thunspa

Care to share more about your pi setup? I've recently started using it (after long-time Claude Code work) and was wondering how you'd achieve these long-running tasks. Do you allow it to spawn sub-agents? Thank you!

Also interested.

This is very cool, congratulations!

This could probably be solved with a better keyboard

Could you share a bit more about this?

I've been doing this more and more over the past year, but I just write on plain white paper and throw it away after the stack on my desk gets too big.

Like the author, I don't seem to ever need to read my old notes. Instead, it works wonders as a mental bucket of sorts and I've found paper to be extremely powerful for this. I tried doing this on a Surface Pro, for example, but it was significantly less enjoyable or effective.

Now with LLMs helping me write code, planning ahead on paper is even more useful.

I work in Python as well and find Claude quite poor at writing proper tests, might be using it wrong. Just last week, I asked Opus to create a small integration test (with pre-existing examples) and it tried to create a 200-line file with 20 tests I didn't ask for.

I am not sure why, but it kept trying to do that, although I made several attempts.

Ended up writing it on my own, very odd. This was in Cursor, however.

I resonate with the phrase: "You never learn to ask good questions"

I think there's an unfinished journey whenever I hear such a story. Like someone who learns just enough about music that they "feel like" Bach is better than Taylor Swift, but then never move to the other side to use their newly acquired competence to understand what's nice about Taylor Swift's music as well.

E.g. a great designer will be able to design houses in very different styles, because they can understand what gives each of those styles its own specific beauty.

There's a lot to say about this, but I think your coffee friend never passed the snob-like point, which is a point I think most people reach when they learn just enough about something and that makes them feel superior. But if you keep going, then you start to understand what makes Italian coffee great as well, for example.

Wrt to coffee, I speak from experience, after going through very expensive equipment, I have learned to enjoy very different styles of roasts, coffees, etc. I still have preferences, I'm just far less judgmental.

Applies to most of my hobbies, I've seen this trajectory very often.

unexpected AI W. Congratulations on the new job!

The "hey if you squint this thing it looks like religion for the non-religious" perspective is just one I've heard countless times

To be fair, we shouldn't bundle Augustine and Thomas Aquinas with John MacArthur and Joel Osteen. Meaning that some religious thought is more philosophically robust than other religious thought.

What makes an artificial mind greater than ours?

Do you assume that someone will stumble into creating a person, but with unlimited memory and computational power?

Otherwise, if we are able to create this person using our knowledge, we will most certainly be able to augment humans with those capabilities.

I mean, it all depends on how one defines "high quality context".

This never occurred to me, but it is a brilliant take.

can you say more? how do you do it?

Or learning to actually code. I can't see why I would ever learn to use these kinds of tools.

As a developer, I can just make it myself. Now with LLMs, if it's very simple and bounded, I can just vibe most of it with very little to lose.

As a lay person, I don't see what the TAM for this is. Who will spend the time to learn how to drag and drop an application?

Saw Rich Hickey say this, that it is a known fact that tested code never has bugs.

On a more serious note: how could anyone possibly ever write meaningful tests without a deep understanding of the code that is being written?

Yes, this is what I'm fearing as well.

That we will end up just trying to review code, writing tests and some kind of specifications in natural language (which is very imprecise)

However, I can't see how this approach would ever scale to a larger project.

In writing the code that is supposed to implement my idea, I find that my idea has many flaws.

Sending that idea to an LLM (in absence of AGI) seems like a great way to find out about the flaws too late.

Otherwise, specifying an application in such detail as to obtain the same effect is essentially coding, just in natural language, which is less precise.

I never used Telegram, but was under the impression that it's similar to Signal. However, Pavel Durov recently tried to meddle in the Romanian elections process[0] in a very bizarre, almost desperate attempt to misinform the electorate.

[0] https://www.lemonde.fr/en/pixels/article/2025/05/23/why-is-t...

I just cmd + shift + p -> disable Cursor tab -> enter

Sure, you could just add a shortcut too.

After a while, it turns into a habit.