Newsletter #004
Gene Da Rocha MSc BSc (Hons)
Hello everyone,
I hope this message finds you well and that you're all gaining more knowledge about AI and related subjects.
My primary focus is on AI technology news, given the rapid pace of change in the field. I also cover seed financing to highlight the substantial investment opportunities available. Additionally, I have a keen interest in AI retail, drawing from my experience at companies like Volvo, Selfridges, and CarPhone Warehouse, among others.
I'm particularly interested in exploring how AI can enhance the customer experience, streamline the creation of new products and apps, and encourage greater participation in the AI industry.
if you want me to include a particular topic or subject, please comment or message me via Linkedin.
News:
AI:
Meta announces a bold, multi-billion dollar investment in AI, signaling a transformative shift in strategy. However, alongside this ambitious push, the tech giant forecasts lighter-than-expected revenue, sending shockwaves through the stock market. Discover how Meta's big bet on AI could redefine the future of technology.
At the TIME100 Summit, leading CEOs from startups and major tech companies convened to discuss the ethical dimensions of AI innovation. Their insights set the stage for a fiery debate on the future of technology, highlighting the critical juncture at which the tech world stands today. Dive into the dialogue that's shaping the next era of AI.
In the US, managing general agents (MGAs) known for their agility in launching specialized insurance products, the dual priorities of speed and customer experience are paramount. Operating with streamlined resources, these nimble entities face the challenge of delivering exceptional service without the luxury of expansive operations. MGAs are redefining efficiency and setting new standards in the insurance industry by using AI
Retail:
In a visionary move, Ikea Retail unveils an ambitious plan to equip 3,000 employees and 500 leaders with AI literacy skills, marking a significant leap towards a tech-driven future. The announcement, made by Ingka Group on April 12, signals a bold commitment from the retail giant, poised to revolutionize its workforce and operations. Dive into Ikea's pioneering initiative, poised to redefine industry standards and empower its workforce for the AI era
As AI transitions from a behind-the-scenes technology to a front-and-center shopping companion, questions arise about its role and reliability. Can AI truly understand our preferences and curate our shopping lists to our satisfaction? While the convenience of AI-driven tools is undeniable, their growing presence in retail sparks a debate about trust in technology versus human insight.
Here are three key considerations for incorporating AI into the consumer shopping experience:
Human Trust Over Machine Insight: Despite the sophistication of AI, shoppers show a strong preference for human recommendations. A vast majority (89%) trust product suggestions from friends or family more than those derived from an e-commerce site's AI analyzing their past purchases (51%) or from a generative AI model (38%).
Seeking External Validation: When making purchasing decisions, consumers often seek validation through reviews and ratings from trusted third-party sources, as well as feedback from their social circles. This indicates a continued reliance on human judgement alongside AI tools.
The Balancing Act: Retailers need to balance AI integration with an understanding of customer preferences for human interaction, leveraging AI to enhance rather than replace the personal touch that consumers value.
Funding:
HealthKey, a health-tech startup headquartered in London, UK, has successfully secured £1.13 million in Seed funding. This investment round was spearheaded by Aviva Ventures and received additional backing from Ascension, Oxford Capital, and Cur8 Capital. The startup plans to utilize this capital infusion to enhance its platform's features and increase integrations with various stakeholders in the health and life insurance sectors, as well as with health plan providers and corporate employers.
SAN FRANCISCO — JUICER, a company specializing in revenue management and pricing solutions for restaurants, has raised $5.3 million in a seed funding round led by York IE. As restaurants increasingly transition to an e-commerce model driven by delivery services and digital customer interactions, traditional approaches to pricing and promotions are becoming obsolete. JUICER is tackling this shift by offering a comprehensive suite of tools that utilize sophisticated machine learning algorithms and data analytics to optimize revenue strategies.
Startups are voicing concerns about raising 'only' $1 million in seed rounds, suggesting a shift in expectations for initial funding. Sam Altman, president of Y Combinator, addressed this sentiment in a series of tweets. He advised that starting a business doesn't require massive capital investment. Altman emphasized that founders should take pride in efficiently utilizing minimal capital to establish and grow their companies, challenging the notion that larger investments are necessary for success right from the start.
Local LLM's -
Microsoft’s Phi-3 shows the surprising power of small, locally run AI language models
Microsoft has unveiled its latest innovation in AI language models with Phi-3-mini. This lightweight model is designed to operate locally, offering a range of benefits for users. Unlike traditional large language models, Phi-3-mini is simpler and more cost-effective, making it accessible even on smartphones without an internet connection.
Key Takeaways:
Microsoft introduces Phi-3-mini, a small, locally run AI language model.
Phi-3-mini is simpler and less expensive to operate compared to large models.
This lightweight model brings AI capabilities to devices without an internet connection.
Phi-3-mini contains 3.8 billion parameters and performs on par with larger models.
Microsoft continues to innovate and optimize AI language models for wider accessibility.
The Microsoft Phi-3-Mini is a fictional device, so I'll provide a hypothetical example of Python code that could theoretically be used to interface with a generic AI chip or device. The example will demonstrate how to send a simple operation command to the device and retrieve its response.
Here's an example setup in Python:
Explanation:
Setup Connection: The setup_connection function initializes a serial connection to the device using a specified port and baud rate.
Send Command: The send_command function sends a command to the device. Commands would be specific to the device's API and capabilities.
Read Response: The read_response function reads the output from the device until a newline character is detected, useful for getting the result of the command sent.
Note:
Ensure that the device (in this case, the fictional Phi-3-Mini) supports serial communication or has a specified API or protocol for communication.
Replace "COM3" with the actual serial port to which your device is connected. On Linux or macOS, this might look like "/dev/ttyUSB0" or similar.
Make sure to install the pyserial library, which is used for handling serial communication. You can install it via pip:
Gemma: Introducing new state-of-the-art open models
Gemma is a new family of open models by Google. It is light and advanced. It is made for safe AI making.
It comes in two sizes, Gemma 2B and Gemma 7B. They have different uses. Gemma helps make AI safe. It is used everywhere and works well with JAX, PyTorch, and TensorFlow.
Create an image of Gemma, the state-of-the-art open model, showcasing her advanced features and capabilities as the newest addition to your tech arsenal. Show her in action, surrounded by cutting-edge technology and equipment, with a sleek and modern design that highlights her superiority over other models. Use cool and futuristic colors to convey a sense of innovation and sophistication, and includes subtle details that emphasize her power and precision.
Key Takeaways:
Gemma is a family of state-of-the-art open models developed by Google.
It is designed for responsible AI development and offers model weights in two sizes: Gemma 2B and Gemma 7B.
Gemma provides pre-trained and instruction-tuned variants, as well as a Responsible Generative AI Toolkit.
The models are available worldwide, support multiple frameworks, and deliver industry-leading performance.
Gemma can be used for commercial purposes by organizations of all sizes.
Example of how to use Google Gemma locally.
Setup
First, you'll need to install the necessary libraries. You can do this using pip:
Text Generation Example
Here's how to run a simple text generation model locally:
Text Classification Example
#ArtificialIntelligence #MachineLearning #DeepLearning #NeuralNetworks #ComputerVision #AI #DataScience #NaturalLanguageProcessing #BigData #Robotics #Automation #IntelligentSystems #CognitiveComputing #SmartTechnology #Analytics #Innovation #Industry40 #FutureTech #QuantumComputing #Iot #blog #x #twitter #genedarocha #voxstar #wiredvibeapp
@ArtificialIntelligence @MachineLearning @DeepLearning @NeuralNetworks @ComputerVision @AI @DataScience @NaturalLanguageProcessing @BigData @Robotics @Automation @IntelligentSystems @CognitiveComputing @SmartTechnology @Analytics @Innovation @Industry40 @FutureTech @QuantumComputing @Iot @blog @x @twitter @genedarocha @voxstar







