eterps

While I appreciate the emergence of new AI coding tools, I've observed that many of them fail to offer significant improvements over Aider's existing capabilities (but I want to be proven wrong).

So this is actual sorting (i.e. the same amount of lines appear in the output)?

Otherwise it might be more or less the same as using https://github.com/TheR1D/shell_gpt ?

Isn't that a non-generic 'reasoning-model' instead of something that is reminiscent of the large language model based AIs we use today?

The question is, is it possible to make reasoning models generic and can they be combined with large language models effectively.

I understand your point, but it's not comparable:

Humans can suddenly "jump" cognitive levels to see higher-order patterns. Gödel seeing that mathematics could describe mathematics itself. This isn't combining existing patterns, but seeing entirely new levels of abstraction.

The human brain excels at taking complex systems and creating simpler mental models. Newton seeing planetary motion and falling apples as the same phenomenon. This compression isn't recombination - it's finding the hidden simplicity.

Recombination adds elements together. Insight often removes elements to reveal core principles. This requires understanding and reasoning.

> and wonder how an intelligent person can still think this, can be so absolute about it. What is "actual" reasoning here?

Large language models excel at processing and generating text, but they fundamentally operate on existing knowledge. Their creativity appears limited to recombining known information in novel ways, rather than generating truly original insights.

True reasoning capability would involve the ability to analyze complex situations and generate entirely new solutions, independent of existing patterns or combinations. This kind of deep reasoning ability seems to be beyond the scope of current language models, as it would require a fundamentally different approach—what we might call a reasoning model. Currently, it's unclear to me whether such models exist or if they could be effectively integrated with large language models.

Can you elaborate, or give some examples? I am having trouble imagining in which situations that would be useful because I tend to put a lot of thought into defining the right prompt before sending it over.

At what scale would this become a significant concern? While some tech startups aspire to achieve Google-level success, such massive scale is extremely rare in reality. I'm curious about the threshold where issues begin to materialize.

Impressive, this actually 'feels' like an extension for HTML (rather than some custom styles).

As a Dutch citizen, I love the expanse in terms of space. Lately, they have been allocating a lot more green areas as well, making the whole experience very enjoyable.

Example: https://zuidas.nl/wp-content/uploads/2021/03/groenstrook-bee...

Are there any frameworks other programming languages (besides this one) that are directly inspired by Hono?

Thank you for pointing this out, will check your link instead.

> Some utility companies bill every 2 or 3 months. This makes monthly stats meaningless (why is September so high compared to August? did I spent too much or is it the effect of the phone bill of the previous quarter?).

According to Claude 3.5, ledger/hledger handles that situation like this:

    ; Automatically allocate $60 quarterly phone bill across 3 months
    ~ monthly
      Expenses:Phone          $20
      Liabilities:Accrued    $-20
    
    ; When actual bill comes
    2023-09-15 Phone Company
      Liabilities:Accrued     $60
      Assets:Bank            $-60

This approach:

1. Accrues monthly portions ($20) to a liability account

2. When the real bill arrives, clears the accrued amount

> Entries in one month's CSV file may be repeated in the previous or following month's CSV file

I am aware of this problem. The approach in the blog post definitely helps with that because the data is imported fresh every time again. So you could export your bank data with overlapping dates and do something like this prior to importing:

    cat *.csv | sort | uniq >clean.csv

In case people are wondering what the outcome/overview looks like:

https://www.google.com/search?q="pol.is"+graph&udm=2&uact=5&...

Also, this project seems to be based on https://pol.is which is open source.

Since mainstream social media platforms often employ algorithms that amplify polarization and division to maximize user engagement, I’m seeking information about initiatives and platforms (like the one in this post) designed to foster constructive dialogue, mutual understanding, and healthy online interactions. I would appreciate any references to existing or emerging solutions in this space.

Look, I'm a fan of HTMX, but if you want to build a collaborative app like Google Docs or Google Sheets, it wouldn't be a good choice.

Interesting point.

Actually there are more interesting parallels to be found between Trump and Mussolini:

- Both displayed arrogant ignorance and avoided in-depth conversations

- Shared a tendency to appear knowledgeable rather than actually being knowledgeable

- Demonstrated hostility toward the press

- Appointed family members to high government positions

- Exhibiting thin-skinned reactions to criticism

- Showing contempt for experts and professionals

- Took credit for successes while blaming others for failures

- Working with existing nationalist movements

- Attacking democratic institutions as "enemies of the people"

But I don't think that would not derail the discussion. Pretty much any comparison with a dictator leads to painful discussion.

The question is, how would it even be possible to address this in a constructive way. I honestly don't know.

5 logical fallacies in 10 words that's a pretty good score!

Huh?? But how is he wrong?

That's exactly how I do it.

Use Twitter lists to bypass the algorithm:

1. Create a list of accounts you want to follow at https://x.com/<your_username>/lists

2. Bookmark your list instead of the main feed

3. Optional: Use a browser extension to auto-redirect to your list

Result: A chronological feed without algorithmic interference.

> PPS: I think the choice of ^ to handle pointers is one of the worst decision in the syntax.

Funny that's my favourite aspect of Odin's syntax.

    p: ^int

Easy on the eyes.

I also like the convention of the type on the left and its usage on the right:

    p: ^int // ^ on the left
    x := p^ // ^ on the right

Very similar to languages of the Pascal family, but less verbose.

And with structs:

   p^.x  // not necessary
   p.x   // this will do

Could use some more (or better?) examples of how this can be useful.

I agree, Rescript is an actual language rather than syntactic sugar.

Reminds me a bit of Pol.is: https://compdemocracy.org/Polis/ but with direct AI assistance.

> This is how I used to do also by embedding hand-written machine code within a BASIC program and calling it to run natively

Me too; that's also the reason why I wanted that possibility back.

Also, running machine code directly can be done like this:

https://github.com/eterps/loader/blob/master/syscall.nim

You should definitely look at: https://github.com/akkartik/mu/blob/main/subx.md

Basically 'naming things'. It makes it possible to address and distinguish strategies in writing and conversation.

H-Factors ( http://amundsen.com/hypermedia/hfactor/ ) have also been useful to me in the past for the same reasons.