5 Comments

A little off topic, but I've ran into a lot of mathematicians, linguists, psychologists etc. etc. who seem palpably disappointed- even a little angry- that machine learning is not theoretically novel- that it's just the equivalent of banging an enormous amounts of computation and data together. These people are disappointed because they wanted big AI advances to flow from, or perhaps lead to, deep secrets about the human mind, the nature of thought etc.

To me, the effectiveness of just ramming data and computational power over ****relatively**** simple models until they work is a testament to the power of evolution.

Expand full comment
author

I take your point. I'm more annoyed that machine learning is being sold as something new, when it's just the same old stuff with more computer power. That means, in turn, that any statistical problem that can't be dealt with by increasing the sample size is going to be just as prevalent with "machine learning" as it would have been with calculations done by hand, or on an adding machine.

Expand full comment

Thank god, I feel like I'm living in the land that invented invisible clothes. As an a former programmer with an interest in statistics, linear regression and predictive models being the main interest, I'm continually aghast at the inflated claims of the unprecedented power of so called AI and ML. It's simplify another IT industry hype cycle the result of which there will be marginal change for vast majority of the world's population and millions transferred to overrated IT firms. Hopefully I can soon retire from what has become a symbiosis of snake oil and ignorance. It's a pity, was programming was once a really satisfying craft, I pity the grads today.

Expand full comment

My experience of Chat, it works best when you know very little about the question you are prompting. If you have some experience/ expertise with the question, then Chat does not provide much more.

Expand full comment

John you are mainly concerned with predictive tasks but I use AI for other purposes that I want to mention.

I use Chat GPT and the visual image programs Midjourney, Dall-E and Stable Diffusion all the time. Your point is well taken - these are language-based learning procedures )sometimes with image prompts) that look for associations and which do not use statistics and do not understand the world. Nevertheless I find them very useful and, for the image programs, a lot of fun - an experience like going to the movies.

Chat GPT gives wrong answers as it always acknowledges it can. I asked it the other day for key economic reforms for the Australian economy and it suggested a goods and services tax and reform of tariffs. The GST was introduced in 2000 and average tariffs are now only around 1% so wrong advice. But I have also used it for projects where it gave very useful outputs - I wanted a survey of the ethics of Thomas Aquinas, an exposition on problems of "conceptual photography" for a talk I had to give and the problems of AI image making. nIt was good at looking at itself and fessing up to limitations. Itgave very useful outputs. In many cases it replicated ideas I had been thinking about for some time - although it often exposited things better than I did.

Obviously you need to check outputs - they are suggestive but can be wrong. They bprovide a starting point.

On image making Midjourney, in particular, has had spectacular success. Boris Eldagsen "won" one of the world's most prestigious photographic competitions (the cSony competition) with an AI image. Once he "won" the competition he told the judges it was AI and he was disqualified but he wanted to make the point.

The non-photographic image you provide in your text is attractive to me.

Now basic photo editing tools (such as Photoshop) include AI components for removing or inserting parts of an image and for changing dimensions. The change has occurred.

I am also involved in preparing audio visual displays and use AI to spin out the draft of a plot line connecting maybe 50 still images and video and it works well. I can even use a language software to present the dialogue.

OK so I have wandered off your track but simply say that for many purposive tasks and for enjoyable experiences, AI is useful.

Expand full comment