Search Ideas

Unlike #3424, however, having a set amount per criticisms means there’s zero incentive for anyone to submit more criticisms, whereas divvying up the amount among criticisms means the incentive is gradually reduced, and it’s up to people to decide for themselves whether the reduction is still worth contributing.

Rather than set a fixed amount for each unproblematic criticism (#3421), the ceiling could be divided among all criticisms equally.

That could be a good thing in that people won’t completely overwhelm OP with criticisms.

But that means that additional criticisms don’t get any payout.

The initiator of the bounty could choose a ceiling for the total they are willing to spend. They could additionally specify the amount per unproblematic criticism.

For example, a user would indicate that they are willing to spend a total of $100 at $10 per criticism.

Yes, people could just start bounties on criticisms.

Idea: voice spaces, like Twitter spaces, except an AI generates a transcript and automatically turns it into a discussion tree, with criticism chains and all.

Fixed as of bd7c1b6.

There’s an encoding bug affecting title previews.

Done, see #3413.

Benjamin suggests making it clearer that you can use Veritula by yourself.

I have implemented 1-4. Give it a try. I think 5 is out of scope for now but I may revisit it at some point. If auto-closing asterisks are a problem at the start of a line (when making lists), use a hyphen instead.

On second thought, implementing a proper text editor would take more work than I initially realized, and is far beyond the scope of what Benjamin is requesting anyway. I can revisit this idea later.

I can take this opportunity to replace manual markdown with a proper text editor. Then there’s no need for autopaired brackets.

The editor will need to support:

• Automatic links to ideas like #123
• Links to @mentions like @dennis-hackethal*
• Safe link formatting
• Disabling of turbo links
• Namespaced footnotes
• Custom blockquote format
• Protection against XSS
• Retention of formatting when pasting

I can take this opportunity to replace manual markdown with a proper text editor. Then there’s no need for autopaired brackets.

The editor will need to support:

• Automatic links to ideas like #123
• Links to @mentions like @dennis-hackethal*
• Safe link formatting
• Disabling of turbo links
• Namespaced footnotes
• Custom blockquote format
• Protection against XSS

I’ve asked Gemini to explain it:

1. Auto-Closure (Insertion State)

When the user inputs an opening delimiter, the system immediately injects the corresponding closing delimiter and places the caret (cursor) between them.

Input: (

Buffer State: (|)

Logic: insert(opening_char) + insert(closing_char) + move_caret(-1)

2. Type-Through (Escape State)

If the caret is positioned immediately before a closing delimiter that was autopaired, and the user types that specific closing delimiter, the system suppresses the character insertion and instead advances the caret.

Context: [text|]

Input: ]

Buffer State: [text]| (Not [text]])

Logic: if (next_char == input_char) { move_caret(+1); prevent_default(); }

3. Atomic Deletion (Regression State)

If the caret is between an empty pair of delimiters, a backspace event deletes both the opening and closing characters simultaneously, returning the buffer to the pre-insertion state.

Context: (|)

Input: Backspace

Buffer State: |

Logic: if (prev_char == open && next_char == close) { delete_range(caret-1, caret+1); }

4. Selection Wrapping (Transformation State)

If a text range is selected (highlighted) and an opening delimiter is typed, the system wraps the selection rather than replacing it.

Context: |selected_text|

Input: [[

Buffer State: [[selected_text]]

Logic: surround_selection(input_pair)

5. Markdown-Specific Heuristics

Obsidian applies context-aware logic for Markdown syntax (e.g., * or _). It often checks word boundaries to determine if the user intends to bold/italicize or use a bullet point.

Context (Start of line): | + * + Space -> Bullet list (autopair disabled/consumed by formatting).

Context (Middle of line): word | + * -> word *|* (autopair enabled for italics).

As of 9087189, the footer automatically hides and shows based on scrolling behavior.

Try it out and let me know if this doesn’t help.

As of c08f508, the footer automatically hides and shows based on scrolling behavior.

Try it out and let me know if this doesn’t help.

I’ve asked Gemini to explain it:

1. Auto-Closure (Insertion State)

When the user inputs an opening delimiter, the system immediately injects the corresponding closing delimiter and places the caret (cursor) between them.

Input: (

Buffer State: (|)

Logic: insert(openingchar) + insert(closingchar) + move_caret(-1)

2. Type-Through (Escape State)

If the caret is positioned immediately before a closing delimiter that was autopaired, and the user types that specific closing delimiter, the system suppresses the character insertion and instead advances the caret.

Context: [text|]

Input: ]

Buffer State: [text]| (Not [text]])

Logic: if (nextchar == inputchar) { movecaret(+1); preventdefault(); }

3. Atomic Deletion (Regression State)

If the caret is between an empty pair of delimiters, a backspace event deletes both the opening and closing characters simultaneously, returning the buffer to the pre-insertion state.

Context: (|)

Input: Backspace

Buffer State: |

Logic: if (prevchar == open && nextchar == close) { delete_range(caret-1, caret+1); }

4. Selection Wrapping (Transformation State)

If a text range is selected (highlighted) and an opening delimiter is typed, the system wraps the selection rather than replacing it.

Context: |selected_text|

Input: [[

Buffer State: [[selected_text]]

Logic: surroundselection(inputpair)

5. Markdown-Specific Heuristics

Obsidian applies context-aware logic for Markdown syntax (e.g., * or _). It often checks word boundaries to determine if the user intends to bold/italicize or use a bullet point.

Context (Start of line): | + * + Space -> Bullet list (autopair disabled/consumed by formatting).

Context (Middle of line): word | + * -> word | (autopair enabled for italics).

I haven’t used Obsidian, so I don’t understand what you are requesting. Is it that, whenever you open a bracket, you want the closing bracket to appear automatically?

It would be nice if I could collapse the 'submit top-level idea' form. It currently takes up a third of my screen when I scroll on PC.

Obsidian autopairs markdown syntax and brackets. I like it a lot and would like Veritula to have something similar!

@dennis-hackethal* see the revision chain on #3164. Revision 5 improved the content but I accidentally removed valuable comments. Revision 6 (a revision of revision 4) brought back the comments but I failed to include the content improvement in revision 5. I then made revision 7 to have both the comments and the improved content.

Maybe it should be possible to amend which comments apply to an idea without needing to make a whole new revision. This could behave weirdly in some edge cases, but it’s food for thought. If you think the way it currently works is going to be best, that seems fine to me.

This is done as of 9b5788c but it’s still free for now. Will make it a paid feature after some more testing and polishing.

The Effective Altruism forum has an interesting way to react to posts.

There’s an ‘Agree’ button and a ‘Disagree’ button. Those are apparently anonymous. Then separately, there’s a button to ‘Add a reaction’ of either ‘Heart’, ‘Helpful’, ‘Insightful’, ‘Changed my mind’, or ‘Made me laugh’. And those are apparently not anonymous.

I wonder why they chose to make some reactions anonymous but not others. I don’t think I’d want a ‘Heart’ or ‘Made me laugh’ button, they seem too social-network-y. Also, ‘Heart’ seems like a duplicate of ‘Agree’. But ‘Insightful’ and ‘Changed my mind’ seem epistemologically relevant. Maybe ‘Helpful’, too.

If I did decide to go with ‘Agree’ and ‘Disagree’ buttons, I wouldn’t make them anonymous, though.