HN user

mometsi

780 karma
Posts1
Comments123
View on HN
OCR4all 1 year ago

How is this different from tesseract and friends?

The workflow is for digitizing historical printed documents. Think conserving old announcements in blackletter typesetting, not extracting info from typewritten business documents.

Seems like a reasonable alternative to Proton for multiple domains: email that costs money instead of being haunted by AI.

That logo is creepy though, and implies exactly the opposite.

Interesting features:

  - No 1 or 0, just use l and O
  - Page width is measured using decimal-divided inches
  - That lowercase g glyph!

It might be that exploring new physics might be best pursued by inhuman alphago-style setups.

A sharp, domain-specific Feynman-flavored LLM would be broadly useful and still worth bothering with.

Poets' Odd Jobs 2 years ago
  I have freed  
  the symbols  
  from your  
  clumsy lines

  which  
  the linter probably  
  considers  
  in error  
    
  Indulge me  
  it's more readable  
  neat functions  
  in neat form

I wish this article featured actual data or a simulation instead of these napkin drawings. Does a set of uniform or gaussian inputs to this interaction model produce power-law output distribution? Is it just because some nodes on the graph are highly connected?