AI: Weekly Summary RTZ #653

1. OpenAI tests upper limits of AI pricing: Price discovery continues in LLM AIs, with OpenAI pondering $20,000/month subscriptions for its most capable AI Reasoning offerings. These would be Phd level agents, capable of results across vertical domains like coding, math, and other science disciplines. These pricing tiers are a possible solution to the high early demand for the company’s higher priced AI offerings, and the variable costs involved to providing the necessary AI compute. Especially in a world where the compute is still in short supply despite the rapid AI infrastructure deployments like Stargate and other AI data centers. More here.

   

2. TSMC next one to commit $100 billion in US: Taiwan Semiconductor is the next big global big tech company stepping up with a $100 billion plus tech/ai outlay in the US. It is a necessary rite of passage in the current geopolitical US/China environment. Apple with its recent $500 billion announcement, OpenAI/Softbank with its similar sized Stargate announcements, were recent examples. These projects in many cases are continuations of investment programs already in place, with incremental adjustments as needed. Expect more of these across the AI Tech stack, from different hardware, networking and power related investments up. More here.

   

3. AI Ad Tech Broadens: The race to build AI applications, tools and services for advertisers expands beyond the core big tech companies like Meta, Google and others. A wider range of AI tech companies are finding ad generation applications of its products, that are then finding partnership opportunities with the bigger tech companies. It all underlines the need for AI innovation not just for end users businesses and consumers, but also the advertising ecosystem to expand their opportunities to reach end customers. Meta in particular remains at the forefront of this trend, having generated more revenue from this focus than its peers for now. More here.

   

4. Apple Intelligence delayed again: New reports on delays around Apple Intelligence for Siri and related services indicate ongoing work needed to revamp Siri with LLM AI technologies. The plan to have some of these features ready for iOS 18.4 in April now get pushed back. Apple meanwhile continues to upgrade all its iphone, mac and ipad hardware upgrades to higher minimum RAM memory configurations of 16 GB or more. Apple Intelligence is also being deployed across Apple’s app ecosystem, and that remains a concurrent effort with the Siri upgrades. More here.

   

5. Amazon Alexa AI releasing: Amazon announced details of is LLM AI upgraded Alexa+ with a series of presentations and demos. This has been a long planned effort, with delays, but now seems to be on track for release later this year. The capabilities will add multimodal voice AI capabilities to Alexa with agentic capabilities to handle user personalized tasks. There is a subscription tier planned of $20 a month, which can be bypassed with a $15/month plus Amazon Prime subscription. Amazon Alexa, along with Apple Siri above remain the two multimodal device contenders this year along with standalone efforts by OpenAI, Google and others. More here.

   

Other AI Readings for weekend:

1. Reinforcement Learning pioneers get Turing awards. More here.

   

2. New autonomous AI Agent Manus from China, with DeeoSeek style innovation.

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)