Time for a catch-up!

Well, I got distracted by various things last week including starting some more training on AI and so I never got around to publishing my newsletter, despite having drafted it, so this newsletter covers two weeks instead of one. However, part of what was happening last week will mean that I’m planning to gradually increase the frequency of my posts.

The pace of change continues unabated, with software updates and new products being released, and yet more significant investments being announced. Despite the huge sums being invested, the big players continue to offer more for less money - makes me wonder how their business models can be sustainable in the long run?

Anyway, let’s dive in and find out what’s been happening!

Big Sharks vs. small fish

These last two weeks have seen yet more announcements of large-scale investments by the “big sharks”, and they each seek a market dominant position.

At the “low-end” Dell was reported to be moving closer to finalising a contract to supply xAI with $5 billion USD of computing capacity, news that came out around the same time that xAI released Grok-3 (more on that later). Then Alibaba announced a $50 billion USD investment in infrastructure. This has mirrored a massive focus in China on embedding AI in businesses, with both Tencent and Baidu looking to utilise DeepSeek to boost their productivity. The support and backing provided by the Chinese government to encourage Chinese companies and local governments to utilise DeepSeek, indicates not only their strategic direction, but also hints at their suspected involvement in the development and promotion of the platform. All of that however, paled in the face of Apple’s announcement that it would be investing $500 billion USD in the US, primarily around infrastructure, including supporting the production of their products (iPhones, iPads, etc.) locally, no doubt in response to the recent tariffs announced by the Trump government.

However, a recent report raised concerns that an “investment bubble” was being created, referencing the (pre-Apple announcement) $340 billion USD invested in AI to date by Meta, Amazon, Microsoft and Google, and questioning whether in fact the market would go the same way as the first DotCom bubble.

The maths become ever more questionable given the recent cuts to charges by OpenAI and others in response to the arrival of DeepSeek. Having released their Deep Research tool only to Professional users a month ago, OpenAI have already started to allow other paid users access, albeit it with certain limits, whilst all LLMs have reduced their costs and/or allowed greater free access in order to counter the low rates of DeepSeek. OpenAI also released ChatGPT 4.5 to Pro users, with a cost of $75 USD per million tokens, compared to $2.50 USD per million tokens with ChatGPT 4. It will be interesting to see how long it is before that price drops significantly, and ChatGPT 4.5 is opened up to other users - I would give it no more than a month!

The battle over costs was not just in relation to the public user interface, as Google launched Gemini Code Assist, a free AI tool for developers with higher usage limits and deeper GitHub integration. It scans pull requests for errors, making debugging easier – making it a direct challenge to Microsoft’s paid GitHub Copilot service.

Legal repercussions and fewer jobs (part 2)?

On the legal front there were continued arguments regarding Intellectual Property Rights, with various parties bringing legal cases relating to LLMs using their materials to train their systems. This was complicated by the news that the new EU AI Act contains a potential loophole that was designed to protect individuals being sued for copyright infringement but may have the unintended outcome of protecting the “big sharks” from copying the work of others – let’s see whether amendments to the final legislation will happen.

Musicians also drew attention to their concerns with the release of a “silent” album, supported by many well-known names.

It wasn’t all doom and gloom though, with a number of reports highlighting how AI was speeding up medical research. In one case, a team used AI to independently research a topic, and it produced an answer within a few hours that not only matched but actually surpassed their own previous research on the same topic, that had taken them more than a decade to reach the same conclusions. The AI even identified some additional areas for future development that had not even been considered before.

The Chipotle Mexican Grill company (popular in the US) also revealed that using AI had helped them manage their seasonal recruitment of 20,000 staff per year more efficiently. The AI helped increase application completion rates from 50% to 85%, and reduced onboarding time for new staff from an average of 12 days to 4 days. This underlines the known benefits of AI in supporting repetitive tasks with clearly defined parameters.

More challenging however, was Wendy’s (US burger company) introduction of a Chatbot to take online and drive-through orders. The initial rollout has faced lots of issues and complaints from customers, but clearly not enough to deter Wendy’s, who intend to proceed with their rollout of the tool to their 6,000 outlets.

Regards

Tom Carter