Welcome to The Daily AI Show Newsletter, your deeper dive into AI that goes beyond the latest news. In this issue:

• Find out how Perplexity can help you move away from custom GPTs

• How AI is helping process vast amounts of data

• What Canvas really brings to your ChatGPT experience

• New ways to use Google NotebookLM in business

Plus, we discuss the challenges of change management and why AI isn’t the job killer some suggest, OpenAI’s new Playground prompt generating feature, building automations for repetitive business problems, Tesla’s stock hit, and all the recent news we found interesting this week.

It’s Sunday morning.

Time to talk about AI, because discussing the weather is just too mainstream.

The DAS Crew

It’s Not About Replacing Doctors . . .Yet

AI is making impressive strides in the medical field, especially in diagnostic accuracy. AI-powered systems in hospitals like the Mayo Clinic can sift through years of medical history in seconds, providing insights that could be missed by human doctors under time constraints. These systems offer not only faster diagnoses but also more precise treatment options, drawing on real-time data from clinical trials and personalized medical records.

While AI’s role in diagnosing diseases is growing, it’s not about replacing doctors. Instead, AI serves as a powerful tool that reduces human error, enhances decision-making, and handles the overwhelming amount of data modern healthcare requires. From helping doctors stay current with treatment options to providing a second opinion, AI is transforming the way we approach patient care.

WHY IT MATTERS

• Improved Diagnostic Accuracy: AI processes data faster than any human, reducing diagnosis times and potentially catching details that would otherwise be overlooked.

• Support for Doctors: AI assists healthcare professionals by reducing the cognitive load and helping them make better decisions based on comprehensive, data-driven insights.

• Personalized Treatments: With real-time data access, AI enables more personalized treatment plans, improving patient outcomes.

ChatGPT Canvas? Claude Knock-Off or Novel Addition?

OpenAI’s ChatGPT Canvas introduces a collaborative, visual workspace that shifts away from the traditional chat interface. With Canvas, users can edit documents, write code, and brainstorm ideas in real-time, all within ChatGPT.

This interactive feature eliminates much of the back-and-forth that typically happens in AI-powered workflows, allowing for hands-on refinement of content.

While Canvas provides clear benefits for developers—such as live code adjustments, instant feedback, and real-time debugging—it’s not entirely a groundbreaking idea. Claude’s Artifacts already offers similar document interaction capabilities, allowing the user to work with multiple formats and collaborate effectively with AI.

Canvas, though useful, feels like a step forward within an existing trend rather than a novel leap.

This means that while ChatGPT Canvas is helpful for a wide range of tasks, it’s not necessarily the game-changer it’s positioned to be.

For users already familiar with other AI tools like Claude’s Artifacts, the real innovation comes down to personal preference and slight differences in functionality.

WHY IT MATTERS

• Streamlined Workflow: Canvas allows for real-time interactive collaboration with your AI writing or coding assistant, working together in the same document or block.. However, its core capabilities are similar to other tools like Claude’s Artifacts.

• Real-Time Coding and Feedback: Developers can troubleshoot and adjust code instantly, although these functionalities aren’t exclusive to Canvas.

• Existing Alternatives: While Canvas enhances productivity, it's not a completely new concept, and users might find similar value in other AI-powered platforms like Claude’s Artifacts.

Can Perplexity Dethrone Custom GPT’s Best Feature?

For the past year, custom GPTs have dominated as a fast and easy way to pull in live internet research, combine it with internal documents (RAG data), and leverage large language models (LLMs) for automated tasks.

Whether it's prospecting, lead generation, or preparing for sales calls, the ability to quickly deploy custom GPTs made them the go-to tool for companies looking for efficiency. However, they lacked deeper customization, especially when it came to controlling the flow of automation.

Enter Perplexity’s API.

While it’s not a direct replacement for custom GPTs, it opens new doors for those using tools like Make or Zapier to build complex automation workflows that can retrieve up-to-date, targeted information from websites relevant to the task objective. By integrating Perplexity’s API into these tools, companies can now pull in up-to-date internet research while maintaining the flexibility to customize each step of their automation process.

The result?

A solution that balances the need for real-time information with greater control over the automation sequence.

WHY IT MATTERS

• Seamless Integration: Perplexity’s API can be embedded into existing automation platforms like Make and Zapier, allowing for real-time research without sacrificing control.

• Customization Flexibility: Unlike custom GPTs, which are limited in their ability to adjust creativity and response paths, Perplexity’s API let’s you flexibly invoke web search at points in your automation workflows, while Zap or Make let you tailor each step and include web search like Custom GPTs which don’t have the workflow or instruction-following capabilities

• Improved Workflow Control: Perplexity allows for more precise control over the automation process, enabling companies to fine-tune their tasks based on current needs without relying solely on a fixed model.

• Filling the Gaps: While custom GPTs are still fast and efficient, Perplexity’s API gives companies an alternative that focuses on real-time updates.

Google NotebookLM Unlocks New Business Potential

Google NotebookLM has been around for a while, but it’s recently gained attention thanks to its new, dynamic features like podcast-style overviews and seamless integration with YouTube and audio sources. While initially seen as a tool for individual study, NotebookLM is now proving to be a powerful resource for businesses, offering innovative ways to interact with and organize complex information.

What sets NotebookLM apart is its ability to function as a knowledge repository where you can upload documents, presentations, and even YouTube links to create study guides, summaries, or audio deep dives.

This can drastically improve internal knowledge sharing, helping employees get up to speed faster on policies, project updates, or product knowledge.

One of the most exciting use cases for NotebookLM in a business setting is how it simplifies the way companies explain complex topics. You can quickly transform your technical documents or detailed reports into easily digestible audio summaries or interactive guides, making it an ideal tool for sales enablement, training, or even customer support.

WHY IT MATTERS

• Improved Knowledge Sharing: NotebookLM allows companies to streamline how they distribute important information, making it easier for teams to get aligned.

• Engaging Content Formats: The ability to create podcast-like discussions from your documents means you can deliver content in ways that people will actually consume—on the go and in a more engaging format.

• Customization Flexibility: Use NotebookLM to create internal FAQ systems, onboarding guides, or even study resources for certifications or compliance training, tailored to your business needs.

• Enhanced Internal Communication: For businesses dealing with complex or technical topics, this tool can simplify and clarify information, making it more accessible to everyone in the organization.

• Cost-Effective Solution: Since NotebookLM is free, businesses of all sizes can start using it without worrying about additional costs, making it a no-brainer for teams needing better documentation processes.

Just Jokes

“AI Is Not The Existential Threat People Think It Is”

Eran shared this recent LinkedIn post written by a super-successful business person he knows in Australia.

Leela makes a good point about how hard change and technology adoption really is for companies.  

OpenAI Playground Takes Some Toys from Claude’s Sandbox

Karl mentioned OpenAI Playground now has a new Generate button that helps users create prompts and schemas from just a few words.

You can find the feature in the Chat section of the Playground and clicking on the two star icon in the upper right.  

Automation for Podcasts

Beth led a conversation with Jyunmi and Eran about further automating the process of using transcriptions from our live shows each week.

While the process isn’t finalized, some of the tools mentioned during the conversation were Google Colab notebook, AssemblyAI’s API, Tally, Make, Airtable, YouTube, Wordpress and others.

Conversations like this are always happening in our Slack channel as it is in businesses around the world.

Teams are coming together to figure out how to automate time consuming processes that yield specific outcomes.

Did you know?

For the first time in history, AI played a pivotal role in research that won Nobel Prizes in 2024.

AI tools helped scientists achieve breakthroughs in fields like physics and chemistry by analyzing vast and complex datasets that would have been nearly impossible for humans to tackle alone.

This marks a significant milestone, highlighting AI’s growing influence in cutting-edge scientific discoveries.

This advancement illustrates how AI is not just a tool for automation but is becoming an essential partner in pushing the boundaries of human knowledge.

Source: CNN

This Week’s Conundrum
A difficult problem or question that doesn't have a clear or easy solution.

The AI Memory Restoration Dilemma:

AI has the potential to revolutionize treatment for memory-related diseases like Alzheimer's by using synthetic memories or AI-generated voices to help patients recall their pasts or maintain a sense of continuity in their lives. By simulating the voices of loved ones or reconstructing lost memories, AI could offer a powerful tool for improving patients’ well-being.

However, critics argue that relying on synthetic memories or AI voices blurs the line between reality and artificiality, potentially creating a false sense of connection or distorting personal histories.

The conundrum: Should we embrace AI technologies that create synthetic memories or voices to help patients with diseases like Alzheimer's, even if it means relying on artificially constructed realities? Or should we focus on preserving the authenticity of human experience, limiting the use of AI to avoid confusing patients with fabricated memories or interactions, even if it means fewer treatment options?

China Develops Photonic AGI Chip

China’s Taichi-II photonic chip has made headlines for its innovative use of light to power AI, resulting in a million-fold increase in energy efficiency compared to its predecessor. This photonic chip is poised to revolutionize AI processing, particularly in terms of speed and energy consumption.

BrainChip Unveils Neuromorphic AI for Edge Computing

BrainChip announced the Akida Pico, a milliwatt-level neuromorphic chip designed for AI inference in power-constrained devices, pushing the boundaries of low-energy AI processing. This technology mimics brain activity and enables real-time AI processing for wearable devices and other small electronics​.

Meta’s MovieGen AI: Personalized Video and Audio Creation

Meta has introduced MovieGen, a new AI-driven video generation model that allows users to place themselves in personalized video content. This model also integrates audio generation, offering users a more immersive multimedia creation experience​.

Adobe Launches Content Authenticity Initiative

Adobe has announced a new tool aimed at content authenticity, helping creators protect their work and manage issues related to AI-generated content. This step could reshape the creative industry by providing long-term solutions to concerns about intellectual property and AI​.

OpenAI Partners with Hearst for Content

OpenAI has teamed up with Hearst to access its extensive content library, which includes popular magazines like Cosmopolitan and Elle. This partnership helps OpenAI enhance its data for model training and is expected to mitigate future legal issues between content creators and AI developers​.

Cerebras Systems Files for IPO to Compete with NVIDIA

Cerebras Systems has filed for an IPO, setting its sights on competing with NVIDIA in the AI hardware space. The company’s unique chip design aims to address AI’s growing bandwidth and memory needs, potentially disrupting the GPU-dominated market​.

Bit Energy AI Introduces Energy-Efficient AI Model Math

Bit Energy AI has developed a mathematical approach, L-Mul (linear complexity multiplication), that replaces floating-point multiplications with simple integer additions, cutting energy consumption by 95%. This could significantly impact how AI models are trained and processed, leading to more efficient hardware.

Kobold Metals Uses AI to Discover Critical Energy Metals

Kobold Metals, leveraging AI technology, has discovered a massive copper deposit in Africa. The company, which started by using AI to survey and research potential deposits, is now expanding into the mining industry, aiming to support the growing energy demands for AI technologies.

University of Waterloo’s AI Agents Build a Self-Sustaining Startup

Graduate students at the University of Waterloo have developed a prototype of a self-sustaining startup composed entirely of AI agents, including a CEO, CMO, and CTO. These agents built a fully functional weather emergency reporting system, showcasing how AI agents can handle complex, coordinated tasks autonomously​.

Anthropic Announces API Batch Processing Discounts

Anthropic is offering a 50% discount on API batch processing for use cases such as content moderation, data analysis, and bulk content generation. This is particularly beneficial for businesses that don’t require real-time processing, making AI more affordable for routine operational tasks​.

Did You Miss A Show Last Week?

Enjoy the replays here on YouTube or take us with you in podcast form on Apple Podcasts or Spotify.