AI Financial Advisor💸 AMDs HBM⚔️ G7 AI conduct 🤝

AI 🤖 

G7 countries and the European Union urge companies to take appropriate measures to identify, evaluate, and mitigate risks across the AI lifecycle, as well as tackle incidents and patterns of misuse after AI products have been placed on the market. 
Full Story

Microsoft will invest A$5 billion in expanding its hyper-scale cloud computing and AI infrastructure in Australia over the next two years – the single largest investment in its forty-year history in the country. This investment will also grow Microsoft’s local datacentre footprint from 20 sites to a total of 29
Full Story

Personalized Maps Feature uses artificial intelligence to analyze photos shared by the community and provide personalized results, helping users find spots that match their preferences. Google is also adding other features, such as an expanded immersive city view and AI-driven 3D scene visualizations. These updates are part of Google's desperate efforts to stay ahead in the AI race.
Full Story

Tech 🧑‍💻 

AMD is set to compete with Nvidia but not with GPU. They will heavily rely on High Bandwidth Memory (HBM). This processor-in-memory (PIM) approach allows data to be processed directly in memory, reducing data movement, enhancing performance, and increasing energy efficiency.
Full Story

Nintendo patented a dual-screen gaming device that could operate independently. Apparently, This is based on leaked information so we need to wait for official announcements.
Full Story

Business 💵 

The consulting industry has experienced a lot of evolution since Generative AI adoption and consultants fear replacement but it is not set to happen anytime soon rather this article covers the ways to leverage it.
Full Story

Where AI meets Financial Advisors AI financial advisors offer cost-effective and 24/7 personalized financial guidance, making financial advice more accessible. They can address a broader range of financial decisions, including savings, insurance, and debt management. However, AI cannot yet replace human advisors, as it lacks some essential qualities for effective financial advice like Bias, Empathy, Accuracy, and Consistency.
Full Story

Tools ⚒️ 

StoryBee is a platform that uses AI to generate stories for kids. Users can provide hints or themes and customize the stories by selecting genres, styles, and visual aesthetics. StoryBee can be used to inspire creativity, help with writing exercises, and introduce kids to the wonders of AI in a fun and engaging way.
Pricing: Freemium

Codecompanion AI is an all-in-one coding assistant that helps developers prototype, code smarter, and scale their productivity. It can read, write, and update existing or new code and use natural language to search the entire codebase. It stores data locally so it’s safe.
Pricing: Free

Research 🧠 

Zephyr Direct Distillation LM paper talks about making a smaller AI language model. The previous paper distilled supervised fine-tuning (dSFT) had issues understanding natural language. To fix this they use AI Feedback. They start with ranked outputs from a teacher model and apply distilled direct preference optimization. This takes fewer hours to train the model without extra data. They claim Zephyr-7B outperforms a well-known Llama2-Chat-70B without human annotation.
Paper

Fact-Checking with LLM: Autonomous fact-checking, using machine learning, has become crucial in combating misinformation. This study evaluates their fact-checking abilities, showing that LLMs perform better with contextual information, with GPT-4 outperforming GPT-3. However, accuracy varies with query language and claim veracity. While promising, caution is required due to inconsistent accuracy. Further research is needed to better understand when these models succeed and fail in fact-checking.
Paper

Tip of the day 💡 

Python: Use Comprehensions for Side Effects

You can use list comprehensions for their side effects, such as printing elements, without creating a new list. Use an underscore _ as the variable name to indicate it's used only for side effects:

my_list = [1, 2, 3, 4, 5]
_ = [print(item) for item in my_list]