Deploy AI 15x cheaper with MosaicML
Plus: US, UK govs step in to regulate AI. Bing Chat released worldwide.
Hello, Engineering Leaders and AI enthusiasts,
Welcome to the 12th edition of The AI Edge newsletter. Today, we bring you 15 times cheaper AI deployment. Thank you everyone who is reading this.
In today’s edition:
🧩 Deploy AI 15x cheaper with MosaicML
⚖️ US, UK govs to regulate AI developments
🌐 Bing AI released worldwide
🤝 Databricks acquires Okera
🌀 Avoiding hallucinations in LLM-powered applications
Let’s go
Deploy AI 15x cheaper with MosaicML
MosaicML, a provider of generative AI infrastructure, has launched a fully-managed inference service to help enterprises easily and affordably deploy generative AI models. They claim that it's possible to serve LLMs for up to 15 times less than other comparable services in the market. A publicly traded customer has already adopted the service in financial compliance & the total cost of ownership for their first model was less than $100,000.
Why does this matter?
Startups will adopt more generative AI capabilities in their products due to the rise of LLMs. However, high costs and privacy challenges when building and deploying such models have made it difficult for them. But with MosaicML inference, they can broaden more growth possibilities.
US and UK governments step in to regulate AI developments
The CEOs of Google, Microsoft, and OpenAI have been summoned to the White House for an AI safety meeting. The meeting will address concerns regarding the rapid growth of AI technology and its potential to cause privacy violations, bias, and the spread of scams and misinformation.
Also the UK's competition regulator has also announced an investigation into the impact of AI on consumers, businesses, and the economy, including the need for new controls on technologies like OpenAI's ChatGPT.
Why does this matter?
The heightened interest of authorities in AI developments could be a positive sign. It might lead to regulations and compliances that provide everyone a fair chance to grow in this evolving landscape. Additionally, it might also help suppress the illicit use of AI for scams, deepfakes, and other anti-social activities.
Bing AI released worldwide equipped with visual search, copilot, and other new features
In an exciting move, Microsoft opens up its AI-powered Bing for all users without a waitlist. Powered by ChatGPT, the company debuted a limited preview version only three months ago. Now, anyone can access it by signing into the search engine via Microsoft’s Edge browser.
Microsoft also revamped the search engine with new features, including the ability to ask questions with pictures, access chat history so the chatbot remembers its rapport with users, export responses to Microsoft Word, and personalize the tone and style of the chatbot's responses.
Why does this matter?
While the move highlights Microsoft’s confidence in the tool and readiness for wider use and feedback, it may prompt other tech giants to make newer, richer AI-powered experiences more accessible to users.
Databricks acquires Okera, an AI-centric data governance platform
Databricks has acquired Okera, a data governance platform that uses AI to automatically discover and classify personally identifiable information, tag it, and apply rules to it (with a focus on the metadata) using a no-code interface.
The company plans to integrate Okera’s technology into its existing data and AI asset governance solution, Unity Catalog. Moreover, it gives the benefit of Okera’s isolation technology, still in private preview, that enforces governance control on arbitrary workloads without significant overheads. The acquisition will also help Databricks offer more APIs to its data governance partners, who can then provide their customers with better solutions.
Why does this matter?
Because the overall AI landscape is changing so quickly and the number of assets is growing too fast, especially with the explosion of LLMs, standard access controls can’t capture these changes quickly enough. A modern, AI-centric governance solution will help enterprises better manage their data and AI assets at scale and across multiple clouds.
Knowledge Nugget: Avoiding hallucinations in LLM-powered applications
Large language models (LLMs) like GPT-4, LLama, and BARD can provide incredible value as personal assistants or chatbots. However, hallucinations, or instances where the generated content is nonsensical or unfaithful to the source, remain a significant concern. LLMs hallucinate for various reasons, including limited training datasets, inability to check factual accuracy, and a tendency to generate probable but not necessarily true responses.
The article discusses how Vectara's Grounded Generation (GG) takes a step further in addressing hallucinations by augmenting the LLM's knowledge with additional data sources. GG, a form of retrieval augmented generation, allows the LLM to provide more accurate answers to user queries based not only on its training dataset but also on up-to-date information from external sources.
Why does this matter?
By understanding the causes of hallucinations and employing solutions like Vectara's Grounded Generation, developers can mitigate the risks associated with incorrect responses and enhance user trust in AI-generated content. This, in turn, paves the way for the safe and accelerated deployment of LLM technology in various industries.
What Else Is Happening
📈 Portfolio Pilot: The new ChatGPT plugin can give you investment advice by looking at your portfolio. (Link)
🤖 Meet Crypto.com’s new AI assistant Amy. (Link)
✈️ Etihad Airways’ AI chatbot will now elevate your flying experience. (Link)
🚚 Pando secures $30M to boost AI-powered supply chains. (Link)
🎶 Have ChatGPT build your music playlist with Plexamp. (Link)
💻 Get comprehensive code explanations from GPTutor. (Link)
👔 LinkedIn brought in AI to assist job seekers. (Link)
Trending Tools
Codemate: Automates debugging and code review process with a click, making developers 10x more productive.
AI Subject Line Generator: Crafts engaging email subject lines and preview text to boost open rates and save time.
Buzzli: Automates LinkedIn content creation through AI. Say goodbye to creator’s block and hello to productivity!
Muzaic Studio: Crafts ideal music for your video. Get a unique soundtrack adapted to your vision, ready in one minute.
Clone my voice: Imitate anyone’s voice with ease for voiceovers, dubbing, and impressions. Have AI read for you.
Folk: The CRM that works for you. Lightweight, customizable and powered with AI to build stronger relationships easily.
ByrdHouse: Multilingual video conferencing platform with AI-powered real-time translation across 70+ languages and meeting notes.
Keyword Camera: Automates photo labeling for quick title, description and keyword generation. Ideal for stock photographers.
That's all for now!
If you are new to ‘The AI Edge’ newsletter. Subscribe to receive the ‘Ultimate AI tools and ChatGPT Prompt guide’ specifically designed for Engineering Leaders and AI enthusiasts.
Thanks for reading, and see you tomorrow.