The Biggest Day Since ChatGPT
PLUS: Databricks uses open-source tools to build unified customer views, addressing key challenges.
Welcome back. Today is Tuesday, November 7th and in today's news OpenAI unveiled major upgrades like GPT-4 Turbo, instruction following, AI assistants, multi-modality, text-to-speech. New to The Intelligence Age? Sign up here.
Drop us a line anytime at [email protected] with feedback.
News & Insights
OpenAI’s New Models, Tools, APIs, & More
In the biggest day of AI news since ChatGPT released, OpenAI extends its already significant lead in the AI market.
Here’s our summary of what was announced, and why it matters.
This new model can process 4-8x more text than previous generates of GPT-4, unlocking many new AI use cases such as reading full industry reports, engaging in longer conversations, summarising entire research papers, and more.
If that wasn’t enough, it’s also 2-3x cheaper.
Better instruction following
GPT-4 turbo is not just bigger and cheaper, but better. It’s been trained to follow nuanced instructions with more care, and more consistently adhere to requests for specific output formats.
What’s more, GPT-3.5 also received these updates, providing an even cheaper and faster, yet reliable solution. We recommend this for simpler use cases that wouldn’t involve complex reasoning or creativity.
Assistants are basically OpenAI’s version of an AI ‘bot’ or ‘agent’ that has access to data, tools and specific instructions. They can be customised to independently perform a sequence of actions, or hold long-term conversations, and can lean on custom documents you provide. This makes building customer support bots, teaching assistants, advisors, coaches, and more a whole lot easier.
The ChatGPT interface has been able to analyse and generate images for a while now, however these capabilities are now available via API. These means they can be integrated into our products and services, and could be used by businesses from custom t-shirt production to marketing content generation to logo design to any number of image classification tasks.
Generating realistic audio has been a long-term challenge of AI developers. While ElevenLabs have clearly led the field for the better part of 2023, OpenAI is on the block now, so expect a whole lot of products to soon be equipped with human-like generated audio.
What would be headline news on anything day, OpenAI’s Copyright Shield means they are following Google, Adobe and others to defend their customers from copyright claims.
Multiple functions calling
GPT can now interact with multiple tools at a time. For instance, you could give it access to a range of products, and depending on the specific request, it would choose which one is most appropriate to use.
Beyond the hype
Today will no doubt go down as one of the biggest product updates in history.
While enabling many businesses and developers, it’s also likely disrupted many startups who’d been working on text-to-speech, AI agents, AI models with more memory… the list goes on.
From a business perspective, today demonstrates how important it is for founders in the current tech-giant-dominated world to ensure that new products and solutions cannot be replaced by a few chatbot upgrades or new features. Strive to build solutions for specific customers with specific needs, knowing anything with general applications across multiple industries will inevitable be solved by one of our tech-giant overlords.
Staying safe in an AI world
If you weren’t yet convinced, it should be clear that we now live in an AI world, and there’s no going back.
With the world gaining increased access to state-of-the-art image and audio generation, we should increase our skepticism of digital content that’s not from trusted sources, and be wary of online privacy as we move closer and closer to lifelike digital clones that can mimic your appearance and voice.
Together with The AI Solopreneur
Grow your business or personal brand on autopilot
If you’ve ever wanted to build an audience for your business or personal brand, AI Audience Accelerator will help you—in a fraction of the time. • Create one week's worth of engaging Twitter, LinkedIn, and newsletter content in one hour • Learn how to create content that builds your authority and generates leads • Spend less time on the content treadmill and more time focusing on growing your business
Please support our sponsors!
Building A Single Customer View Using Open-Source Tools And Databricks
In the evolving landscape of customer data management, building a single customer view (SCV) is a critical undertaking. The process, however, can be fraught with complexities arising from data quality limitations, cost considerations, and the need for compliance with data handling regulations.
Databricks, an open-source, Spark-based cloud computing platform, offers a solution. It enables the construction of a data quality and record linkage pipeline. This approach addresses the challenges of data quality, cost, and compliance, and it does so without the need for expensive third-party master data management (MDM) tools.
The workflow involves implementing data profiling and validation checks using the 'Great Expectations' library, ensuring compliance with data protection and retention policies, and using the 'Splink' library for probabilistic record linkage within customer datasets. The final step involves the application of golden record logic to deduplicate and link disparate customer datasets, resulting in the coveted SCV.
However, building an SCV isn't the end of the journey - it's just the beginning. An SCV opens the door to numerous impactful data science and analytics applications. These can help a business market more efficiently, understand their customers better, and deliver superior service. towardsdatascience
Around The World
Written by Isaac R. Ward, Casey Clifton, and Alex Brogan.
Send us feedback at [email protected] and help us provide the best coverage of Artificial Intelligence possible.
Interested in reaching smart business leaders like you? To become an Intelligence Age partner, apply here.