Showing 1 - 10 of 20
Caveat emptor is undefeated in online shopping
Life, James Hein, Published on 11/03/2026
» It is becoming more common to buy things online. The majority of my shopping, not counting groceries, is now done that way. In the past I've warned about prices that are too good to be true, like a 4TB thumb drive for a few dollars from sites like Temu and AliExpress. There is now a kind of middle ground where the price could be correct and it's coming from, say, Amazon. Recently, even though I had some doubts, I bought a 5TB SSD drive from Amazon for around half of what I'd expect it to be. I did this knowing I can easily send things back to Amazon.
Android is now the new Apple
Life, James Hein, Published on 10/09/2025
» Some may be wondering why have I used an Android phone up to now compared to, say, Apple iPhone? In the beginning, it was for the following four reasons -- a headphone jack, a removable battery, the ability to insert an SD card for storage and the ability to load a program into the computer. That last part may sound a little strange but a 12GB device with 1TB of storage and a graphics unit built in is a computer now. The "program" loading here means you can put your own operating system on the device and install applications, bypassing the Play Store. So where are we now? No headphone jack, no SSD support, no removable battery and based on a recent announcement, no ability to "side-load" programs any more. In other words, Google phones have now or will soon be turned into iPhones.
LLMs are left-leaning models
Life, James Hein, Published on 07/05/2025
» A while back I wrote about the political bias in Large Language Models (LLMs). Since then the models have evolved and David Rozado has conducted more recent tests based on four of the popular political orientation tests. Using the Political Compass, Political Spectrum, Political Correctness and Eysenck tests, he worked with xAI Grok 3 beta, Google's Gemini 2.5 pro, Deepseek V3, OpenAI GPT 4.1 and Meta's Llama 4 Maverick. In all but one of the tests Grok 3 was closest to the centre, and on average was the clear leader. All the models were still located in the Left Libertarian quadrant, with Grok just sneaking into a more Conservative area with the Eysenck test. These tests are of course but one way to measure the political leanings of any LLM. Overall however, it does still indicate the left-leaning bias in all models tested so far. If you want to see more details, you can visit David Rozado's substack.
Rise of fake AI content
Life, James Hein, Published on 23/04/2025
» We have unofficially entered the time of the fake AI generated content. It's not perfect yet, but it's good enough to fool many people. I use a YouTube alternative, which means I have more time to watch videos across a wide range. In the last few weeks, I've noticed an increase in AI generated product that is fake. One standout is the Got Talent franchise where people have grafted AI generated acts with cuts from responses from the panel. They look great, like a woman transforming into a lion, but it's all fake. Even some of those cute animal videos are now fake. Why? More clicks so more income all based on fabricated videos.
What's the state of AI today?
Life, James Hein, Published on 28/08/2024
» Recently, I've talked with a few people about artificial intelligence and watched a few presentations. The gap between what some people think are AI capabilities and actual capabilities is a big one.
Can you trust search engines?
Life, James Hein, Published on 14/08/2024
» So how do conspiracy theories start and how does the internet and major search engine players contribute to them? As I type this, depending on where you are in the world, if you type "assassination attempt" in your search engine, and in particular one associated with Alphabet, the autofill options will have everything except "Trump" in the result set. You can get Kennedy, Hitler, Putin and George Wallace, but not the most prominent one so far this year that was one of those "where were you when you heard" events that some people saw in real time on their TV. The reason Alphabet offered for not giving the result was something along the lines that their policy is not to show political violence. You can of course find a plethora of political violence videos and examples from their search results, just not for this particular instance. Another example if you type "President Donald", the autofill adds Duck and Reagan but not Trump. Or if you Google Donald Trump you get a bunch of Kamala Harris results.
Headlines are disappointing
Life, James Hein, Published on 22/11/2023
» If you have been in any way involved in social media over the past few years, there will be words that immediately have you wondering if the headline is in any way real. My favourite recently was "SpaceX Launched REAL UFO In ISRAEL! HAMAS And Iran Shocked!" on YouTube.
The new normal of AI fakes
Life, James Hein, Published on 11/10/2023
» Following on from my last article, consider the following scenario. You've grabbed enough clear speech of someone to make a good AI model of them. You write up some text, pass it through the model and verify that the result sounds exactly the same as that individual. This is a little different from the previous example because it's a text to speech model, but essentially the same as using one voice to change to another. You now take a speech or interview from that person, change one word that will essentially change the context, and process this.
Microsoft stock booming on AI run
Life, James Hein, Published on 07/06/2023
» It's not a good time to be working for Microsoft, but it is a good time to be a shareholder or executive. Stock is up over 30%, net income is up and the CEO Satya Nadella got a nice 10% raise. Regular workers received no pay rise, or effectively a 5% pay cut due to inflation. Microsoft has rationalised it as pat and generic reasons like a "competitive environment" and the "global macroeconomic uncertainties". In reality, Microsoft is using the money to jump into the AI wave through a multibillion-dollar partnership with OpenAI.
YouTube ramps up ads, moderation
Life, James Hein, Published on 01/03/2023
» If this was a YouTube video you would have to sit through up to 30 seconds of ads before you could even start. YouTube seems to be stepping up its advertising while at the same time providing less service. I still use it because it has things I'm interested in, like Chinese martial arts series and info on music products I like. If I'd written this using ChatGPT you would not see some of the material because the trust and safety filters on the AI product have repeatedly been found to be biased towards the US political left in the content it will return. Some people associated with ChatGPT have acknowledged this but it remains to be seen if anything will change.