News for hackers. Stuff that matters.

blinkbat

could be an interesting idea if the article wasn't slop as pure as the driven
Re:Talk Without Talking on Friday October 30, 2026 @04:53PM (#50310011)
Faq doesn't seem to indicate whether it works or not
Does this shit actually work lol
Re:Poll: How do you run your coding agents day to day? on Friday October 30, 2026 @11:52AM (#50305705)
several sessions in one or multiple instances of vscode.
Terrible slop copy.
Re:What are you using for agentic coding now? on Tuesday October 27, 2026 @12:33AM (#50254734)
LLMs
Bit vague innit
Re:The Art of Unix Programming on Saturday October 24, 2026 @01:58PM (#50223589)
The art of https
Lol
Re:Ask HN: Time to read those programming books? on Thursday October 22, 2026 @02:54PM (#50199515)
Writing is the slowest part
Re:Hello World, AI Doompop on Wednesday October 21, 2026 @10:42AM (#50181877)
at the risk of sounding dismissive, fuck this lame shit
Re:Ask HN: Are agent benchmarks the new credit ratings? on Tuesday October 20, 2026 @04:14PM (#50172019)
Tests and community are far better indicators
Re:Ask HN: Time to read those programming books? on Tuesday October 20, 2026 @03:51PM (#50171718)
yes, you get better _at programming_, specifically your own programs, and in a general sense. but is there a way to hone the reading of code that is not yours?

> Now we are using LLMs to write the code to the point where we need to assign ourselves homework to keep our skills?

yes!

> Why not just use the LLM in a way that doesn't prevent us from absorbing the knowledge?

the economic viability of this is unproven in our new future.

Re:Ask HN: Is Agentic Coding really faster? on Sunday October 18, 2026 @06:46PM (#50146274)
typing speed was never my strong suit, I also was heavily into readable formatting which I often did by hand in a mixed team. so I'm not one of these stays-in-vim people that can pop out reams of code. in this way I appreciate the manual lift an LLM gives me.
Re:Ask HN: Is Agentic Coding really faster? on Sunday October 18, 2026 @06:27PM (#50146129)
I will never be able to type out a large feature at even 1/100th of the speed of today's frontier LLMs -- even if I literally had a complete plan of every line of syntax I'd use.

now, the correctness of said feature is still up for debate, but we cannot argue that the thing came into existence faster than human hands could have wrought.

Re:Show HN: Vibe coded a city builder set in ancient Rome on Sunday October 18, 2026 @06:26PM (#50146119)
panning is choppy, I would expect to be able to arrow or wasd pan as well.
Re:My Obituary for Programming on Sunday October 18, 2026 @06:20PM (#50146084)
I mean, if the AI can improve your program's speed by 2-3x, and you want people to use the program happily, you should probably run it past that AI at SOME point.

maybe it can teach you, if you do it methodically, how such speedups are possible for you to handmake in the future.

there's a real disconnect between "I liked making this" (which is great!) and "my users would be happier if this were faster". the users, sadly, do not care how much you liked writing the code if that code is their bottleneck.

Re:Ask HN: How are you all ensuring your AI code isn't slop? on Saturday October 17, 2026 @12:45PM (#50132646)
I do extremely regular audits
Re:AI companies plot "day after" scenarios for public revolt on Saturday October 10, 2026 @01:41AM (#50029944)
Yep seems like we could wrap that up in a day
Re:Need Unbiased and Real Assessment on Saturday October 10, 2026 @01:27AM (#50029860)
If you find it useful someone else might.
Re:Software's centaur age may last decades on Saturday October 10, 2026 @12:53AM (#50029688)
Wow a good post I agree with almost entirely.
Re:Midterms 2026 on Wednesday October 07, 2026 @01:42PM (#49996180)
can we get the version with a lot of the slop refined out?
Re: on Tuesday October 06, 2026 @05:16PM (#49984282)
the guarantee that people on the internet could be trusted to leave sound, safe comments would persuade me to add one, but that'll never be
I assume yes they would have to verify it somehow...
what could go wrong
Re:Ask HN: Vibecoding Follies on Tuesday October 06, 2026 @12:51PM (#49981060)
I make games if that counts.
Re:Senior Software Engineer on Tuesday October 06, 2026 @12:14PM (#49980621)
Mysterious HN Post
Re:How funny are the frontier AI models? on Tuesday October 06, 2026 @12:02PM (#49980466)
claude also recently tried to incorporate my dead mom into a joke after I offered it that knowledge. the apology after wasn't exactly heartfelt.
Re:How funny are the frontier AI models? on Tuesday October 06, 2026 @12:01PM (#49980448)
I talk about this with my wife a lot. what I keep coming back to is "astra/opus/fable has jokes, but they aren't funny." recent exchange, from memory:

me: lua was an odd choice for that script

astra: it's the scripting lang for aseprite, convenient for pixel editing

me: ahhh. also convenient if you think arrays should start at 1...

astra: perfect pixels, but it steals your zero index

like, this has the timing or cadence or something of a joke, but it isn't exactly funny.

Older »