3 ms·

Corporations have been a form of AI Now together with actual programmable AI, they will capture ALL of our public institutions. Because nearly all our systems

by EGreg 3y ago

Corporations have been a form of AI Now together with actual programmable AI, they will capture ALL of our public institutions. Because nearly all our systems depend on the inefficiency of an attacker, to even function. For example if thousands of sleeper accounts join HN since 2021 and use ChatGPT to amass karma, then one day they can start to gradually shift opinion, constantly downvote and dispute anything that is of the opposite opinion, and also do reputational attacks on dissenting views, and have some accounts get them embroiled in pointless discussions, making them look bad. Everything from voting to politicians to the legal system can be corrupt using the current LLM technology deployed at scale!

fortran77 3y ago
You don't need ChatGPT to get "karma" on HN. Most people here are so banal and shallow that 50 lines of well-crafted Perl will get all the upvotes.
- EGreg 3y ago
  Alright let’s see As a Large Language Model, I can run do in 2 seconds what takes you 2 hours… bitch! (Is something I would never say.) Also, as of my knowledge cutoff date in September 2021, since there is no official public API for Hacker News, I can demonstrate how you can use the unofficial "Hacker News API" by Firebase to fetch comments and analyze them using Perl. Please note that this example uses a third-party API, and the availability or functionality of this API might change over time. Always review and comply with the API's terms of use and documentation before using it in your application. To perform HTTP requests in Perl, we'll use the LWP::UserAgent module, and for parsing JSON responses, we'll use the JSON module. If you don't have these modules installed, you can install them using CPAN. Below is an example Perl script that fetches the top 10 comments from the Hacker News API and analyzes them by calculating the average comment length, identifying the most common words, and highlighting comments containing specific keywords: perl #!/usr/bin/perl use strict; use warnings; use LWP::UserAgent; use JSON; # Function to fetch JSON data from a URL sub fetch_json { my ($url) = @_; my $ua = LWP::UserAgent->new; my $response = $ua->get($url); die "Failed to fetch data: " . $response->status_line unless $response->is_success; return $response->decoded_content; } # Fetch top stories from Hacker News API my $top_stories_url = 'https://hacker-news.firebaseio.com/v0/topstories.json'; my $top_stories_data = fetch_json($top_stories_url); my @top_stories = decode_json($top_stories_data)->@*; # Fetch comments from the top 10 stories my $comments_limit = 10; my @comments; foreach my $story_id (@top_stories[0..$comments_limit-1]) { my $story_url = "https://hacker-news.firebaseio.com/v0/item/$story_id.json"; my $story_data = fetch_json($story_url); if ($story_data->{type} eq 'story' && exists $story_data->{kids}) { foreach my $comment_id (@{ $story_data->{kids} }) { my $comment_url = "https://hacker-news.firebaseio.com/v0/item/$comment_id.json"; my $comment_data = fetch_json($comment_url); push @comments, $comment_data->{text} if exists $comment_data->{text}; } } } # Calculate average comment length my $total_comments = scalar @comments; my $total_length = 0; foreach my $comment (@comments) { $total_length += length($comment); } my $average_length = $total_length / $total_comments; print "Average comment length: $average_length\n"; # Count occurrences of words in comments my %word_count; foreach my $comment (@comments) { my @words = split(/\W+/, $comment); # Split by non-word characters (e.g., space, punctuation) foreach my $word (@words) { next unless $word =~ /\w/; # Skip empty strings $word_count{lc($word)}++; } } # Sort words by frequency and print the top 5 most common words my @sorted_words = sort { $word_count{$b} <=> $word_count{$a} } keys %word_count; print "Most common words in comments:\n"; for my $i (0..4) { print "$sorted_words[$i]: $word_count{$sorted_words[$i] }\n"; } # Define keywords to search for in comments my @keywords = qw(perl hacker news code); # Find and print comments containing the specified keywords print "Comments containing specific keywords:\n"; foreach my $comment (@comments) { if (grep { $comment =~ /\b\Q$_\E\b/i } @keywords) { print "$comment\n\n"; } } In this script, we first fetch the top stories from the Hacker News API, and then for each story, we fetch the associated comments. The comments are then used to calculate the average comment length, identify the most common words, and display comments containing specific keywords. Remember to review the terms of use and documentation of the API you're using before deploying this script in a real-world scenario.

by EGreg 3y ago

fortran77 3y ago

You don't need ChatGPT to get "karma" on HN. Most people here are so banal and shallow that 50 lines of well-crafted Perl will get all the upvotes.

EGreg 3y ago
Alright let’s see As a Large Language Model, I can run do in 2 seconds what takes you 2 hours… bitch! (Is something I would never say.) Also, as of my knowledge cutoff date in September 2021, since there is no official public API for Hacker News, I can demonstrate how you can use the unofficial "Hacker News API" by Firebase to fetch comments and analyze them using Perl. Please note that this example uses a third-party API, and the availability or functionality of this API might change over time. Always review and comply with the API's terms of use and documentation before using it in your application. To perform HTTP requests in Perl, we'll use the LWP::UserAgent module, and for parsing JSON responses, we'll use the JSON module. If you don't have these modules installed, you can install them using CPAN. Below is an example Perl script that fetches the top 10 comments from the Hacker News API and analyzes them by calculating the average comment length, identifying the most common words, and highlighting comments containing specific keywords: perl #!/usr/bin/perl use strict; use warnings; use LWP::UserAgent; use JSON; # Function to fetch JSON data from a URL sub fetch_json { my ($url) = @_; my $ua = LWP::UserAgent->new; my $response = $ua->get($url); die "Failed to fetch data: " . $response->status_line unless $response->is_success; return $response->decoded_content; } # Fetch top stories from Hacker News API my $top_stories_url = 'https://hacker-news.firebaseio.com/v0/topstories.json'; my $top_stories_data = fetch_json($top_stories_url); my @top_stories = decode_json($top_stories_data)->@*; # Fetch comments from the top 10 stories my $comments_limit = 10; my @comments; foreach my $story_id (@top_stories[0..$comments_limit-1]) { my $story_url = "https://hacker-news.firebaseio.com/v0/item/$story_id.json"; my $story_data = fetch_json($story_url); if ($story_data->{type} eq 'story' && exists $story_data->{kids}) { foreach my $comment_id (@{ $story_data->{kids} }) { my $comment_url = "https://hacker-news.firebaseio.com/v0/item/$comment_id.json"; my $comment_data = fetch_json($comment_url); push @comments, $comment_data->{text} if exists $comment_data->{text}; } } } # Calculate average comment length my $total_comments = scalar @comments; my $total_length = 0; foreach my $comment (@comments) { $total_length += length($comment); } my $average_length = $total_length / $total_comments; print "Average comment length: $average_length\n"; # Count occurrences of words in comments my %word_count; foreach my $comment (@comments) { my @words = split(/\W+/, $comment); # Split by non-word characters (e.g., space, punctuation) foreach my $word (@words) { next unless $word =~ /\w/; # Skip empty strings $word_count{lc($word)}++; } } # Sort words by frequency and print the top 5 most common words my @sorted_words = sort { $word_count{$b} <=> $word_count{$a} } keys %word_count; print "Most common words in comments:\n"; for my $i (0..4) { print "$sorted_words[$i]: $word_count{$sorted_words[$i] }\n"; } # Define keywords to search for in comments my @keywords = qw(perl hacker news code); # Find and print comments containing the specified keywords print "Comments containing specific keywords:\n"; foreach my $comment (@comments) { if (grep { $comment =~ /\b\Q$_\E\b/i } @keywords) { print "$comment\n\n"; } } In this script, we first fetch the top stories from the Hacker News API, and then for each story, we fetch the associated comments. The comments are then used to calculate the average comment length, identify the most common words, and display comments containing specific keywords. Remember to review the terms of use and documentation of the API you're using before deploying this script in a real-world scenario.