a performer on stage playing a guitar next to large representations of a potato chip bag labeled "Selling Out Potato Chips" and an open pizza box labeled "Corporate Jingles." In the audience, people of various emotions are reacting with expressions ranging from shock to sadness. The city skyline in the background adds to the urban setting of this performance.
Newsletter

Perplexity AI is Selling Out! 😱🚫💸

Happy Wednesday from AI Smart! 😊 Join our amazing community at AI Smart Insiders where you’ll get the inside scoop on saving time and enhancing productivity with cutting-edge AI tools. When

Happy Wednesday from AI Smart! 😊 Join our amazing community at AI Smart Insiders where you’ll get the inside scoop on saving time and enhancing productivity with cutting-edge AI tools. When you sign up, you’ll receive an invitation to an exclusive workshops an opportunity you definitely won’t want to miss!

This Week’s Insights

Perplexity AI is Selling out! 🌟🤖

Selling out can often feel like that moment when you open a bag of chips, only to find it half empty—disappointment galore! 🍟 To put it plainly, it’s the equivalent of your favorite indie band signing with a major label, suddenly swapping their soulful lyrics for commercial jingles. 🎤✨ When companies cash in their ideals for cold hard cash, we’re left with a sense of betrayal, much like discovering your go-to pizza place started using pineapple as a topping (just why?! 🍍❌). So let’s face it: selling out blends the excitement of a product launch with the cringe of an awkward family reunion! 😬🤦‍♂️ So beware, friends! When selling out is on the table, do we really need more mediocre options crowding our choices?

Perplexity AI, with backing from notable investors like Nvidia and Jeff Bezos, is gearing up to roll out advertisements on its AI-driven search platform by the fourth quarter. This move comes hot on the heels of their recent program launch for publishers, which has already onboarded big names like TIME, Der Spiegel, and Fortune. The initiative aims to share revenue from user interactions that reference a publisher’s content, offering a fresh revenue stream in the publishing industry. This means you pay for the media…again. As AI continues to reshape digital landscapes, Perplexity AI’s entry into ads search could promise a more tailored and interactive experience for users while providing a new platform for advertisers. Do you believe that? 🤦‍♂️

This strategic push is designed to position Perplexity AI as a formidable player in the ongoing AI-search race—a field currently dominated by giants like Google and Microsoft. Microsoft has been leveraging OpenAI technology for Bing, while Google has introduced AI-powered summaries. With a recent funding boost of $62.7 million and a valuation now exceeding $1 billion, Perplexity AI is poised to potentially alter the competitive dynamics of the industry, maybe. For more details, check out the original article.

a performer on stage playing a guitar next to large representations of a potato chip bag labeled "Selling Out Potato Chips" and an open pizza box labeled "Corporate Jingles." In the audience, people of various emotions are reacting with expressions ranging from shock to sadness. The city skyline in the background adds to the urban setting of this performance.

AI Bot 🤖 for Mayor?

In a unique twist to traditional politics, Victor Miller is campaigning for mayor of Cheyenne, Wyoming, with the promise of letting an AI bot, VIC, take the helm if he secures the position. Built on OpenAI’s ChatGPT 4.0, VIC is designed to navigate the intricacies of governance more efficiently than a human counterpart. Miller envisions VIC sifting through hundreds of documents from city council meetings, offering policy recommendations, and ensuring decisions are well-informed and transparent. That doesn’t sound so bad.

Despite challenges like legal restrictions on AI candidacy and concerns expressed by state officials, Miller is set on showcasing the potential of AI to revolutionize public service. The bot aims to prioritize transparency, economic development, and the integration of new technologies to improve public infrastructure. (and create a lot of buzz for Miller) By processing vast amounts of information rapidly, VIC promises a productivity boost that could redefine efficiency in governance.

For more details, dive into the original article.

AI robot dressed in a formal mayor's suit, complete with a shiny name tag and a confident smile. The robot is wearing a campaign ribbon with a futuristic design, and the setting is a bustling political rally or press conference, with a large crowd of people from diverse backgrounds holding signs and cheering.

Team Collaboration Boosts with AI 🤝

MIT’s CSAIL has introduced an innovative AI team coordinator to streamline teamwork by ensuring alignment among team members’ roles and beliefs, to enhance overall collaboration. This AI assistant, unveiled in a recent study at the International Conference on Robotics and Automation, operates by assessing the team’s understanding of the overarching plan. By intervening when necessary, the AI provides real-time updates and instructions. This is particularly valuable in high-stakes environments like search-and-rescue operations, medical procedures, and strategic video gaming. For instance, it can notify rescue teams about covered areas. It can also update medical staff on patient status. This ultimately reduces misunderstandings and improves decision-making.

The potential applications of this AI assistant are vast. It can enhance task organization and boost productivity by preventing role confusion. Researchers at MIT are exploring future improvements. These include integrating machine learning to generate new hypotheses and extending its capabilities to more complex scenarios. By incorporating probabilistic reasoning and recursive mental modeling, this technology enhances current collaborative efforts. It also sets the stage for more advanced applications, promising significant improvements in various fields requiring intricate teamwork and coordination.

HeAR-ing Health Improvements 🎧✨

Google’s latest AI innovation, Health Acoustic Representations (HeAR), is helping in healthcare by analyzing coughs and breathing sounds to assess health conditions. Developed through self-supervised learning, it is trained on over 300 million audio clips. HeAR leverages the Transformer architecture to generate compact representations of audio data. This capability is crucial for detecting health-related sounds, classifying coughs, and estimating lung functions. Notably, HeAR outperformed existing models in a series of benchmarks. It achieved a commendable AUROC of 0.739 for tuberculosis detection, outshining the previous leader, TRILL. Its promise extends to estimating critical lung function parameters like FEV1 with impressive accuracy. This suggests its potential as a tool for accessible screening of lung diseases such as COPD. The system’s ability to process short audio clips could transform remote healthcare, especially in resource-limited areas.

Still in the research phase, HeAR’s diagnostic applications await clinical validation. Current limitations include the processing of only two-second audio clips. Google’s vision includes refining these capabilities for mobile use through model distillation and quantization. This paves the way for more efficient applications. The initiative is backed by the StopTB Partnership, targeting tuberculosis eradication by 2030. Researchers can access the trained HeAR model and anonymized data for further exploration. For those interested in delving deeper into this promising technology, more details can be found here.

Why You Need To Give AI Templates For Better Outputs💡

In a recent study conducted by researchers at UCLA and Amazon, large language models (LLMs) have shown distinct capabilities in different types of reasoning. The research highlights that LLMs excel in inductive reasoning, where they are adept at learning rules from examples (aka templates). However, their prowess diminishes in deductive tasks that require following specific instructions. This is where the innovative SolverLearner framework comes into play. By utilizing an external code interpreter, SolverLearner isolates inductive reasoning, enabling a clearer analysis of how LLMs learn patterns.

The study’s implications for real-world applications are substantial. While LLMs can effectively discover patterns, their performance may falter when confronted with unfamiliar scenarios. This means that despite their apparent ability to follow logical instructions, they might rely heavily on familiar patterns from their training. Providing them with patterns to follow using templates can help. However, this reliance can lead to decreased performance in novel situations. SolverLearner offers a pathway to ensure the accurate learning of input-output mappings. Its use is limited to environments with verification tools like code interpreters. This research serves as a reminder of the complex nature of these models and their evolving role in technology. It also highlights the super simple fixes we can use to help AI with its growing pains. 🤝

Tools We Used This Week  🧰

  • Jasper.ai – (uses multiple LLM’s) Generate high quality marketing copy for emails, ads, websites, listings, blogs while maintaining your brand voice and security.
  • Opus Clip – Chop up long form videos into short form content for social media.
  • Gamma – AI tool for creating presentations, documents, and websites quickly.

Join the Conversation  🔊

This isn’t just about sending newsletters into the void—we’re building a lively hub here. If you’ve got insights itching for daylight, questions that stump the wise, or ideas that buzz louder than a caffeinated bee, send them flying to info@aismartmarketing.com. Your contributions are the fuel for our content engine, driving us to enlighten and inspire.

See you in the Lab!

Share with a friend or Subscribe!