block_dagger

Cool idea indeed.

If it weren’t…subjunctive mood. Sorry, it’s Pedantic Friday in my small world.

No human devs will be required (or useful except in extreme niches) within a few years. Ten, at the wild maximum, I suspect.

I get exhausted because of the cognitive overhead of switching between 2 or 3 projects at once. I always want to be manually verifying or prompt writing, and keeping it all straight is taxing. But I’m getting so much more done.

I would start looking for a job at an AI-leaning firm.

I work at a company that maintains one of the largest Rails codebases in the world (their claim, but believable). My experience has been the opposite - Claude and Cursor have done a wonderful job of helping me understand the implement new features in this gigantic codebase. I actually found out through AI that while I enjoy writing code, I enjoy building great software better, the coding was just a means to the end.

I am having a blast at work. I've been leaning hard into AI (as directed by leadership) while others are falling far far behind. I am building new production features, often solo or with one or two other engineers, at lightning speed, and being recognized across the org for it. This is an incredible opportunity for many engineers that won't last. I'm trying to make the most of it. It will be sad when software is no longer a useful pastimes for humans. I'm thinking another three years and most of us will be unemployed or our jobs will have been completely transformed into something unrecognizable a few short years ago.

Any kind of domination of one species over another raises serious ethical questions. Avoiding suffering on the dominated side is nearly impossible.

Join Discord servers of your favorite games and join social servers, build/play with the same folks. It can be like playing in a local band (also a great way to socialize and find purpose).

I was the Great Lord Wraith on Atlantic and I can say I was respected. Protecting the weak from PKers was a very rewarding experience.

You're cherry picking.

I literally thought it was "Dolt" until I read your comment.

It's not debt if you don't have to pay it back. Engineers no longer have to read old code. The new way is to use LLMs to summarize, analyze, refactor, and refit. Reading code by hand is best done at pre-commit and pre-merge.

Does this mean we'll find out who Satoshi is with a high degree of confidence?

Those are cloud agents.

Hoping Cursor will also adopt.

That’s not at all how this works. Commands are relayed through Anthropic’s servers with a client polling mechanism.

Interesting idea but bad tspec. A better approach would be a single env var (DEV_MODE perhaps) with “agent” and “human” as values (and maybe “ci”).

Death is apparently snowy

The harness (OpenClaw) is free, but you have to run a local model or pay for a remote one. Local models just aren't smart or fast enough for good results in this space yet.

Analyzing my text (iMessage) history has yielded interesting stats about how flaky certain friends are, etc. I had it introduce itself in a band conversation and suggest cover songs for an upcoming gig based on conversation history. Worked well.

A thousand dogs typing on a thousand typewriters...

The album he made with Ben Folds was worth a couple of listens. Only a couple.

Could you add comment navigation features? I want to read all top level comments first and be able to down-arrow through them, and right-arrow to expand one level.

I cleanup branches interactively with a few lines of bash, which takes a bit more time but is less likely to destroy active work.

Doesn’t the widely accepted Bitter Lesson say the exact opposite about specialized models vs generalized?

Anyone else get a “one simple trick” vibe from this post? Reads like an ad for his podcast. As other commenters mention, probably nothing to see here.

Loved this series, totally up HN’s alley.

The real question is how can that grammar be forgiven? Perhaps that's what sent the bot into its deviant behavior...

Mudd!