenAI and Google offerings.

• Name recognition on another level: The U.S. Justice Department’s suit against Google for alleged search monopolization said ChatGPT might have come sooner if not for the company’s stranglehold on the market.

• Viral Music duo: 10M TikTok views and 600k Spotify streams later, a popular deepfake of a The Weeknd and Drake called “Heart on My Sleeve” was taken down due to a request from one of the music labels.

The Generative AI News (GAIN) rundown for March 23, 2023, was packed with significant announcements. Bret Kinsella hosts this week along with Eric Schwartz. The top stories in generative AI land this week include:

• Picasso and NeMo: Nvidia isn’t going just to cash checks for GPU sales related to the generative AI tsunami. They now offer text-to-image and text-to-text models that compete directly with OpenAI, Google, and Stable Diffusion.
• Omniverse Upgrade: Omniverse is an open platform for 3D design collaboration and real-time physically accurate simulations. It announced new services for creating virtual factories mirroring real-world facilities, the option to stream simulated experiences and train to AI-powered robots, and a simulator for testing autonomous vehicles.

• Show your PaLMs: Google is now letting testers use its giant large language model (LLM) PaLM. This is different from LaMDA and might eventually be a replacement for the brains behind Bard. Right now, it is Google’s answer to GPT-3/GPT-4. A tool to help developers using PaLM called MakerSuite was also announced.
• Gmail and Docs to Get AI writing assistant: Docs and Gmail are getting PaLM-enabled text generation features.
• Bard announced again: Google says it is now offering access to Bard to the general public. But there is a waitlist.

• Midjourney 5 is here: There are several minor upgrades, but the key benefits are enhanced quality, more coherence, better photorealism, and more detail.
• Bing adds DALL-E: You can now create images in Bing through a new DALL-E integration. The quality seems better to me than DALL-E 2, which you can access today on OpenAI. Might this be the long-awaited DALL-E 3?

• TikTok to tighten deepfake rules: TikTok announced new policies around synthetic media and deepfake use on the platform before its CEO’s Congressional testimony. The policy description grew from 30 words to nearly 400. Will others social platforms take this as a cue to make their own updates?

• Roblox wants to make development easier: Roblox added generative AI tools that enable developers to use natural language to create objects and generate code.
• Unity wants generative AI NPCs: Unity didn’t make any concrete announcements. However, its CEO told Reuters that generative AI would help game makers write dialogue and enable non-player characters to interact more naturally with human players.

• LinkedIn goes generative: The Microsoft-owned company added new generative AI features to create user profiles and job descriptions.
• SoundHound shows Chat AI assistant: The new mobile app offers an assistant that blends SoundHound’s NLU-based assistant with new LLM features.

The show is recorded live and streamed via YouTube and LinkedIn at 12 noon EST on Thursdays. You can re-watch each week’s discussion on Voicebot’s YouTube channel. Please join us live next week on YouTube or LinkedIn. Also, participate in the live show by commenting.

OpenAI introduced GPT-4o as a new model and the foundation for ChatGPT. The company also offered more than a dozen videos and other use case examples, which enabled us to break down many of the nuances enabled by the new model. Is this the voice assistant everyone always wanted?

A day later, Google debuted its latest updates for Gemini and offered a preview of Project Astra, the upgraded future of Gemini assistant, which they say will turn it into an agent for users. Google provided just a couple of videos and examples, but it also showed off a much broader set of impacts across its applications. These include search, Workspace, photos, and smart glasses (?). In all, Google had more than 100 announcement at I/O and most related to AI.

Below are 20 articles and blog posts that provide added context on the biggest news across the generative AI landscape. There were a few stories we did not get to that were also of significant impact and we will likely work them into the next episode of GAIN.

• ChatGPT 4o is the Voice Assistant We Always Wanted, with Two Exceptions
• 6 Things Announced at Google I/O, One That Mattered A Lot, and One Mystery

• Accolades for OpenAI's GPT-4o Are Rising
• OpenAI Shows Off New GPT-4o Generative AI Model and More ChatGPT Upgrades
• Hello GPT 4o
• Introducing GPT-4o and more tools to ChatGPT free users
• OpenAI Chief Scientist Ilya Sutskever Leaves Company
• ChatGPT 4o is the Voice Assistant We Always Wanted, with Two Exceptions

• Google’s ‘Gemini Era:’ The Top Generative AI Google I/O 2024 Announcements
• Google Unveils Gemini ‘Gem’ Custom Generative AI Chatbots to Rival OpenAI’s GPT Store
• Introducing VideoFX, plus new features for ImageFX and MusicFX
• Gemini breaks new ground with a faster model, longer context, AI agents and more
• Generative AI in Search: Let Google do the searching for you
• 3 new ways to stay productive with Gemini for Google Workspace
• 6 Things Announced at Google I/O, One That Mattered A Lot, and One Mystery
• Announcing Trillium, the sixth generation of Google Cloud TPU
• Ask Photos: A new way to search your photos with Gemini
• 100 things we announced at I/O 2024

• Anthropic Taps Instagram Co-Founder for Chief Product Officer to Expand Enterprise Generative AI Services

• AWS Launches Amazon Bedrock Studio for to Support No-Code Generative AI Projects

• YouTube’s AI ‘Jump Ahead’ Tool Helps Users Skip to the Best Parts of Videos