Happy New Year everyone and welcome to 2026!
As we’re only 4 days into the year, there isn’t a huge amount of generative AI news to report, but I thought it worth sharing a few of the annual wrap ups (I’m sure more will be published this coming week too) as well as some of the interesting things I’ve seen over the festive break.
Enjoy!
2025: The year in LLMs
This is a great wrap up on 2025 from Simon Willison, one of my favourite AI commentators. Highlights include:
How ChatGPT is Changing How We Speak
I’ve been fascinated with language for a while now and have shared news reports previously on the rise of words like ‘delve’ as well as a whole new vocabulary we’re having to invent around AI like word of the year ‘slop’.
What’s important for people to be aware of though is that there’s a very strong link between language and beliefs. This is a great video that explains this better than I can, but essentially we are now in a place where generative AI models are trained on a representation of reality (from data on the internet), but these AI models are now being used by so many people that they’re influencing reality (hence why people are adopting ‘AI-speak’ in everyday language).
Worth a watch.
2025: The year of the recap
I thought it would be fun to share some of the stats and commentary from my Year with ChatGPT, which summarised everything I’ve been doing with it throughout 2025.
To put this in context, I was a heavy Claude user in the first half of the year, but switched to ChatGPT being my daily driver in the summer as I was finding it more capable and was getting frustrated by the limits on the Claude Pro plan. Despite that, turns out I have used ChatGPT a lot (who knew?!). Here are some of the interesting things I found in my 2025 wrap up:
I’m in the top 3% of messages sent and the first 20% of users to try out new features, which is probably a good thing given my job!
Don’t tell my boss (🤓) but apparently I’m most likely to automate my job, but then at least I’ll tell everyone about it and how I did it!
My archetype is ‘The Strategist"‘, which I share with 3.6% of users. This is defined as using ChatGPT to synthesise information, explore new concepts, and guide direction which is a pretty good summary of what I use it for.
Just a bit of fun mostly, but will be interesting to see how this compares to the 2026 wrap up in 12 months time!
Web 4.0
With the rise of the chat interface, I’m convinced that we’ll finally see voice become a primary mode for user interactions. We haven’t seen this really take off with any of the smart speakers, home assistants, or mobile assistants and it’s been a very slow burn over the last 10+ years. But, I think this will change in the next few years, so it was unsurprising (to me at least!) to see that OpenAI is betting big on audio as Silicon Valley declares war on screens.
AI Ethics News
Bernie Sanders criticizes AI as ‘the most consequential technology in humanity’
Elon Musk’s Grok AI generates images of ‘minors in minimal clothing’
AI boom adds more than half a trillion dollars to wealth of US tech barons in 2025
Nvidia insists it isn’t Enron, but its AI deals are testing investor faith
OpenAI’s ChatGPT ‘Wrapped’ Is a Good Reminder to Check Your Privacy Settings
Long Reads
Understanding AI - 17 predictions for AI in 2026
Peter Steinberger - Shipping at Inference-Speed
“The future is already here, it’s just not evenly distributed.“
William Gibson