Yes, and this is what I mean by "which context the prompt runs in". The subagent approach is different and has pros and cons, and it may in some situations be better (but perhaps not in others). On the other hand, I can also just create a new conversation and paste my own review prompt into it; then take the last turn's summary output and feed it back into my main conversation thread in the unusual event I would need to do so. Spawning a subagent is a convenient shortcut for this, but ultimately, it's the same thing.

> I think the intent behind 'install a language server plugin' is that these tools should lint automatically after every edit, without waiting for an explicit call from the LLM.

This is a great point and I had only checked my session logs for explicit tool calls. I went back and looked for diagnostics injected automatically by the harness after every edit, and whether the agent made use of them.

Claude: neither the Rust or Dart LSPs ever inserted any diagnostic events, but Ty did. Across 627 sessions, ty-lsp injected diagnostics blocks in 186 sessions, with a total of 33 findings. Out of those 33, 32 were dismissed as unrelated (13) or pre-existing (19). Only 1 finding was acted upon. The model is in the habit of running the batch analysis tools (ruff, ty, cargo clippy etc.) and prek anyway, so it would have caught that diagnostic regardless.

Codex: no diagnostic events were inserted by any of the LSPs.

So I won't be reinstalling those LSPs.

I find the mix between slash commands that are programmatic harness configuration and control commands (/config, /model, /feedback, /fork, /usage, etc.) and ones that are little more than prompt template insertion (/code-review, /, etc.) to be a little confusing and unnecessary. A slash command should be one thing, and one thing only: a command for the harness, not the agent.

When I invoke a slash command like /code-review, I should be invoking some additional harness functionality, something above and beyond the agent's sphere of influence - not just pasting some hidden text into the next turn. Otherwise, why wouldn't I just say "Claude, review this code"?

Yet most of these "added value" commands bloating the slash command list, are just shortcuts for copy and paste. I don't want to go to have to learn the syntax of a special /code-review command (which options are positional args, which are --flags, etc.), and I'm much less likely to use or even be aware of a command like this, when I can just ask "Do a balanced code review and fix the issues", or use the GUI to set the effort level to xhigh before asking "Review my code." That way I can also be more specific about exactly what I need, rather than relying on what's in the canned prompt - a prompt which I'll probably never read and vet myself anyway. The value added by the slash command needs to be really high compared to just typing a prompt, for it to justify the friction of discovery and learning the syntax.

So I suppose I'm advocating for a different system. Keep slash commands for meta-level harness control and configuration, and add a new mechanism for canned prompt insertion, one which is tailor made for that purpose rather than overloading the slash command system. Let the user see what's in the canned prompts, and even make adjustments or edits as needed before sending them, one-time or persisted. Provide a GUI in the app with the user's favorite prompts, where the user can add, delete, and edit them, making it easy to invoke and insert them as needed. Or let the agent automatically discover and use them as needed, rather than requiring the user to remember and recall their magic shortcuts and their arguments. That's just one idea.

Skills, plugins, commands, and so on, need to be consolidated not just for code review of course but across the full architecture of how prompt templates are managed.

- Write a .claude/commands/review.md. Simple but deprecated.

- Use a /code-review skill, either one you install or one you just write yourself (it's just Markdown, after all).

- Use the /pr-review subagent. Also just Markdown, but it runs "in the background" and "in parallel", so it must be better, I guess.

- Install the /code-review plugin. This just installs the skills and subagents above.

- Simply ask Claude to review the code. Probably works almost as well as the above in most situations.

They are all just variations of "insert a canned prompt", varying only along the dimensions of (a) how and where the prompt is installed and from where it is sourced, and (b) which context or contexts the prompt runs in. There's not much advice here about which option is best, and no clear best practices seem to have emerged yet either. Personally, I find just asking Claude to review the code works well enough.

Some of the advice here is also off. For example:

"Install a language server plugin. Type errors and unused imports caught after every edit. Highest-impact plugin you can install."

I work mostly with Rust, Python, and Dart, and followed similar advice, installing LSPs for all three in both Claude Code and Codex. Two months later, after heavy development in all three languages and hundreds of sessions - and frequently running out of RAM due to all the Rust analyzer, Dart analysis server, and Ty LSP servers the harnesses were spinning up - I checked the session logs to see how often the agents were actually invoking the LSP tools. The answer was they had invoked them literally once the entire time. I uninstalled all my LSPs and haven't looked back. The agents do just fine using ripgrep and calling cargo clippy, dart analyze, ty check, etc. themselves.

Everybody has free access to Google, Wikipedia, and Bing - that doesn't make sharing quotes from those sources worthy of abusive language telling people they shouldn't have children. Some people really do find value in another person opening up their ChatGPT subscription, crafting a suitable prompt, and passing on the response. There is absolutely some value in that, and many AI models are not even free.

Even just with Google, some people aren't very good at crafting search queries. Crafting a query or a prompt, deciding which model to use, vetting and sharing the response - these all add value beyond just "proxying it slowly and poorly".

Replying with an AI response can absolutely be relevant and valuable to a discussion, just as replying with a quote from Wikipedia or some other authority can be.

On the other hand, flaming people for trying to help you is, frankly, obnoxious, even if you don't appreciate the help or the manner in which it was offered.

People quote other sources sometimes. That's entirely OK. In fact, sometimes it's completely appropriate. We have to get used to the idea that sometimes, that source will be AI, and pretty soon (if not already) it will be just as authoritative and correct as Wikipedia or any other expert the person might quote.

If you don't like it, instead of responding by sending them a link to an aggressive, insulting, disrespectful and frankly low emotional IQ site like this, you can just say, "OK, thanks, that's great, but what's your opinion? I'm genuinely interested in hearing what you think." Unfortunately, if you send a link to this site, you are more than likely to lose the person entirely from the conversation anyway.

Comments URL: https://news.ycombinator.com/item?id=48173030

Points: 3

# Comments: 0

git-hunk provides an enumerate-then-select workflow: list available hunks with stable content hashes, then stage or unstage specific hunks by hash. Hashes are deterministic and remain stable as other hunks are staged or unstaged, so multi-step staging workflows produce consistent results.

Comments URL: https://news.ycombinator.com/item?id=48171716

Points: 2

# Comments: 1

Futures are marked to market every day and gains and losses are paid out daily in cash. So in the example you gave, you would have received $500K profit between when you bought the futures at $1M and when they expire at $1.5M. For cash settled futures that's the end of the story - a final small adjustment happens, in cash, your position disappears, and that's it.

For physically settled futures, the story is the same, but on the day of delivery, you (as the buyer) would be required to pay $1.5M cash to receive the oil, and whoever held the other side of the contract (as the seller) would be required to deliver it to you at some named location, typically a storage facility.

Who loses the $500K? The seller (who was short the contract) has been paying the losses daily as the price moved against them. The payments are handled through the exchange/clearinghouse and their brokerage.

What if they can't deliver? The exchange steps in and provides guarantees.

What's your take on the probability of a recession given the oil price shock and currently high interest rates? Historical parallels do not look good.

Spot prices: https://www.eia.gov/dnav/pet/pet_pri_spt_s1_d.htm

Futures prices: https://oilprice.com/oil-price-charts/

Then to see Fidelity take a political stand - well, given my terrible experiences with that company, I can't say I'm surprised.

Comments URL: https://news.ycombinator.com/item?id=47927938

Points: 4

# Comments: 1

Yet prediction market fraud is made an enforcement priority, except to say that nobody close to Trump's own cabinet will be prosecuted - the little guys will be made an example of to make it seem like those at the top are taking the moral high-ground. "Every accusation is a confession."

I think we all can guess at the truth here.