A lot of AIs seem to be focussed on how to create. new things, the interesting thing that Readwise is up to is that they are focussed on using AI to better understand what is already there ….. I think that is an interesting problem to solve.

Ghost Reader, which is definitely NOT Ghost Rider IS definitely interesting.

I have been playing round a little with it and decided that the best way of describing what it does is with a real world example.

I do like reading Napkin Math - but don’t always have the time.

Take a piece that came out this past week. At over 3,000 words it’s a 12 minute read - and THAT is if you get it first time round. I don’t always get it - and on a topic like this - I knew it would be a long job (I am slower than the average bear).

It was written by Evan Armstrong who summarized the article thus;

“Over the last few months, I’ve been obsessed with AI—chatting with AI researchers in London, talking product development with startups in Argentina, debating philosophy with AI safety groups in San Francisco.”

Sounds good, So here’s what I did.

Evan’s original piece on the site is here.

I dropped it in to Readwise and then applied Ghost Reader, which took about 20 seconds to come up with the summary. (My bold and line breaks added to help you understand how it broke down the article and make it easier to compare to the original.) AND - if you read the article you will see that my choice of the use case might not exactly be random since Evan actually quotes Readwise Daniel Doyon in item 4 - META!

Anyway, here’s the Ghost Reader summary …

“The text discusses six micro-theories about how artificial intelligence (AI) will change human society.
The first theory is that AI will make it easier for people to create new things.
The second theory is that companies will compete more on sales and marketing than on the quality of their AI models.
The third theory is that AI will make it easier for people to find and consume custom-generated content.
The fourth theory is that AI will amplify existing power law dynamics in the digital media world.
The fifth theory is that the most successful AI companies will be those that use AI to make their products more delightful.
The sixth and final theory is that AI will enable entirely new modalities of digital interactions.”

Compare Here are my one line extracts coped out of the article

  1. Fine-tuned models win battles, foundational models win wars

  2. Long-term model differentiation comes from data-generating use cases

  3. Open source makes AI startups into consulting shops, not SaaS companies

  4. Most endpoints compete on GTM, not AI

  5. AI will not disrupt the creator economy, it will only amplify existing power law dynamics

  6. Invisible AI will be the most valuable deployment of AI

For a 20 second ‘read, analyze, write’ a 3,000 word article - not a bad start. Not sure I think that it got to all the salient points - I wonder what Daniel and/or Evan think … but remember - as the article itself points out MOST AIs appearing are focussed on how to create … the interesting thing that Readwise is up to is that they are focussed on using AI to better understand what is already there ….. and they are only just now getting going.

I think that is an interesting problem to solve.

What do you think?