Neo-French Revolution

The European Union has been resistant in its approach towards AI. While ChatGPT, the poster child of AI grew exponentially across the world, it had a meagre 3.8% of users from Europe. The GDPR and European AI Act have obstructed the expansion of AI in Europe. But it’s going to change and France is set to lead another revolution — the new-age tech revolution.

Away from the limelight, France has already been catching up in the AI race. Hugging Face, headquartered in New York, has a strong presence in France. The company has a number of partnerships with French universities and research institutions and also, and its research team happens to be based in Paris.

Similarly, Meta’s FAIR (Fundamental AI Research) hub, launched in 2015 in Paris, has been responsible for building and deploying LLaMA – a state-of-the-art foundational large language model designed to help researchers advance their work in this subfield of AI.

Early this week, a seed funding of $113 million in a month-old French AI startup, Mistral AI, pulled everyone’s attention towards the growing AI ecosystem in France. Mistral AI is building an open-source solution in the LLM space and “making AI useful”. 

Besides private companies, the government too is pushing for the ‘AI-fication’ of the country. This week, French President Emmanuel Macron, at the VivaTech 2023 event, announced that the country would invest 500 million additional euros in AI to go “much faster and stronger”. “The worst case scenario would be a Europe that invests much less [in AI development] than the Americans and the Chinese, and that would start by creating regulation. This scenario is possible, it would not be the one I would support," he noted. 

"We must simultaneously accelerate research, innovation, and the creation of a powerful ecosystem and participate in regulation whose scope must be as wide as possible."

Read the full story here.

Quantum Breakthrough 

IBM, in collaboration with UC Berkeley researchers, has made a significant breakthrough in quantum computing. They conducted an experiment that suggests quantum computers will soon outperform classical computers in practical tasks. The team developed error-mitigating techniques to overcome quantum noise, which affects calculation accuracy. Using the Eagle processor's 127 qubits, they achieved accurate expectation values for complex circuit volumes, surpassing classical brute-force computation. 

This showcases the potential of quantum computing, even before achieving fault tolerance. The success was attributed to advancements in coherence, calibration, and noise control across the large device. This development is crucial for quantum algorithmic design and represents a major step forward in manipulating noise.

Read the full story here.

LangKit Completes Missing Link in LLMs

Recently, WhyLabs, an AI observability platform, released LangKit, an open-source monitoring toolkit. LangKit is designed to monitor the outputs of LLM-powered chatbots like Bing Chat, GPT-4, Baidu, and Google Bard. It provides measures for evaluating the quality of input and output text from chatbots. The toolkit works with ecosystems such as Hugging Face and LangChain to improve its usefulness. 

Enterprises can use LangKit to obtain access to LLM behaviour during deployment and mitigate concerns such as hallucinations, toxicity, and harmful prompts. The toolbox provides steps to prevent against security threats and assure specification compliance. As more businesses embrace LLMs, solutions like LangKit help to make these models more secure for deployment.

Read the full story here.

Microsoft’s New Peak

Microsoft’s investments in AI are paying off. The tech giant, at the helm of the AI race, on Thursday saw its shares surge by 3.2% taking the company’s capitalisation to $2.6 trillion. Besides AI, Microsoft's cloud service is also soaring with profits and customers that have added to the growth of the company. 

In the last four quarters, Microsoft generated around $208 billion in revenue. The executives of the company predict that it would record a $10 billion revenue this year.

Read the full story here.