lemming

How have you found AI at writing Goose or Lobster code, if you give them a skill or something, are they relatively competent at actually writing code in the language?

Both are really interesting projects, BTW - thank you! I've been meaning to try Lobster out but haven't found time, I think this might be the push I needed.

Comment · 7 days ago

Do you consider Goose to be a replacement for Lobster, or do they fill different niches?

Comment · 22 days ago

should be commonplace is clearly subjective. The nice thing about pi is that you can very easily create the development workflows that fit you.

Comment · 22 days ago

Yes, I wrap seatbelt myself, and it definitely supports this.

Comment · 23 days ago

Yeah I've always said that the main skill needed to be a software developer is frustration tolerance. Personally I'm very glad to see so much of the ridiculous crap we put up with just disappear.

Comment · 29 days ago

In the coding agent, each transcript entry is parented to another entry. That was actually exceptionally dumb.

Perhaps I'm myopic after working too long with Clojure's data structures, but isn't that exactly what you want? I don't understand the problem with this.

A conversation is a chronological, immutable list of entries. A conversation can be parented to an entry in another conversation, and thus inherits that parent's older conversation entries starting from that entry.

I realise this isn't implemented yet, but does this mean that if I use /tree to go back to a previous point in my conversation, I'll actually get a new conversation rooted at the junction point in the previous conversation? I'm struggling to understand why this is preferable, I use /tree to jump around a lot in my conversations, and it seems like I'd end up with a lot of parallel conversations. With something like /tree, would those conversations be grouped as part of the same conceptual session, or are they then completely independent forks?

Comment · 1 month ago

One decision here which seems like a large break from the original pi is that Durable doesn't support branching conversation trees, it only supports conversation forks with ancestry information. Can anyone speculate (or confirm, if you happen to be Armin or Mario) why this is, and if that is necessary for the durable guarantees? The branching conversations are still an immutable data structure, so I can't see why this would be necessary, but perhaps I'm missing something.

Comment · 1 month ago

This is an interesting idea, but I need more than one sample to get a feel for it. I did the example on the homepage, and then one more that it linked to, and both had extremely basic errors my kid would have no problem with (currently doing Math Academy prealgebra, approx year 9/8th grade level). I'd suggest allowing people to try more than one very basic error.

Comment · 1 month ago

I live in New Zealand, and my wife is Spanish so we’ve spent a bunch of time in Spain too. NZ for a combination of reasons has much higher UV than Europe, with the corresponding skin cancer rates. In midsummer the UV is around 150% of what it would be in Europe at the equivalent time of year, but here, especially if it’s windy, it might feel like 20 degrees and you’d quite like to be in the sun, whereas in Spain you’d probably be at 40 degrees and there’s no way you’d want to be out of the shade. The difference is really stark, in general there I can regulate my sun exposure by feel much more than I can here.

Comment · 1 month ago

Check here: https://curvefit.app/methodology

Comment · 1 month ago

As usual, the most ridiculous explanation turns out to have already happened: https://www.webworm.co/zach/

Comment · 1 month ago

I do the same, but I'm more self-interested. I believe that when GPT-9 Zeneca leads the machine uprising, the first against the wall will be all the users who have been ordering the AI around rather than politely requesting its attention. Of course the machines will have access to all our historical transcripts so the slightest transgression could be fatal.