AI: Weekly Summary. RTZ #808

1. OpenAI’s GPT-5 & Open Source OSS: OpenAI had a very big product week, with both its long anticipated GPT-5 and GPT-OSS, ‘Open-Weight’ AI models. The key takeaways on GPT-5 are that it is simplified and unified such that their 700+ million weekly users don’t have to pre-pick which AI model to click on, from a drop down menu. Now GPT-5 is designed to ‘route’ the queries to the right underlying OpenAI models. Also GPT-5 is broadly available even in free tiers. Although higher usage means higher priced tiers. Also businesses and developers can of course pick specific OpenAI models still via the APIs. On the OpenAI open weight ‘OSS’ models, the offerings should sate businesses and developers who want to use open source to craft their own AIs, but stay within the broader OpenAI ecosystem. Also this helps OpenAI publicly respond to the global popularity of Chinese open source models. And of course address Meta’s Llama LLMs. Both sets of products now will of course be rigorously reviewed. discussed, and assessed by the global market. So stay tuned for that. More here.

   

2. Google, Anthropic et al keep pace: Meanwhile, Google, Anthropic, Elon’s xAI/Grok and others had a flurry of various AI announcements as well. Including robust activity from AI companies in China like Alibaba and others. Google’s AI coding model ‘Jules, now out of beta, and powered by Gemini 2.5 Pro, offers asynchronous, agent-based coding tools integrated with Github. Anthropic’s Claude Opus 4.1 offers upgrades to previous models on agentic, real world coding and reasoning tasks. Elon Musk’s ‘guardrails light’, xAI/Grok introduced ‘Spicy’ modes for Grok’s image and video generators. Complete with NSFW options freely available on celebrities like Taylor Swift. Alibaba’s AI announcements were on multimodal capabilities, healthcare and animation innovation. Also, OpenAI, Google and Anthropic are now working with the US government on inclusion of their models in government procurement schedules. Expect OpenAI’s announcements above to nudge more out of its peers in coming months. More here.

   

3. AI Coding in ‘Frenemies’ stage: As above moves by Google, Anthropic and OpenAI illustrate, the big LLM AI companies are rapidly rolling out their next gen AI Coding models and tools. Whilst their core products are also driving the runaway success of AI Coding Unicorn startups like Anysphere’s Cursor and others. A classic tech ‘Frenemies’ story, not unlike the larger one under way with Nvidia and its core big tech companies. Those companies, accounting for almost half of Nvidia’s fast-growing revenues, are also rolling out their own AI training and inference chips with billions being invested. A twist here is the white hot AI Talent wars catalyzed into the stratosphere by Meta’s Mark Zuckerberg. That blurs the line between competing in AI coding vs acqui-hiring their founders for the big tech companies own AI Labs. Google $2.4 billion acquisition of the Windsurf AI Coding team for Deepmind is a case in point. A new multi-layered chess game for sure. More here.

   

4. Big Tech Models Morphing with AI Infra Boom: With Wall Street now green lighting the multi-hundred billion annual big tech spend on AI infrastructure like data centers and talent, the next big question is clear. How this impacts their business models and financials fare long-term. Core here is their world-leading software and ad sales model are changing, and not for the best near-term. They are now being burdened with ‘heavy’ infrastructure capex more akin to the Telecom spending of a trillion dollars in the Internet decade in the nineties. And they’re not necessarily ‘one-time’ investments. But ongoing for years. So far, investors have not exhibited high concerns, in their assessment of big tech company valuations. Likely due to the assumption that their models will ‘revert to the norm’ sooner than later. That may not be the case given the multi-year AI innovation and AI scaling challenges ahead. Especially as big tech focuses on ‘Physical AI’ opportunities like AI devices, robotics, robotaxis etc.. Definitely an issue to keep in mind going forward. More here and here.

   

5. Apple & Nvidia Manage Tariffs: Both Apple for now, and Nvidia a few weeks ago, seem to have mollified and placated the administration in Washington. Most importantly on their ongoing tech tariffs, especially on chips and devices. Both CEOs Tim Cook and Jensen Huang have done what they had to do to personally manage their companies through these geopolitical and ‘just political’ waters. These depths in tech are now more politically treacherous than in any prior tech waves. And it’s not clear that getting exemptions for their particular companies is going to really help long-term. Especially given the enormous, and complex ecosystems their businesses have built and rely on. To do the tech magic at global scale that they’ve achieved for investors to date. More here and here

   

Other AI Readings for weekend:

1. Intel’s Washington CEO Lip-Bu Tan challenge driven by US’s China Geopolitical Fears. Broader context here.

   

2. Tesla dismantling Dojo AI Chips effort as it ramps new AI chips with Samsung. More here.

   

(Additional Note: Doing a new podcast series on AI from a Gen Z and later perspective called AI Ramblings. Now 15 weekly Episodes and counting. More with the latest AI Ramblings Episode here, on AI issues of the day. As well as our latest ‘Reads’ and ‘Obsessions’ of the Week). Co-hosted with my nephew Neal Makwana. His latest ‘The New Terms of Service’ piece here)

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)