$ links -dump https://news.ycombinator.com/item?id=17205865 | grep -Eio '(php|python|react|javascript|java|objective-c|objective|object-c|object|haskell|rails|mongo|lamp|docker|aws|devops|android|ios|django|ruby|rails|react|swift|node|docker)\b' | tr A-Z a-z| sort| uniq -c | sort -rn
  87 react
  63 python
  42 javascript
  39 aws
  33 devops
  29 rails
  28 ruby
  28 node
  23 android
  22 java
  19 ios
  19 docker
  15 django
  12 haskell
   8 php
   7 swift
   5 mongo
   2 object
   1 objective-c
   1 objective

Used links to get just the text and not the HTML markup. Used grep -O to emit just the part of the regex which matches so we can feed it all into a pipe that emits a single histogram. By ordering the pattern we can match "javascript" before "java" and not need to add the space or "," after "java" to exclude javascript (you could also use "\b" with grep -E).Case-insensitve and case-folded. Added a few other keywords.

Since grep only emits the first match on a given line, this won't count properly if multiple keywords appear on one line.

Why did you search for "mean"?

Edit: here's another approach that splits the output into words first so that we can count multiple keywords on one line, along with a couple tweaks to the pattern to eliminate some false positives:

  $ links -dump https://news.ycombinator.com/item?id=17205865 | tr -s '[[:punct:][:space:]]' '\n' | tr A-Z a-z | grep -E '\b(php|python|react|java|object|haskell|rails|mongo|lamp|docker|aws|devops|android|ios\b|django|ruby|rails|react|swift|node|docker)' | sort | uniq -c | sort -rn
  87 react
  63 python
  42 javascript
  38 aws
  33 devops
  29 rails
  28 ruby
  27 node
  23 android
  21 java
  19 docker
  18 ios
  15 django
  12 haskell
   8 php
   7 swift
   6 nodejs
   5 reactjs
   5 mongodb
   5 mongo
   2 objective
   2 object
   1 reactnative
   1 reactive

Another issue: this only grabs the first page of comments.