📈 How are companies using AI?

➜ Citi Partners With Google Cloud for AI Infrastructure Modernization. Two tech giants hand in hand. Citi is partnering with Google Cloud to modernize its infrastructure and enhance AI capabilities, with plans to migrate multiple workloads to Google's cloud platform. The collaboration will enable Citi to execute millions of daily computations in its markets business while developing new generative AI tools for document processing and customer service. 

PYMNTS

➜ Morgan Stanley’s GenAI Assistant For Supporting Staff. Morgan Stanley has rolled out a GenAI-powered assistant called AskResearchGPT, to support staff in surfacing and filtering insights from its research library. AskResearchGPT can be used to look for data, obtain insights and summarize information from more than 70,000 proprietary reports published annually by the bank. Morgan Stanley had pioneered the use of generative AI on Wall Street via a partnership with OpenAI - looks like the other banking major players have to keep up!

Finextra 

➜ Next-Gen Shopping with AR from Meta. Shoppers may soon browse digital clothing racks and try on virtual jewelry from their living rooms with the help of AI, as Meta pushes its vision of a new AR prototype headset called Orion. The technology combines AI-powered recommendations with augmented reality shopping experiences. The new glasses “seamlessly integrate contextual AI that can sense and understand the world around you in order to anticipate and proactively address your needs”, Meta wrote in its announcement. Available in a couple of years. Love to try out.   

PYMNTS

➜ IBM’s Granite 3.0: New Open-Source LLM for Enterprise.  The new Granite 3.0 8B and 2B language models are designed as ‘workhorse’ models for enterprise AI, delivering strong performance for tasks such as Retrieval Augmented Generation (RAG), classification, summarization, entity extraction, and tool use. The models are designed to be fine-tuned with enterprise data and seamlessly integrated across diverse business environments or workflows. 

IBM

➜ Claude AI Can Now Control Your Computer. One prompt to rule them all? Claude announced a refreshed Claude 3.5 Sonnet — available now for you to use on Claude.ai. The new “Computer Use” API, allows an LLM to control your computer and do arbitrary things like browsing websites, downloading and running files, essentially imitating a person sitting at a PC. A slew of startups doing browser agents feel obsolete overnight. 

TechCrunch 

➜ Midjourney Launches AI-Powered Image Editor. Now you can finally transform your creative process with Midjourney’s web editor, that allows users to modify, retexture, and expand both generated and uploaded images using natural language prompts. During the initial release they have opened the feature for yearly subscribers and members with at least 10 000 images. Fun! 

VentureBeat 

💡 What are the nerds up to?

➜ On the Path to Superintelligence. Alexandr Wang from Scale.ai described his vision of the future of AGI superintelligence at the WebexOne 2024 event during a panel focused on elevating human connection with AI. Wondering how much it is that we’re spending on training AI models this year? It’s $200 billion (which is approaching the US defense budget of $800 billion). Building superintelligence “is one of the biggest human projects of time” remarks Alexandr.  

YouTube

➜ OpenAI CFO Says AI Isn't Experimental Anymore. OpenAI CFO Sarah Friar says AI isn't experimental anymore. Banks, financial institutions and fintech's are using it everyday in their business. According to Sarah “AGI is closer than you think, it will blow your mind what’s already coming, we have learned to take reasoning models on top of the GPT models and the interplay between those. You’re starting to see some incredible outcomes on a PhD level, where you have it in your pocket. The power there blows my mind”. Listen to the interview held by Ed Ludlow at the Money 20/20 Conference on Bloomberg Technology. 

YouTube

➜ Bruce Schneier: Public AI Models Need Full Transparency. Speaking at Atlanta's SOSS Fusion Conference, security expert Bruce Schneier addressed the critical differences between human and AI capabilities, emphasizing that AI isn't better at training than humans are – they are just faster. Schneier’s take is that AI's superior speed and scale are transforming fields like high-frequency trading, stating that even if AI could do a mediocre job at reviewing all of the source code, that would be phenomenal.

ZDNET

➜ Godfather of AI on AI Exceeding Human Intelligence. Geoffrey Hinton, former Vice President of Google and sometimes referred to as the ‘Godfather of AI’, speaks to BBC’s Newsnight about how AI is trying to take over human intelligence. Hinton claims that “almost everybody he knows believes that AI will exceed us, it’s just a question of when”. Futuristic vision or the inevitable future? 

YouTube

➜ How to Write Code Using Cursor: A Comprehensive Review. Interesting read by Tom Yedwab in his latest blog post  —  a personal review of Cursor. Yedwab summarizes: "I am confident that at the time of writing this, Cursor is the best example of the potential of LLM coding assistants”. Nicely written, definitely worth your while. 
Arguing with Algorithms

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.