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

Apple doesn't ever need to make much money on this software, when they make money on hardware needed to use it.

It’s kind of funny that apparently most of work that’s left after you automated software development is summarizing meetings and building slide decks.

Hey, don't forget booking your flights! Because everyone who has ever flown knows it's very safe to let an RNG machine book something like a flight for you!

Now they can start saying 90% of the meetings will be done by Claude agents by 2027 (And we will all get free puppies)

Then there's the shuffling around of atoms.

As of yesterday OpenAI seems to explicitly allow opencode on their subscription plans.

Yeah, but that would mean me giving money to Sam Altman, and that ain't happening.

Also GPT 5.2 is better than slOpus

Not in my expericence. Gpt 5.2 acts like a desperate teenager, while opus acts like a composed adult - most of the times.

Subjective.

can you point me to this claim? also last i checked trying to connect to OpenAI seems to prompt for an API key, does openAI's API key make use of the subscription quota?

just wanted to make sure before I sign up for a openAI sub




Multiple instances of agents are an equivalent to tabs in other applications - primarily holders of state, rather than means for extreme parallelism.

I have not used Claude. But my experience with Gemini and aider is that multiple instances of agents will absolutely stomp over each other. Even in a single sessions overwriting my changes after telling the agent that I did modifications will often result in clobbering.

See the agent as a coworker ssh-ing on your machine, how would you work efficiently ? By working on the same directory ? No

You give each agent a git worktree and if you want to check, you checkout their branch.


You should try Claude opus 4.5 then. I haven’t had that issue. The key is you need to have well defined specs and detailed instructions for each agent.

Proper sandboxing can fix this. But I didn’t see op mention it which I thought was weird

Op mentions in the follow up comments that he does a separate git checkout, one for each of the 5 Claude Code agents he runs. So each is independent and when PRs get submitted that's where the merging happens.

Personally I just use /resume to switch back to other states when I need to.

> This reliance on readily available solutions, particularly for familiar problems, creates a real risk: engineers may inadvertently atrophy their own problem-solving skills, hindering their ability to tackle truly novel challenges.

Yes, that will happen. But it also happens every time we move up the abstraction ladder. Most engineers go through their entire careers and never do anything TRULY novel.


Agreed, most engineers never do anything truly novel. But the few who did, brought much value to everyone else.

I think the new question here is, if the new status quo is to offload your creative thinking to LLMs, will now any engineers do anything truly novel?

If you're not engaging your mind to create and think on a day-to-day basis, will you be in position to have some new insight?


It's helpful to distinguish problem solving from creative thinking. The main goal of problem solving is to make a problem go away. The main goal of creative thinking is to come up with new problems to solve. Some also call this convergent vs. divergent thinking.

When I want to think creatively but need to solve problems that feel more like housekeeping or toil, LLMs are a useful tool to stay in the right mindset. I have yet to successfully engage with an LLM to help with creative thought. All I've gotten is uninspiring brainstorming.


Have you posted the right link? It seems to be a 2009-2012 collection, when the question was about the 90s.

In what sense? I have a Switch 2 without NSO and it does not allow online play. It also, in the default configuration, needs to authenticate before even entering the online shop. In that scope, it doesn’t even need to have any parental controls.

I buy cassettes. Mostly old, period-correct ones, but some new. I also have a fairly high end tape deck, that these days can be had for rather good price. Our perception of cassettes are mostly warped by the experience of badly recorded tapes played on horrible, unmaintained players, but inherently the tape is much less of a limiting factor to quality than most of the things people use to play music nowadays. In fact, when comparing my vinyl and cassette purchases, I have higher change of getting a bad sounding vinyl than a bad sounding cassette.

Notably, tape decks with separate play and record heads let you listen to the recorded signal, while it's being recorded and quickly switch between the tape and source signal. Even on a good pair of headphones, when correctly dialled in, vast majority wouldn't be able to recognise which signal is the tape.


How well does nix-darwin survive macOS updates?

Installed nix-darwin on 26.0, on a fresh M4 air. I have updated since updated macos to 26.2 through the normal method, no wacky nix stuff there. no issues. no clue on major version changes, but nix-darwin is essentially the nix config language parsed to then run the necessary set of scripts.

I installed on macos14? Maybe 13? And have had minimal trouble since.

Comparable to the amount of breakage I regularly had with homebrew.


There’s also a very real utility to non-streaming media. It turns out that a system that lets you listen to anything is terrible for actually building a collection. Your „library” fills up with tons of stuff you „liked” at some point and saved as some sort of a bookmark. Over time it actually works against the goal of keeping track of the group of records you enjoy. When you introduce friction to the system, whether it’s having to buy something, or even hunt down and download an mp3, it results in better libraries.

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

Search: