Outsmart The Future
Today in Everyday AI
7 minute read
🎙 Daily Podcast Episode: Adding GenAI into an enterprise setting may not be easy. So how do industry leaders like IBM do it? We asked one of their leaders and are giving you their blueprint. Give it a listen.
🕵️♂️ Fresh Finds: A video generator GPT for ChatGPT, Microsoft’s AI playbook for sustainability, and having an AI-powered doctor’s office. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: Microsoft Copilot’s new customized copilots, YouTube unveils voice cloning, and Google adding watermarks to AI audio. For that and more, read on for Byte Sized News.
🚀 AI In 5: Bing Chat and almost all LLMs have a problem with lying. BUT we found a way to stop Bing Chat from lying to you! See it here.
🧠 Learn & Leveraging AI: Are you an enterprise looking to add AI? Are you a small business looking to add AI? See how IBM does it. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked to an NVIDIA leader about how NVIDIA uses AI to improve lives, Microsoft announces competitor to NVIDIA chips, Argentina’s AI-driven election, and creating custom podcasts in ChatGPT. Check it here!
IBM Leader Talks Infusing GenAI in Enterprise Workflows for Big Wins 🏢
If you work at an enterprise company, using GenAI might not be as easy as you think.
From different departments using different LLMs to people who may have specific needs, what's the best way to integrate AI in enterprises?
We asked an industry leader, IBM, who helps shape GenAI enterprise tools.
Ben Mandelstein, Worldwide Sales Leader at IBM watsonx Orchestrate, joins us to discuss how to infuse GenAI into enterprise workflows.
Also on the pod today:
• How the IBM watsonx platform works 🧑💻️
• How IBM helps businesses 🤝
• Using IBM watsonx Orchestrate 🧠
It’ll be worth your 37 minutes:
Subscribe and listen on your favorite podcast platform
Upcoming Everyday AI Livestreams
Friday, November 16th at 7:30 am CST ⬇️
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – Meshy-1 lets you create stunning 3D models with AI, VideoGPT generates amazing videos directly in ChatGPT, and Polywork can convert your LinkedIn profile into a website.
Read This – There’s many “talks” parents need to have with kids as they grow up. Never would we have thought that one of those included a talk on AI. This article thinks so.
1. Microsoft Copilot Studio: Customized Copilot 🚀
Microsoft has unveiled Copilot Studio at Microsoft Ignite 2023, a groundbreaking low-code tool for customizing Microsoft Copilot in Microsoft 365 and creating standalone copilots. This innovative platform combines custom GPTs, generative AI plugins, and manual topics, enabling users to tailor Copilot to their specific enterprise scenarios. Copilot Studio, integrated with various Microsoft technologies, allows for the building, deployment, and management of customizations and standalone copilots within a unified web experience.
2. YouTube's Cloning Voices and Creating Music 🎵
YouTube is experimenting with innovative generative AI features, allowing users to create music tracks from text prompts or simple hums. The highlight is 'Dream Track', which auto-generates 30-second music tracks mimicking the style of nine collaborating famous artists. Additionally, YouTube is developing tools to generate music from a hummed tune. These advancements are part of YouTube's strategy to balance AI-generated music with copyright norms and its relationships with major music labels.
3. Google DeepMind's Watermarking AI-Generated Audio 🗣
Google DeepMind's AI model, Lyria, is introducing SynthID, an inaudible watermark for AI-generated audio tracks. This feature, used in YouTube's new audio generation tools, ensures the identification of AI origins in music without affecting the listening experience. SynthID remains detectable even after audio modifications like compression or adding noise.
4. Bing Chat Transforms into Copilot 🧑✈️
Microsoft has rebranded Bing Chat to 'Copilot in Bing' at Microsoft Ignite 2023, signaling a strategic shift in its AI offerings. This change aims to create a unified Copilot experience across consumer and commercial platforms. Starting December 1, Copilot in Bing will offer commercial data protection for users signing in with corporate accounts, ensuring data privacy and security.
5. Spotify Enhances Personalization with Google's AI 🎧
Spotify is expanding its partnership with Google Cloud to leverage large language models (LLMs) for enhancing user experience. This collaboration aims to analyze users' listening habits across podcasts and audiobooks, enabling Spotify to offer more tailored recommendations. By integrating Google's AI technology, Spotify is set to refine its recommendation system, providing a more personalized and intuitive listening experience for its users.
Secret to keep Bing Chat From Lying! 🤫
Bing Chat, soon to be Copilot, and almost all LLMs have a problem with lying.
BUT we found a lil work around to stop Bing Chat (which will soon be called Copilot) from lying to you!
We’re showing you how this secret hack works!
🤷♂️ What’s Going On and Why It Matters:
Integrating Generative AI in an enterprise setting can be a mess.
Emily needs OpenAI tools in marketing.
Keenen uses HuggingFace models for R&D.
Vernon taps into Meta’s Llama 2 for customer service.
And Jordan somehow has to pull it together and make sure to tear down silos. (Like MJ’s dunk tearing down Ewing in ‘91).
Enter: watsonx Orchestrate — the powerhouse GenAI platform from IBM that can help pull it all together.
Ben is a long-time Everyday AI listener and first time guest, and he’s the Worldwide Sales Leader for IBM watsonX Orchestrate.
(Yeah, we’ve got some heavy hitters hangin in the LinkedIn livestream comments, come join them!)
So how can Enterprise companies untangle this mess? After all, different departments need different models.
And as companies struggle to acclimate to Generative AI and tackle governance, sometimes employees and departments just start building their own solutions.
So how can watsonx bring it all together, and how can that lead to some big wins?
How watsonx is designed to be user-friendly for nontechnical users, particularly small businesses and startups
AB testing and gradual incorporation of generative AI into workflows
How companies should determine areas where Watson X Orchestra can have the largest impact
How watsonx Orchestrate combines generative AI and automation for efficient work processes
How demonstrating value through initial wins increases organizational readiness and excitement
Why IBM is committing to upskilling 2,000,000 people worldwide on AI
Let’s break it all down. 👇
🦾How You Can Leverage:
Did you have your notepad ready for today’s convo with Ben from IBM?
Here’s the reality: all of those gems won’t be worth a dime if you don’t implement what you learned into your day-to-day.
We get Enterprise GenAI aint easy. And watsonx might not be the solution for everyone.
But, there’s still plenty to unpack and apply to get your company up and running before the race is over.
1 – Break down the tech barrier 🚧
Don't you just hate it when you need a PhD in computer science just to understand how to use those large language models?
Large Langauge Models usually require a bit of tech know-how. And if that’s not your strong suit, you (or your department) might not be able to squeeze much juice outta the GenAI fruit.
Ben shared how watsonx Orchestrate makes the process easier with prebuilt skills, custom skills, and integrations to the software you already use, like Salesforce, Workday, Slack, Google Workspace, Microsoft products and more.
Try this: Orchestrate is just one of the products in the watsonx suite. Go check out if/how watsonx might be able to help your company break down the tech barrier.
2 – Bring your own systems ⚙️
Emily needing OpenAI, Keenen needing HuggingFace, Vernon swears by his Llama2…. Happens all the time.
You betchyour sweet chatbot it is.
Ben talked about how watsonx can solve that problem. Different teams need different tools. He said, as an example, 95% accuracy in a model might be good enough for Department B, but might not slice the cheese for Department C.
Try this: Yeah, that first integration of Generative AI across an Enterprise company can be a headache. Need to relate? Check this article on the 9 headaches when integrating Generative AI (and how to begin to solve them!)
3 – Measure a low-hanging win 🍊
Okay, picture this because this is gold.
One of the most valuable things you can do when implementing GenAI across your enterprise?
Ben said start with a low-hanging win. Picture it as picking the juiciest, ripest fruit from the tree of success. In today’s show, he even detailed the process IBM went through in its own implementation.