News for hackers. Stuff that matters.

ritcgab

Same experience early this year but I did halve the dosage to 5mg for a month, which was of near zero difference. But the full withdrawal? Oh man it is no joke.
Re:After the AI Crash on Wednesday July 29, 2026 @11:42PM (#49105914)
It is senior programmers that will be in demand again, but at junior price.
Re:It's getting harder to focus every day on Friday July 24, 2026 @10:54AM (#49036609)
And yet people are still happily driving 10 LLM sessions concurrently.
Re:I love LLMs, I hate hype on Sunday July 12, 2026 @10:33PM (#48887168)
I mean, Pfizer has a good margin on every pill they produce too.
Re:Write code like a human will maintain it on Friday July 10, 2026 @10:25PM (#48868018)
Heat death is inevitable in a codebase built mostly by coding agents.
It's a documentary.
Re:The AI backlash is only getting started on Friday June 26, 2026 @11:19AM (#48687707)
We cannot talk about net benefit until we agree on the objective: net benefit for whom?

I don't care if my current job is being replaced or the whole industry I am working in vanishes. But as an individual, if the new technology says it aims to wipe out all the possibilities of my future career, it is not net benefit for me.

Just saying "net benefit" "overall" sounds like some collectivism propaganda.

Frontier models do not trade on 100x earnings either.
I wonder why the apartment buildings in the US never have mini splits but always use PTACs or window units. In my opinion, they suck.
Re:In praise of memcached on Tuesday June 23, 2026 @10:35AM (#48645639)
Because a database is a kv store. Most workloads won't tell the performance difference as long as the store works.
Re:What job interviews taught me about Kubernetes on Monday June 15, 2026 @10:21PM (#48549770)
Multi-cluster federation is still hard.
Re:DNS Is for People – Not for IT Infrastructure on Thursday June 04, 2026 @11:05AM (#48399769)
DNS is a federated, read-optimized, geo-replicated key-value store with eventual consistency.
Re:An AI Hate Wave Is Here on Sunday May 17, 2026 @09:51PM (#48174868)
An AI hate wave is here, yet when I opened the link:

> Axios AI+: Catch up on what's new and why it matters in just 5 minutes.

> Sign up for Axios AI+ to continue reading for free.

People run websites on a VPS with 32MB RAM decades ago and it was fine. What is new about this?
Re:LLMs Are Not a Higher Level of Abstraction on Monday May 04, 2026 @09:28AM (#48008489)
Is "Mostly-reliable" TCP connection a real thing? A TCP connection is either reliable or not working at all. That is what a proper abstraction should be like.
Re:AI can cost more than human workers now on Monday April 27, 2026 @11:07AM (#47922625)
Using AI: pay-as-you-go via token usage.

Using human workers: monthly subscription but rate limited due to biological nature.

They aren't. They act like seniors. But that is all the business needs.
They all know that, and we all know that.

So we are all in this "scheme".

I hope the voyagers can last longer. We are trapped on Earth, but it is just fascinating (and relieving) thinking of them expanding the boundary of human's space adventure.
Re:AGI Is Here on Sunday April 05, 2026 @12:54PM (#47651314)
If humans are not GI, we cannot judge what is AGI.
Re:OpenClaw privilege-escalation bug on Friday April 03, 2026 @10:29PM (#47635050)
Isn't OpenClaw itself a privilege escalation?
Interestingly mixed usage of en dash and em dash across the article.
Well deserved.
Re:What makes Intel Optane stand out (2023) on Sunday March 15, 2026 @02:38PM (#47390414)
All those nice numbers are just beaten by the unit cost. And the ecosystem is a mess.
Re:My Homelab Setup on Sunday March 08, 2026 @03:46PM (#47300568)
Hard pass whenever you host long-term storage without ECC memory.
Re:Verification debt: the hidden cost of AI-generated code on Saturday March 07, 2026 @03:30PM (#47291184)
At the end of the day, it's about liability. Whether you use AI tools to generate the code or not, you are the author of the code, and such authorship implies the liability that you are being paid to take.
But you will never commit them via GitHub's web interface one file at a time :)
Re:Ask HN: Is Linux Safe to Daily drive in 2026? on Wednesday January 21, 2026 @11:21AM (#46707794)
Usually no, at least not out of the box.

If you don't sandbox your desktop applications, it is just not safe by principle. You are running binaries using the same user account with zero audit into what it is doing. You need access control.

I just don't see the point of denying that. People use LLMs for coding for sure and it is not a bad thing, and no one can automate a full project 100% with LLMs.

What I saw is that you created initial commits one file at a time via GitHub's web UI, with multiple follow-up commits just to remove emojis from your scripts:

https://github.com/whispem/minikv/commit/c611897f05235ea4410...

https://github.com/whispem/minikv/commit/59369fcbb2a94c5dd07...

https://github.com/whispem/minikv/commit/0e86e9e328b21170e77...

https://github.com/whispem/minikv/commit/9a4cedcd30c3a00e7ea...

Care to share why those emojis were there in the first place? I guess the next step is to say the Rust codes are 100% handwritten but scripts/docs are AI-assisted? ;)

Older »