Scrivener Manual Project

AI isn’t really required; Pixelmator had it more than a year ago. Affinity also used Lanczos when upscaling. There are also web sites that perform upscaling on-line. Interestingly, there is one that changes the metadata so an image reports 300 dpi when it is in fact less. Not sure what the point of that is.

1 Like

They call it ‘Machine Learning’, which sounds much like ‘AI’ to me.

Regardless of what you want to call it, it does what I said it does, very well.

I’m not buying into either the AI enthusiasm or the hysteria. I tried using it for simple searches and it made egregious errors. For coding, it’s OK, at least as far as I can tell. There has been a significant drop off in its quality over the past six months and no one knows why. It’s now become a sales gimmick: add the term “AI” to your product copy and get more sales. “Jello enhanced by AI-calculated flavors in the new family pack!”

1 Like

I had a LaTeX question that the savants at Stackexchange couldn’t answer. On the third try, ChatGPT came up with the answer.

However, it confused Clarence Darrow with William Scott Stewart (“Chicago”) and had to be corrected.

I wouldn’t call anything I’ve said ‘hysteria’.

AI has a place and will improve over time.

I’ve tried it out and while, yes, it can make some serious stuff-ups, it can be a useful tool with prospects into the future.

It is no worse than the wonder that is Google. ‘I saw it on Google’ can sometimes be no more accurate than ‘I made it up’.

I vividly remember ‘bendgate’. ‘It’s all over Google’, when several hundred images of bent iPhone 6s were analyzed and found to be 6 or 7 phones with photos of different angles and some judicious cropping.

You didn’t indeed.
Many people feel however, that AI means the end of the world while the other extreme thinks its useless. I think the truth will be somewhere in between.

1 Like

Congratulations on your printed manual(s).

To pick up on other comments, when Pixelmator first came out, I bought it but rapidly discovered I couldn’t get on with it—only one of a number of photo/image manipulation apps I’ve bought only to abandon quickly. I still use Graphic Converter for most things. However, around 2016 or so, when I had to find a way to enlarge a ca 400px scan to 2400px while retaining detail, I found Topaz Labs Gigapixel AI which did a brilliant job. So now, I shoot in Raw develop in Topaz Photo AI and then, if necessary, do things like cropping to the size I want in Graphic converter. Topaz Labs are not cheap, but Photo AI, Sharpen AI, and Video AI are brilliant.

On another track, for those who want to set things up for themselves, in the past I used Cheap Impostor to do the imposition for stuff I produced for my students in China. I will probably use it again. I used it to convert portrait A4 pdfs into properly imposed 2-up landscape A4s to 16-page signatures, with it handling the “margin creep”, though I never bothered to guillotine the page edges of my own printouts.