Hacker Newsnew | past | comments | ask | show | jobs | submit | jspdown's commentslogin

Domestic mass surveillance might feel tolerable when you live in the country conducting it. But how would you feel about other countries adopting similar policies, and thereby mass-surveilling the American people? Because that's exactly what these policies authorize when applied to the rest of the world.

Americans always think they're exceptional so they have the divine right to do things that the rest cannot.

Maybe that’s why they like Israel so much.

The way the anthropic statement was written really stood out to me. How they posture themselves in favour of surveillance for foreign countries or the existence of fully autonomous weapons if they don't threaten US citizen lifes.

I wonder if this is how some non minority of American thinks or was just worded like that to try to appeal to the "most radical patriots"


The bad news for American people is that "others" are pretty good at these technologies. When I read an important AI paper chances are all the names on it are non-American, even for papers from American labs. In a real war, this becomes problematic.

Every nation has some bias but I think Americans have power poisoning for being the dominant power for so long. They think they are entitled to do anything and believe they are the good guys in the history. Well...


On my personal coding agent I've introduced a setup phase inside skills.

I distribute my skills with flake.nix and a lock file. This flake installs the required dependencies and set them up. A frontmatter field defines the name of secrets that need to be passed to the flake.

As it is, it works for me because I trust my skill flakes and skills are static in my system: -I build an agent docker image for the agent in which I inject the skills directory. -Each skill is setup when building the image -Secret are copied before the setup phase and removed right after

All in all, Nix is quite nice for Skills :)


I pay a Max subscription since a long time, I like their model but I hate their tools:

- Claude Desktop looks like a demo app. It's slow to use and so far behind the Codex app that it's embarassing.

- Claude Code is buggy has hell and I think I've never used a CLI tool that consume so much memory and CPU. Let's not talk about the feature parity with other agents.

- Claude Agent SDK is poorly documented, half finished, and is just thin wrapper around a CLI tool…

Oh and none of this is open source, so I can do nothing about it.

My only option to stay with their model is to build my own tool. And now I discover that using my subscription with the Agent SDK is against the term of use?

I'm not going to pay 500 USD of API credits every months, no way. I have to move to a different provider.


I agree that Claude Code is buggy as hell, but:

> Let's not talk about the feature parity with other agents.

What do you mean feature parity with other agents? It seems to me that other CLI agents are quite far from Claude Code in this regard.


Which other CLI agents are that? Because I've found OpenCode to be A LOT better than Claude-Code.

Whats better with opencode? Never tried it. I like that claude code has double escape, shift + tab, team of agents

I haven't used opencode but pi agent runs rings around claude code. Never eats tons of CPU on big outputs, no flickering, open source, tree-based context instead of claude's linear context, easy to toggle collapsing/expanding tool outputs, built for extension with runtime reloading of extensions and skills, etc. You can easily build your own amp-code like handoff mechanism, customize the UI (i see models' edit diffs syntax-highlighted with delta, and just added a keybind to list session-edited files + files from git status in fzf), etc.

Meanwhile with Claude Code I've had to get claude to decompile the editor (extract JS from the bun executable) _twice_ to diagnose weird things like why some documented config flags were not taking effect.

Opus is great - but I'd rather use a different model than be forced back into Claude Code.


The great force of claude code is that you can use claude sub, you can’t with pi unfortunately

I regret ever promoting that Claude Code crap. I remember when it was nothing but glowing reviews everywhere. Honestly AI companies should stick to what they are good at: direct API interface to powerful models.

We are heading toward a $1000/month model just to use LLMs in the cloud.


They still have that. If that's all you wanted then what are you complaining about

>I'm not going to pay 500 USD of API credits every months, no way. I have to move to a different provider

It's funny, you are probably in the cohort that made Antropic have to pursue this type of decision so aggressively.


I had a Claude code instance using 55 GB of RAM yesterday.

I got so tired of cursor that I started writing down every bug I encountered. The list is currently at 30 entries, some of them major bugs such as pressing "apply" on changes not actually applying changes or models getting stuck in infinite loops and burning 50 million tokens.

I tried to have Cursor change a list of US States and Provinces from a list to a dictionary and it did, but it also randomly deleted 3 states.

It's not HTML purism. It's simply recognizing that HTML and CSS have evolved a lot and many things don't need (or are close to not need) JS anymore. This shouldn't be taken as an anti-JS article, everyone benefits from these gradial improvements. Especially our users who can now get a uniform experience.


This article compares a single ChatGPT query against 1h of video streaming. Not apple to apple comparison if you ask me.

Using Claude Code during an hour would be more realistic if they really wanted to compare with video streaming. The reality is far less appealing.


Consider how many folks use Claude Code for an hour vs. streaming many hours. Globally, not among HN readers.


My bad, in the context of the article you are definitely right.

I think I was biased by the fact that this argument was used in an HN comment where people tend to be heavy users of LLM based agents.


You can type Option+Enter. A more standard Shift+Enter would have been better but until then that's the best we have


Looks like it's OpenAI response to Google's Agent Payment Protocol (AP2) but without the micro transaction part.


Do you mind telling us a bit more? I never used OpenCode, what makes it better in your opinion?


I'm consistently hitting weird bugs with opencode, like escape codes not being handled correctly so the tui output looks awful, or it hanging on the first startup. Maybe after they migrate to opentui it'll be better

I do like the model selection with opencode though


- opensource with an SDK so you can build things on top of

- supports every LLM provider under the sun, including Anthropic

- has built-in LSP support https://opencode.ai/docs/lsp


There's a bit of UI around it where you can accept the plan. I personally stopped using it and instead moved to a workflow where I simply ask it to write the plan in a file. It's much easier to edit and improve this way.


Yeah, I just have it generate PRDs/high-level plans, then break it down into cards in "Kanban.md" (a bunch of headers like "Backlog," "In-Progress", etc).

To be honest, Claude is not great about moving cards when it's done with a task, but this workflow is very helpful for getting it back on track if I need to exit a session for any reason.


i've experienced the same thing. usually i try to set up or have it set up a milestone/phase approach to an implementation with checklists (markdown style) but it's 50/50 if it marks them automatically upon completion.


I have this in my CLAUDE.md and it works better than 50/50. Still not 100% though:

### Development Process

All work must be done via TODO.md. If the file is empty, then we need to write our next todo list.

When TODO.md is populated:

1. Read the entire TODO.md file first 2. Work through tasks in the exact order listed 3. Reference specific TODO.md sections when reporting progress 4. Mark progress by checking off todos in the file 5. Never abbreviate, summarize, or reinterpret TODO.md tasks

A TODO file is done when every box has been checked off due to completion of the associated task.


I'm wondering if it's not just: spawn multiple time the same prompt and take the best


It is exactly that


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: