that will become a big issue for Google & other data aggregation companies - including those offering public AI services. A major part of the content available online will be generated by those AI engines from now on.
At first glance, the quality of the result seems surprisingly good. But it is not. The Tinderbox review is a very good example. Really, the text does not represent a gain in any way. Either the text just repeats stereotypes that can be found everywhere or the statements are flat and meaningless. There is practically no gain in knowledge with this kind of text. The opposite is the case. The masses of such empty bubbles make really good texts hard to find. This is not only true for us personally, but especially for Google & Co. and for the AI engines themselves. How are these machines supposed to build their own ānetwork of trustā and thus avoid that nonsense or such non-content will become the basis for the next generation of generated texts in the future? The problem will drag these machines into the abyss. Avoiding the problem would have required appropriate concepts and a lot of coordination between the players. Instead of tackling this, the hype of the hour was relied on - according to the motto: what do I care about tomorrow.
Good points. Now I know that the website canāt be trusted, I can avoid it in future, but of course I probably wonāt know in most cases. Weāve always had to verify sources, but this isnāt going to make it easier, and Iām not sure thereās any great benefit to readers to offset that.
Then make that clear on the web pages themselves, particularly when the text states explicitly that you wrote the review?
Oh itās simple just go to the docuchat.io page , create an account and add your files. Take like 3 mins to create a bot.
See Itās an experiment to test Google organic SEO. As for my readers, they are intelligent enough to know that it reads flat.
Honestly I havenāt even read the review. Still halfway.
Well, itās up to you, but to me that sounds like the perfect approach to lose credibility for a subscription website plugging a newsletter and book.
My main thought at the moment is to remove the Creative commons licenicing and replacing with one requiring explict acknowledgement of LLM use, one that doesnāt regard text like this, as proper disclosure:
This review is based solely on my research on acrobatfaq.com.
No mention of the AI/LLM aspect.
Regardless, the article is comically bad/wrong in parts and the images are out of date and sourced from elsewhere. Tinderbox v7 came out in 2017! As a user, why not take a screen grab from the current app, surely thatās no too much effort? It doesnāt reflect well the accuracy of the source material.
Today is one of the few times in the 28 years Iāve written FAQ content for the Web that Iāve had pause to regret sharing.
AI is interesting but it isnāt a labour-saving toy and use needs careful consideration and disclosure.
As I mentioned earlier. Itās an experiment to see if Google recognises AI content. My intent was to make some Tinderbox reviews available which I feel is lacking on the web.
I feel even if itās 70 per cent good, it is ready to ship. This is perhaps the reason Tinderbox though being the oldest tool around is overtaken by upstarts like Obsidian and roam research or Evernote.
The site is 3 weeks old, and the subscription is dummy. Haha. Seems it does look professional, I am working on it however, no product-market fit yet, the entire site is a AI content experiment. I have a day Job otherwise.
Youāll notice that Iām the only person responsible for the development of Tinderbox, and my participation on this topic has been very sparing! So, donāt blame other people for Tinderbox being āovertaken by upstartsā.
As for your list of those who you think have overtaken us: Iām not sure I agree! (Evernote?)
You know the images arenāt right, but my audience is someone who has never heard about Tinderbox. This is about getting their feet getting wet, this is my second review in the last 15 days.
I am trying to see how my community and readers react.
None of them are public companies so one never knows which apps is the most popular.
In the community I run, nobody has heard of tinderbox. All are notion, roam, obsidian users. Evernote has lost its prime.
Fair enough! I agree: we were here before Evernote, and we will be here after Evernote.
I do appreciate you work to spread the word to people who can benefit from Tinderbox.
I just read this carefully, I wasnāt aware of that aTbref could be downloaded as a .tbx file, could you please let me know, how to download the same?
Try reading the aTbref home page - you might find it more useful tham the LLMs:
Tip: the TBX has always been available for the 18 years aTbRefās been around. Read the website for more detail.
Expectation management. aTbref is my own volunteer initiative done in my free time, which means some prioritisation and I get to choose it. In order of up-to-date-ness:
- The website. This is often edited near daily, usually in response to conversations here in the form. Edits are normally in the form of clarifications, addition of new insights and correction of typos
- The TBX available for download from the website. Updated when I get around to it, i.e. if Iāve time to spare. Remember the website is the prime record.
- PDFs. Occasionally, usually for significant releases as opposed to every public release (so likely out of date for LLM use).
My tip - website is always the pace to start and donāt overlook the live site map as a quasi-index. Web search engines are, consistently over the decades, pretty rubbish at accurately spidering documentation sitesāand no better for the likes of Adobe of Microsoft products. That suggests those coding web search indexes just donāt really care about documentation as a source of information .
I honestly donāt know where the home page is? Why would you screenshot and not post a link also? 
Why donāt you try atbref.com
I should have guessed, thanks @archurhh . I just find the design so pre 2000 of the resource. No offence meant.
Dear @mwra , not able to download the .tbx file, getting a warning. Check screenshot below.
Hereās an idea:
-
Download the Tinderbox document. Use Safari if Chrome is complaining, or fix Chromeās security settings.
-
Run out a full HTML export.
-
Youāll see that thereās a single css stylesheet for screen formatting. Itās quite straightforward!
-
Modify the design by revising this stylesheet. Be as radical as you like!
-
Show us new redesign.
Itās Mark Andersonās call, but Iām confident heād be receptive to good ideas ā and CSS makes it fairly easy to allow people to skin sites like aTbRef if people cannot agree.
Amen. Indeed, the sharp-eyed will have spotted the aTbRef site go a CSS bump just recently due to some helpful, implementable, CSS code suggestions.
Please be aware, aTbRef is a reference, not a blog or such. It is designed to be consistent in layout and not visual design-heavy, and to even work for those reading it in a phone. Not is it designed for monetisation as many blogs are. The only real visual style is the yellow bar that old hands will recognise s the style of old Mac Help files. Still, Iād rather spend the free time I can devote to the project on useful content. so my apologies to all if itās not āstylishā.
I have zero interest in making it look like TikTok or Pinterest.
But, that doesnāt mean what you think it does when written like that. In truth, actually, I am offended, as I put a lot of effort into aTbRef and running the forums: Iām sad if it seems so sub-standard. But weāre a tolerant bunch here and Iāll let this slide as a misunderstanding. Iām sure weāre not here for an argument 