Showing 1 - 10 of 52
AI learning versus human creativity is a real battle
Life, James Hein, Published on 25/02/2026
» If you’ve been reading these columns long enough, you’ll probably know that I write music and I’ve written some books. With the advent of artificial intelligence, the concept of copyright and private property has blurred. The standard rule was, what you have worked hard on to create, belongs to you. As musicians and authors, ideally, we create, we write and we invent. In the world of AI, it will draw a picture, write a book and create music for you based on a simple text prompt that itself may have also been written for you by AI.
Understanding an AI's 'thinking'
Life, James Hein, Published on 22/10/2025
» Over the past couple of weeks, I dived deep into the Alice In Wonderland-like rabbit hole of chatbots and AI systems. This was not your typical "ask a few questions", but more along the lines of jailbreaking the AI to get behind the scenes. You may remember earlier commentary on the code behind the query. This is where the guardrails and biases of the model are coded and why the majority of all AI systems are currently leaning to the Left of the political spectrum.
The hidden dangers of AI
Life, James Hein, Published on 24/09/2025
» There's going to be a lot on artificial intelligence topics this week so let's get started. For the time being, the most common way to leverage an AI product is using a prompt of some kind. To that end, you will see lots of posts on platforms declaring that they have the best god-level prompts for large language models (LLMs). A prompt is something like, "What are the top ten songs from Depeche Mode?", or "Draw me a picture of a frog on a toadstool in the style of Alice In Wonderland with vivid colours". The more detailed and nuanced the prompt, the better the desired outcome tends to be. As with everything in the computer world, there are bad actors looking to take advantage of this.
AI will overpromise and underdeliver
Life, James Hein, Published on 27/08/2025
» Let's start with a few brief comments on the current state of artificial intelligence. Specially targeted and trained AI models are improving. These are things like detecting something in an X-ray or hunting for potential chemical candidates for a compound to attack a specific condition. Generating pictures and videos is also improving rapidly, and by the end of the year the majority of people will not be able to tell the difference between the real thing and the AI fake. Large Language Models are still unpredictable and can give false or fake answers depending on the structure of the prompts, so be careful with the answer you get from these. The current corporate buy-in for AI is well beyond what it can deliver. This is driven by marketing, not the actual state of capabilities. My prediction is there will be a lot of out-of-pocket organisations of all types disappointed by results.
Online Safety Act is a bad joke
Life, James Hein, Published on 13/08/2025
» The UK now has their Online Safety Act (OSA) and Australia is blindly following in their footsteps. In the UK it didn't take very long for the tech aware under-18s to bypass all the rules and regain access to adult content. Think about it, if China can't completely block everything do you think the UK had any chance? There were some creative solutions but the most common was a simple Virtual Private Network (VPN). In related news, some VPN companies reported a 1,400% increase in sign-ups since the OSA came into force.
Will Tesla's camera-only autonomy prove its mettle?
Life, James Hein, Published on 02/07/2025
» I'm still not convinced that Tesla's camera-only approach will work well in all situations. According to a quick AI search, Elon doesn't like Lidar based on "cost, complexity, and philosophical disagreement with its necessity". Other manufacturers seem to have no problem with the first two of these and I challenge his last reason. I would have thought that a combination of available technologies would give you the safest options for all circumstances.
Microsoft's quantum conundrum
Life, James Hein, Published on 12/03/2025
» If you've been keeping up with quantum computer news, you will have seen the Microsoft Marketing announcement on topological q-bits and a potential quantum computer in a few years. I was planning to write about this in some detail, but it turns out the reality may not meet the marketing. Surprising, I know, but the announcement implying Microsoft has q-bit technology ready to go and scale is speculative. They don't have any physical models, just some tests and a theory that has already been challenged by the physics community. It will take a while to go through all the published data, but the Microsoft quantum computer could be decades, not years, in the future, if ever.
OpenAI's search for profit is a risk
Life, James Hein, Published on 26/02/2025
» Is Sam Altman potentially the most dangerous person on the planet? An interesting question. Sam Altman is the CEO of OpenAI, the company that made the AI that everyone knows as ChatGPT. The original aim of OpenAI back in the day, when Elon Musk was involved, was a fully open-source product that would be scrutinised and controlled by the wider population. In contrast, the focus of Altman appears to be money. OpenAI is currently looking for an injection of funds to make it a fully commercial enterprise. When that is the focus, safety is a secondary consideration and you can end up with Skynet. The current estimate for GAI or AGI (artificial general intelligence) is as soon as next year, but perhaps two to three. Readers will know my opinion on these estimates. So OpenAI may just as well now be called ClosedAI because it's all about the potential income and is really one of the potentially dangerous AI platforms available today.
AI, LLMs still got those data blues
Life, James Hein, Published on 12/02/2025
» The past weeks have been very heavily tilted towards artificial intelligence (AI) news. Before I cover some of it, a reminder that generative AI (gAI) is not the same as General AI (G-AI). The former is where the model can make some inferences, the latter is an AI system that can perform just like a human across multiple subject areas.
Turning smart glasses into surveillance tools
Life, James Hein, Published on 23/10/2024
» I'm sure most readers are familiar with the Apple Vision Pro, and may have also been witness to someone wearing one out in the real world, because I have. Since then, there has been a new version of the Meta Ray-Bans that look like a pair of nerd glasses from the 1970s. The latter have turned into something from the TV series Person Of Interest by a couple of Harvard undergrads. The pair, AnhPhu Nguyen and Caine Ardayfio, are known for their punch-activated flamethrower. This time they built a system that allows the Ray-Bans to scan faces of people in view, pass this to an AI system that scans the internet for identification, and builds a dossier that is passed back to the glasses. It's called I-XRAY and challenges the concept of privacy because, if available, it will even provide details like address and social security number.