AI: Weekly Summary. RTZ #304

Welcome to the Saturday weekly AI summary here on ‘AI: Reset to Zero’. 

1. Nvidia Developer GTC 2024 cements Jensen Huang long-term AI Strategy: Nvidia rolled out its formidable AI strategy in a stadium filled with over 11,000 customers, partners and developers. This product strategy points to a multi-platform strategy beyond mere ‘GPUs’, and underlines a well-thought out strategy against rising competition from its best customers and startups. Lots in the keynote and in follow up write-ups. But key highlights are that the new Blackwell family of AI GPU products offer 4x the training and 30x the Inference capabilities of current Hopper H100 GPUs, with far more efficient power consumption. Key new service of note was Nvidia NIM, Nvidia inference Microservices, which when combined with Nvidia NEMO, a software set of additional microservices for enterprises, allow new ways to deploy AI with their own data and choice of AI models, open or closed. Finally, Nvidia Omniverse based service updates and upgrades offer new ways for enterprises across industries to train their own AIs at an industrial scale, using ‘Digital Twins’ on synthetic environments and data. Then hosting them on AI DGX cloud services, offered through partners like Amazon AWS, Microsoft Azure, Google Cloud and others. Deeper dive here.

   

2. DOJ vs Apple now makes Four US Big Techs with Government Actions: The DOJ’s lawsuit against Apple’s smartphone franchise was announced with a full press conference and details. This action and other actions now brings four big US tech companies under US regulatory scrutiny and actions. It runs parallels with similar actions in the EU for US big tech as well. Most will take years to resolve, but it all represents accelerated headwinds as big tech tackles Big and Small AI. Further dive here.

   

3. Microsoft Inflection Deal highlights new way for Tech M&A: Microsoft surprised the AI industry with a creative way to acquire the core parts of Inflection AI, an LLM AI Unicorn, along with its key founders and employees. Highlights mounting pressures and scaling challenges for LLM AI startups. Lots of moving parts to the deal that may help against regulatory speed bumps on traditional tech acquisitions. Lots of implications for the ‘Big AI’ companies, competitors and investors. More here.

   

4. OpenAI GPT-5 release soon this year?: Media leaks point to the next major release of OpenAI’s GPT-5 later this year. It’s a release long anticipated by the industry at large, especially given the rapid pace at which competitors Google, Anthropic and many others are ramping up their LLM AI offerings this year. More to read here.

   

5. Reddit another greenlight for IPOs in 2024: The US IPO market may be open for other technology companies with a successful Reddit IPO. Another ‘overnight’ success over two decades. Reddit of course was the first YC company, and founder Paul Graham has a heartfelt personal story to tell of how it inspired YC itself to get started. Broader context for US IPOs beyond tech here.

   

Other AI Readings for weekend:

1. Google Gemini 1.5 now offers over 10 million tokens in user ‘context windows’, which rapidly ramps the utility of LLM AI models for customers large and small.

   

2. Given the focus on the DOJ action on Apple this week, I thought it’d be useful to highlight my take again on the AI opportunities ahead that are taking shape for the company this year and beyond.

   

Thanks for joining this Saturday with your beverage of choice. 

Up next, the Sunday ‘The Bigger Picture’ tomorrow. Stay tuned. 

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)