tstrimple

I find it incredibly frustrating how effective conservative propaganda is even on non-conservatives. The "fiscal responsibility" and "small government" stuff are just examples of the many convenient statements they have with absolutely zero bearing on reality in how they legislate. So many liberals even believe conservatives used to "just want a slow and measured rate of change" and Trump is somehow different. All that has changed is the amount they feel comfortable saying out loud. They are dispensing with the euphemisms and saying what they really mean now.

Have you tried a custom output style? CLAUDE.md instructions are appended to the system prompt. A custom output style can replace the system prompt. At least the part of it pertaining to voice and persona. The reason it forgets over long chats is the context size starts getting too large. Instructions weigh more strongly the later they appear in the context. This is necessarily true otherwise you couldn't change your mind in a conversation. The model would stick with what you originally said. For the output styles, there is a per-turn "reminder" that gets added to the context asking it to "remember" the content in the system prompt. That's why it has more staying power than the CLAUDE.md instructions in long conversations.

https://code.claude.com/docs/en/output-styles

That's why I mentioned hooks in particular. That feels like the right layer for this sort of adjustment. A PostToolUse hook on Edit|Write would be much more reliable than just a CLAUDE.md instruction. The consistency I get from CC comes from instructions at multiple layers.

CLAUDE.md heirarchy: At the top level you've got general instructions you want all contexts to follow and each subdirectory can add more specific instructions in their own CLAUDE.md files. References in CLAUDE.md are not fully loaded into the context. They are loaded opportunistically. So keep important instructions in the CLAUDE.md file itself and not a referenced or linked file.

Rules files: These offer path scoped rules via frontmatter. So you could have specific rules for certain types of files Claude Code interacts with. Certain rules for handling all .cs or .js files for example.

Auto-memory: You cannot rely on this one. I use auto-memory as a cache for potential future CLAUDE.md instructions. I have an audit process that kicks off when the auto-memory gets beyond a certain number of entries.

Skills: On demand context. I don't tend to use /skills explicitly. I tend to have them used in context. I've got a task tracking system I call threads. So whenever I say "Create a thread for X" it has always reliably followed the specific instructions. I've got skills for managing my NAS for searching historical session for sharing content and other things. I use them a lot of times in place of MCP servers.

Hooks: Deterministic scripts run on lifecycle events. I've got hooks that run linters on code files post edit and hooks which tie into the request / response events to push my history into a SQLite database.

Output Styles: CC ships with a few different styles, but you can create your own. This is key for changing the default voice. CLAUDE.md instructions are appended to the system prompt and can fight against the system prompt. A custom Output Style would let you replace the instructions in the system prompt with your own instructions. This can be done at the user level or per project.

I think that was absolutely true for a time. But now outside of Apple TV who is doing high budget quality TV? Certainly not Netflix. Best they can offer is a Korean or Spanish soap opera.

Distribution got better for a time and then became the same old bullshit. When Netflix was the only dominant streaming service it was great. Now ever other person has their own custom streaming service and content is distributed across them. Back to the high seas for me. I don't mind paying. I mind the inconvenience. There was a brief window when Netflix was improving TV quality. But now it's almost all non-native US made originals that are cheap as fuck. They aren't all bad. My wife has a weakness for Spanish and Korean dramas for some reason. But it's all noise to me.

And thus microservices were spread across the unsuspecting world.

I don't know that we want to resort to comparing which sellout is worse.

MBA syndrome. People who allegedly know "business" but know fuck all about the domains they frequently ruin.

There is no realistic path to human oversight for the vast quantities of data these models are trained on. Imagine the cost of having every Reddit comment ingested human reviewed. Insane.

I'm curious about this because I've had Fable decompile games and help me understand what's going on inside the game itself and it never complained. I'm not sure what it takes to trip the "safety" guards but digging into game code and data files doesn't seem to be a barrier at all. I've used CC to build some personal game mods a few times now. Once for a game with no modding capability explicitly exposed.

I have to wonder if everyone else is just running these models raw without any custom instructions. I hear all these things about voice and code comments and those are all things I've dealt with long ago via claude.md instructions, rules, and hooks. My claude can already respond in any "voice" I want and the quantity and quality of comments is within my control.

I switched between Windows and OSX and Linux frequently over the last 20 years. The longest lived Linux experience is the one I'm currently running. NixOS with Hyprland for a desktop environment. None of the annoying opinions of KDE or Gnome from either an app or shortcut or configuration standpoint. Super easy to map keybindings to be similar to OSX. Hyprland makes my ultrawide monitor actually a useful replacement for dual monitors. I'm using NixOS for both my desktop environment as well as my UGREEN NAS and being able to have shared configuration across them is very helpful. Add in Nix and Home Manager for OSX and my terminals have never been more consistent across all my systems.

A lot of people are stupid though. The law had zero age verification in it for example, but tons of breathless conspiracy theorists are claiming that's what it is about.

The easier a thing is to do, the more often people will do it. And downloading a file and pressing print is a far, far lower barrier than machining steel.

The baseline is already being established. In the majority of US states now you have to send your photo ID to a 3rd party to access "adult content". The conspiracy theories around the California and Illinois laws are senseless given the speed run the rest of the country is doing to ending privacy online. We get threads with hundreds of comments about California "fighting against privacy" and basically silence on the dozens of other states implementing far more privacy invasive legislation.

It's easy when you just ignore all the externalities! I'll be paying for the decommissioned nuclear plant near us for the next 20+ years through my electric rates. The plant is closed and no longer operational but still has to be staffed and guarded for decades during the decommission process. Imagine this issue scaled up across thousands of SMRs.

The party of "personal responsibility" literally never hold themselves accountable for anything. It's always someone else's fault.

Conservatives are the problem. Always have been. From the time they started a civil war to maintain slavery to the time they fought against women's suffrage to their fights against LGBT rights. When public sentiment turns on them too much, they retreat to "states rights" and other dog whistles. When they feel like they can get away with it they say "Haitians are eating your pets!" and "Mexico is sending us their rapists and murderers!" But fundamentally they have always wanted these things and believed these things. Political Correctness just kept them behaving more civilized for a short time.

Step debuggers can 100% cause or hide race conditions that are exposed via other mechanisms. Especially in multi-threaded apps as I stated. If you've never encountered it, I'd suggest you've never debugged really nasty race conditions or serious bugs.

My NixOS + Hyprland desktop largely has the same shortcuts as OSX. I super+space for launching things. Super+Q for quitting. Custom keybinds are trivial. I always thought KDE and Gnome were kind of garbage, but Hyprland gets out of the way and really lets me build the desktop environment I want.

What sorts of integrations are you looking for? I use a few tools:

mautrix-imessage - imessage bridge so I can send/receive messages from my Linux computers

vdirsyncer - Pulls iCloud calendar data for display in my waybar.

I also have a "headless" (the screen broke) M1 MBA. I'm using it as a bridge to my Apple ecosystem. I use EventKitCLI to pull / update reminders for example. I have a Claude Code context focused on cooking and adapting recipes. It can add stuff directly to my shopping list via a skill.

> make super sweet tea/lemonade. Maybe it's slightly healthier.

Not the way my family in Oklahoma made sweet tea. Basically so much sugar that the tea has to stay warm or it precipitates out. It's disgusting.

Setting a goal seems like it would be more appropriate here.

https://code.claude.com/docs/en/goal

The specific reason I reach for print debugging first is it works literally everywhere across all technology stacks. I regularly work with different projects written in different languages. Printing works across all of them and the logs are easy to aggregate across different parts of the stack (microservices or other service composition patterns) giving you a broader range of data than you typically get from a step debugger. I've personally found that step debuggers are more likely to create Heisenbugs. Especially in multi-threaded applications or dealing with race conditions.

It's only ever brought up by people with absolutely nothing to contribute to the conversation. So you've got that going for you.

Nope. It's a stupid take. Especially considering over half of the US states have jumped right to looking for that verified answer but no one on HN gives a shit when red states do things. They only ever bitch and moan when a blue state is doing something far less invasive and then they have to turn it into a conspiracy theory that the blue states are secretly trying to do what the red states are just blatantly doing in the open with zero push back.

This is a nonsense argument considering over half the states have already put actual ID verification requirements in front of "adult content". Blue states are trying to ratchet to what red states just pass? Why? They could just follow suit and require you to upload your ID to a 3rd party service and scan your face. Your conspiracy theories don't seem to hold up to reality.

That and there was basically zero uproar on HN for any of the 20+ red states requiring 3rd party ID verification and or face scanning for access to adult content. But when blue states try to implement something far less privacy invasive, that parents can opt into, people lose their minds.

https://dailycitizen.focusonthefamily.com/pornography-age-ve...

Sure buddy. None of those changes to voting are "manipulating".

Counties that used to need federal preclearance closing 1,688 polling places once the Supreme Court said they didn't have to ask permission anymore is "just good fiscal sense".

North Carolina's legislature pulling data on which voting methods black people disproportionately used, then writing a bill restricting exactly those methods, is "election integrity". The Fourth Circuit threw the whole law out and said it targeted "African Americans with almost surgical precision". But sure. To you it's not manipulation.

Alabama passing voter ID and then closing 31 DMV offices, including every single one in counties over 75% nonwhite, is "saving taxpayer money". The federal DOT found it violated Title VI and made them reopen. Budget cuts, though.

Ohio's redistricting commission being told seven times by its own Supreme Court that the maps were unconstitutional, then running out the clock and holding the 2022 election on them anyway, is "federalism".

And then there's the fake electors, which people gloss over because they don't know how the process works. When you vote for president you're not voting for the president. You're voting for a slate of electors. Your state certifies which slate won, sends that certificate to Congress, and on January 6 the Vice President opens the certificates and they get counted. That's the whole election. It's paperwork.

So in 2020, in seven states Biden won, Republicans got a second set of people together, had them sign certificates declaring themselves the real electors and Trump the winner, and sent those to Congress and the National Archives. 84 people signed. Ron Johnson's office tried to hand-deliver the Michigan and Wisconsin ones to Pence on the floor on January 6. The plan, written down in a memo a month beforehand, was for Pence to count those and reject the real ones. That's "pragmatism".

Louisiana's governor suspending a congressional primary that was already underway so the legislature could delete a majority black district is just... I don't know. I'm out of ridiculous fucking excuses anymore.

And the third that didn't vote "said sure, whatever"? Some of them did. Some of them got purged off the rolls, or showed up to a precinct that isn't there anymore, or needed an ID from a DMV their county no longer has. Those are not the same group of people. Keeping the second group home is the point of everything above. Republicans have spent a decade making it harder for specific people to vote, and then pointing at the resulting turnout as proof nobody cares. Well it worked on you at least.

Oh great. You've summoned the Sinophobes. You literally cannot say anything positive about China without these people coming out of the woodwork with their tired talking points.