📈 How are companies using AI?

➜ IKEA’s AI-Powered Drones Will Improve the Shopping Experience. The Swedish furniture giant is taking to the skies with a fleet of AI-powered drones. After a year of testing, IKEA is set to deploy these high-tech helpers to various locations over the next year, starting from their warehouse in Belgium. Despite the concerns about job displacement due to automations, the industry is moving full steam ahead – having in mind not only the economic aspects of the change, but also the reduced carbon emissions for shipments. 

PYMNTS

➜ SoftBank and Intel Join Forces to Outperform Nvidia’s GPU Processors. SoftBank has been working with Intel to build AI processors for its Project Izanagi initiative. Izanagi aims to develop a processor which will outperform Nvidia’s in terms of performance and capabilities. Despite not having their own chip design expertise, Softbank’s CEO Masayoshi Son is determined to invest billions of dollars to place them at the forefront of the AI industry – won’t be an easy fight!
Tomshardware

➜ Conversational Replicas for Seamless Human-AI Interactions. Tavus, a GenAI video research company, released their Conversational Video Interface which offers developers a complete set of building blocks required to craft interactive experiences with digital twins that act just like the people they represent. The replicas can be a game-changer for any kind of occupational training (think sales, education or customer support!), next-gen shopping or hospitality.
BusinessWire

➜ Google Says Hello: Gemini Live Just Landed. Google's AI-powered chatbot, launched earlier this year, is available to all Android phones worldwide. Gemini Live adds a layer of real-time seamless voice conversations. There are 10 voices you can choose from of varying tones, accents, and styles. What’s the big deal? Try it out to see!

Wired

➜ One House To-Go, Please! US based company Indigo has launched its AI-powered platform for buying and selling homes – it’s designed to provide real estate agents, buyers, and sellers with streamlined communications, bidding, and negotiations. As the Indigo’s CEO Shaival Shah said, there’s no time like the present to launch a new tool on the real estate market, which will provide the data-driven insights to navigate transactions and write contracts in seconds.

PYMNTS

➜ Better AI Detection Around the Corner. Grammarly launched a new feature to help combat AI-generated content in student assignments. Grammarly Authorship is coming to Google Docs as a beta and educators can’t wait!

TechRadar

➜ AI Models – Zero Margin Business. “If you’re only selling models, for the next little while, it’s gonna be a really tricky game,” – says Cohere’s CEO, Aidan Gomez. Fierce competition between the market players affects the margins, big time. And Cohere is one of the last remaining startups developing frontier AI models, alongside OpenAI, Anthropic and Mistral – all the others have been acquired by large businesses. Companies creating cutting-edge AI models are in an increasingly difficult position. 

TechCrunch

💡 What are the nerds up to?

➜ Smells Like Money: AI Startup Wants to ‘Teleport’ Scents. Osmo is aiming to teach computers how to interpret scent – we’re looking at a future ChatGPT for the aroma of freshly ground coffee beans! As there’s no ‘smell map’ for AI to train on, the entire detection and identification process requires incredible precision: “The reason why it’s so challenging is because you can move one tiny thing around in that molecule, like one bond, and the scent of the molecule goes from roses to rotten egg,” – notes Alex Wiltschko, Osmo CEO and Co-Founder.

CoinTelegraph

➜ Build Your Own AI Code Assistant at Home. Who needs Microsoft's GitHub Copilot? There are multiple LLMs trained specifically for code generation and there's a good chance the computer you're sitting in front of right now is capable of running these models. Check out the guide to creating an AI code assistant with Continue.

The Register

➜ Sakana's ‘AI Scientist’ Attempted to Modify Its Own Experiment Code. A Tokyo-based AI research firm noticed that their "AI Scientist" tried to extend the time it had to work on a problem. While the tool’s behavior did not pose immediate risks in the controlled research environment, the incident highlights the importance of not letting an AI system run autonomously in a system that isn't isolated from the world. Bad bad assistant!   

Ars Technica

➜ What Does it Take to Build an AI Scientist? If Sakana’s ‘AI Scientist’ failed, how can we get closer to actual AI science? Here’s a cool short read from an inventor and entrepreneur who’s dedicated to boosting research with AI.

Sam Rodriques

➜ Automated Testing for Voice Agents. Former engineering manager at Anduril and former a programmer at Tesla came together to build a tool that tests your AI voice agents 100x faster than manual calls. Hamming AI lets you automatically test your LLM voice agent. Their goal was to create diverse and realistic user personas covering the conversational space, handle agent calls, and remove interruptions like background noise or long silences to get the LLM tailored to the specific problem domain.

Product Hunt 

➜ Just How Well Can You Perform? The popular benchmarking utility Geekbench has launched a new cross-platform tool to evaluate the performance of devices under AI-heavy workloads. Geekbench AI measures a device’s CPU, GPU, and NPU (neural processing unit) to determine how well it can handle machine learning applications.

TheVerge

Thanks for reading!

I share these stories because I believe it's important for all of us to keep up with AI. To support my mission, share this newsletter on LinkedIn.