gtirloni

I'm a pretty good documenter so it helps.

I wasn't but now, doing a lot of work inside a harness that is logging everything and can easily create docs from that, I'm finally persisting a lot of knowledge.

Considering how volatile things can get depending on who's president, I'd say even American companies need to "compete against America" if they don't want to get their rug pulled from under them (which, apparently, the legal system allows to easily happen in the US).

This is irrelevant and simply whataboutism.

Vscode fell asleep, Zed won, it's not even a competition

I feel like I'm missing something. Zed won?

That is, if they can even get a work visa through the proper channels while following all requirements.

I had read Kafka's The Castle before dealing with the German immigration office but that experience gave me a new perspective.

I've tried Fable on a few complex tasks and couldn't tell the difference over Opus 4.8, tbh.

Could you share some of these 0-star github repos?

The catch is you didn't know about this until today?

Screw those poor copyright holders without the means to stop frontier AI labs, amirite?

The Ever Given was stuck in the Suez Canal.

Narrowest width:

- Strait of Hormuz (33 km)

- Bab-el-Mandeb (26 km)

- Strait of Gibraltar (13 km)

- Strait of Malacca (2.8 km)

- Bosphorus (700 m)

- Suez Canal (205 m) <- Ever Given

- Panama Canal (55 m)

I hope these people are properly attributing their work to the IDE/editor they used as well. /s

I do spend an enormous amount of time working on projects with AI assistance. When people imply that my output is any less "worthy" because of that, it gives me pause.

The reality is that there are people using AI in areas they would have no chance of producing much quickly. I understand the sentiment against that kind of situation but one shouldn't assume that's every situation.

This is not a limitation that can be overcome by LLMs. Their generative value is in their unreliability. If you turn temperature down to zero, you get a deterministic machine - but you also break every meaningful application I know of in production.

This is not a reasonable argument. Setting the temperature down to zero does NOT give you a deterministic machine. And I have never seen that break any application in production, quite the contrary.

People are still figuring things out, there's a lot of wasted tokens, etc.

This is like complaining a student isn't as productive as a senior engineering.

I think we as an industry haven't even graduated to junior level when it comes to figuring our how to use AI to improve things.

The compute is still real. The VRAM is still real. And the memory bandwidth is where it gets genuinely surprising.

sigh

As someone constantly nerd-sniped, the difficulty is that our instincts are still being formed about what this current era of AI tools can and cannot do.

So when a blocker or an idea pops up, it's very easy to use that magic-like tool to solve it quickly and then go back to whatever it's you were doing before.

However, if you care about the quality of your output, that won't be a quick detour. It will pile up with the other "quick" tasks you were doing simultaneously and that's how you end up with 5-10 sessions working on totally unrelated projects.

This website looks so terrible that I can't tell if it's really owned by Google or a scam.

Check back in 5-10 years, not joking.

I think people are missing the "hidden gems" part of the question here.

How you get deterministic output though? t=0? Pydantic AI outputs?

People who like it aren't vocal, agreed. I think this works for every outrage wave we experience online. Something to keep in mind.

Could you emit a virtual credit card for them that you then cancel after the verification is done?

What "unfucking" looks like though? The law is mandated in the UK. What other way of age verification would work better for them?

Yes. I sometimes had to specifically ask it to NOT add any code to the specs because that would be done at a later stage.

That's definitely not a CLI wrapper. But people are calling Claude Code (clearly a TUI) a CLI so :shrug:

GSD is a collection of skills, commands, MCPs(?), helper scripts, etc that you use inside Claude Code (and others). If anything, Claude Code is the wrapper around those things and not the other way around.

Re: helper scripts. Anyone doing extensive work in any AI-assisted platform has experienced the situation where the agent wants to update 10k files individually and it takes ages. CC is often smart enought to code a quick Python script for those changes and the GSD helper scripts help in the same way. It's just trying to save tokens. Hardly a wrapper around Claude Code.

I think this shows that the model alone isn't the complete story and that these "harnesses" (as people seem to be calling them) shape a lot of the experienced behavior of these tools.

GSD and superpowers aren't CLI wrappers?

Yes, it does help in that way. Maybe I'm still struggling to let go and let AI take the wheel from beginning to end but I enjoy the exploratory part of the whole process (investigating possible solutions, trying theories, doing little spikes, etc, all with CC's assistance). When it's time to actually code, I just let it do its own thing mostly unsupervised. I do spend quite a lot of time on spec writing.

I was using this and superpowers but eventually, Plan mode became enough and I prefer to steer Claude Code myself. These frameworks are great for fire-and-forget tasks, especially when there is some research involved but they burn 10x more tokens, in my experience. I was always hitting the Max plan limits for no discernable benefit in the outcomes I was getting. But this will vary a lot depending on how people prefer to work.

What are you investing time in instead?

I wish them good luck but they will become even more of a niche project.

Worth posting here: https://github.com/ghostty-org/ghostty/discussions