frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Claude Sonnet 4 now supports 1M tokens of context

https://www.anthropic.com/news/1m-context
940•adocomplete•11h ago•517 comments

Search all text in New York City

https://www.alltext.nyc/
127•Kortaggio•3h ago•29 comments

Ashet Home Computer

https://ashet.computer/
201•todsacerdoti•8h ago•43 comments

Show HN: Building a web search engine from scratch with 3B neural embeddings

https://blog.wilsonl.in/search-engine/
363•wilsonzlin•11h ago•59 comments

Journaling using Nix, Vim and coreutils

https://tangled.sh/@oppi.li/journal
86•icy•13h ago•29 comments

Bezier-rs – algorithms for Bézier segments and shapes

https://graphite.rs/libraries/bezier-rs/
16•jarek-foksa•3d ago•0 comments

Training language models to be warm and empathetic makes them less reliable

https://arxiv.org/abs/2507.21919
221•Cynddl•13h ago•221 comments

A gentle introduction to anchor positioning

https://webkit.org/blog/17240/a-gentle-introduction-to-anchor-positioning/
49•feross•4h ago•13 comments

Show HN: Omnara – Run Claude Code from anywhere

https://github.com/omnara-ai/omnara
220•kmansm27•10h ago•111 comments

Visualizing quaternions: An explorable video series (2018)

https://eater.net/quaternions
11•uncircle•3d ago•3 comments

Multimodal WFH setup: flight SIM, EE lab, and music studio in 60sqft/5.5M²

https://www.sdo.group/study
190•brunohaid•3d ago•81 comments

Blender is Native on Windows 11 on Arm

https://www.thurrott.com/music-videos/324346/blender-is-native-on-windows-11-on-arm
125•thunderbong•4d ago•50 comments

WHY2025: How to become your own ISP [video]

https://media.ccc.de/v/why2025-9-how-to-become-your-own-isp
107•exiguus•10h ago•13 comments

LLMs aren't world models

https://yosefk.com/blog/llms-arent-world-models.html
242•ingve•2d ago•129 comments

Blender on iPad Is Finally Happening

https://www.creativebloq.com/3d/blender-on-ipad-is-finally-happening-and-it-could-be-the-app-every-artist-needs
20•walterbell•1h ago•7 comments

Launch HN: Design Arena (YC S25) – Head-to-head AI benchmark for aesthetics

61•grace77•11h ago•24 comments

A spellchecker used to be a major feat of software engineering (2008)

https://prog21.dadgum.com/29.html
140•Bogdanp•4d ago•129 comments

Go 1.25 Release Notes

https://go.dev/doc/go1.25
134•bitbasher•5h ago•25 comments

Why are there so many rationalist cults?

https://asteriskmag.com/issues/11/why-are-there-so-many-rationalist-cults
410•glenstein•12h ago•614 comments

RISC-V single-board computer for less than 40 euros

https://www.heise.de/en/news/RISC-V-single-board-computer-for-less-than-40-euros-10515044.html
131•doener•4d ago•75 comments

Fixing a loud PSU fan without dying

https://chameth.com/fixing-a-loud-psu-fan-without-dying/
22•sprawl_•3d ago•26 comments

The equality delete problem in Apache Iceberg

https://blog.dataengineerthings.org/the-equality-delete-problem-in-apache-iceberg-143dd451a974
47•dkgs•8h ago•23 comments

Evaluating LLMs playing text adventures

https://entropicthoughts.com/evaluating-llms-playing-text-adventures
94•todsacerdoti•11h ago•58 comments

Weave (YC W25) is hiring a founding AI engineer

https://www.ycombinator.com/companies/weave-3/jobs/SqFnIFE-founding-ai-engineer
1•adchurch•10h ago

Debian GNU/Hurd 2025 released

https://lists.debian.org/debian-hurd/2025/08/msg00038.html
190•jrepinc•3d ago•102 comments

Dumb to managed switch conversion (2010)

https://spritesmods.com/?art=rtl8366sb&page=1
40•userbinator•3d ago•17 comments

Galileo’s telescopes: Seeing is believing (2010)

https://www.historytoday.com/archive/history-matters/galileos-telescopes-seeing-believing
18•hhs•3d ago•7 comments

The Missing Protocol: Let Me Know

https://deanebarker.net/tech/blog/let-me-know/
81•deanebarker•7h ago•61 comments

Is Meta Scraping the Fediverse for AI?

https://wedistribute.org/2025/08/is-meta-scraping-the-fediverse-for-ai/
8•nogajun•1h ago•0 comments

Australian court finds Apple, Google guilty of being anticompetitive

https://www.ghacks.net/2025/08/12/australian-court-finds-apple-google-guilty-of-being-anticompetitive/
335•warrenm•13h ago•125 comments
Open in hackernews

Fixing a loud PSU fan without dying

https://chameth.com/fixing-a-loud-psu-fan-without-dying/
22•sprawl_•3d ago

Comments

mkayokay•2d ago
A new semi-passive 850-watt fully modular PSU is around EUR 130, the Noctua fan around EUR 30.

I guess if you know electronics and how to safely handle the PSU internals, the risk of injury is low, but I personally would not risk it for EUR 100.

Also, if the only problem was the noisy fan, I guess selling it used would have returned most of the investment, leaving him with like EUR 50 in added cost. Compared to the price of a modern gaming PC, that's nothing (also avoiding not risking your life).

jojobas•2h ago
C grade high school physics understanding makes the risk exactly zero. Selling defective stuff without declaring is bad mojo, with declaring it just delegates the pain in the bum to someone else..
doubled112•2h ago
I’ve been tearing things open after ignoring the “dangerous if opened” stickers since I was 8 years old. I’m in my 30s, you’d think something would have caused me harm by now, but no.
jojobas•1h ago
There's a fair bit of survivorship bias in this. That said, modern power supplies use "bleeder resistors" to discharge the capacitors once powered off.
scottlamb•1h ago
> There's a fair bit of survivorship bias in this.

Is there? This kind of statement has the potential to exhibit survivorship bias, but I feel like the opposite—"12-year-old dies replacing a power supply fan"—would make headlines. Definitely haven't seen that.

Capacitors should be respected for sure, but people don't routinely die in DIY electronics tasks.

jojobas•58m ago
Electrical injuries claim some 1000 lives per year in the states, and 20% of all electrical injuries are sustained by children (lethal and non-lethal). I don't think every electrocuted child makes even local headlines.
heelix•2m ago
As a child, I remember being up in the attic with my grandfather. He touched some wires, swore (which was unusual for him), and said 'that was 220'. I still have a healthy respect for power.
tcdent•2h ago
I will put hundreds of dollars in Noctua fans into a second-hand chassis without thinking twice.

It's sometimes uneconomical from a cost-ratio perspective, but it is crucial to making datacenter-grade equipment actually useable at home.

ahartmetz•2h ago
If you need pretty good fans for cheap as dirt, there is also Arctic Cooling.
PaulKeeble•2h ago
They just never last. Arctic fans perform really well especially for their price but they all seem to develop problems. I have probably bought about 15 different Arctic fans from the F and P range and none of them survived 5 years, most were dead or developed noise within 1-2 years. Noctua on the other hand the old 80mm fan from the early 2000s still works just fine and remains quiet. Noctua fans are crazy reliable, they cost more too but I would suspect over the life of the fan they end up similar priced or cheaper.
ahartmetz•1h ago
Arctic fans seem to last 10+ years for me. I haven't bought any recently though and I run them at 500 rpm or so, that's my strategy: many and large fans, running slowly.
aaronmdjones•2h ago
I have a 12U 19 inch rack built into my computer desk, and I have a couple of NASes in it (2x HPC-8316SA-55RB1). The 40mm fans in the included CRPS PSUs are loud, whiny, and rattly, all at the same time.

I replaced all 4 of them with Noctua NF-A4x20s, wired to run at full speed all the time. They still report their speed so the IPMI management interface doesn't consider the power supply fan to have failed, but the PSU can no longer control the fan's speed.

The PSUs don't run any hotter and I can't hear them now.

I have a used Eaton PW9130 UPS in the bottom of the rack. The 80mm (exhaust) and 60mm (inverter heatsink) fans were likewise louder than I'd like. I replaced them with Noctuas too, again wired to run at full speed all the time, and the UPS' Web/SNMP card confirms it's still no hotter than 30'C internally. I can't hear that now either.

Hilariously, the most critical fan, the original inverter heatsink fan, is a 2-pin fan, so it probably can't even detect when it has failed (unless it's detecting fan failure by monitoring current consumption). The original rear exhaust fan uses a locked rotor sensor rather than a tachometer, which required a bit of bodging to convince the UPS that it has not failed. Oh well.

ydj•27m ago
I’ve also done a noctua fan replacement on my ups. My worry is that they are rated for lower airflow than the original fans they replaced. Have you checked whether it stays cool when running on battery?
hobs•2h ago
Hilarious to read this on "hacker news" - replacing a standard fan is too dangerous! This is not styropyro.
sonofhans•2h ago
An important side note is that Cooler Master’s “10-year Warranty” is garbage:

> I looked at Cooler Master’s warranty, and for issues within the first two years you have to deal with the retailer. That would be Amazon in my case. So I looked at Amazon’s information on warranty issues. Their policy is that if it’s more than 30 days since purchase, you have to send it off to a third-party repair center and wait for them to diagnose and try to repair it. Here’s the kicker: Usually repairs take up to 20 business days (including delivery time), but could take slightly longer

Everyone knows electronic devices tend to die early or last forever. Cooler Master and Amazon are working a shitty dodge here, and I bet they avoid most DOA warranty claims because of it.

doubled112•2h ago
I ran into this same setup with a microwave that failed in a few months after purchase and Walmart.

Tried to deal with the manufacturer, but they couldn’t help and sent me to the retailer.

Went to the store, popped it up on the counter, had a short conversation and got the expected “you have to deal with the manufacturer”.

Is there anything else I can help you with today?

Actually yes. Can you throw that out for me?

The confusion on the guy’s face was great.

Spent more than enough time on the $100 microwave. Their problem now.

jakeinspace•1h ago
My experience with cheap modern microwave failures has been the door sensor failing, which for safety ofc prevents the magnetron from running. I had one fail in about 2 months, thankfully fixable with a $10 sensor and 15 minute of work. Same goes for a lot of appliances, repaired a dryer that had its door sensor fail (in fact, they all tend to use identical door sensors as far as I've seen, dryers and washing machines and microwaves).
PaulKeeble•1h ago
I am pretty certain that the G2 Noctua fans will start at 5V and they also in the 4 pin model will turn off at 0-19% and turn on at 20%. Its a bit of a shame Noctua don't make 2 pin fans anymore but it is possible on the latest most efficient models to get the low start voltage.

The G1 Noctua fans don't, I have a 120mm still in my case that is a decade or so old and it starts at a bit under 7V, but all the G2 140mm will start at 5V IIRC the Noctua presentations on this. Wish it was laid out in the specifications however.

userbinator•1h ago
I've "fixed" countless noisy computer fans over the years with simple lubrication. Peel off the label, remove the bung, pry out the circlip and pull the rotor + blade assembly out, clean off any remaining old grease/oil, and then pack the bushing with petroleum jelly before reinserting the shaft. Reassembly is the reverse of disassembly.
serf•1h ago
doing this on a fan that used a dry lubricant can burn a wire lead from increased startup torque requirements to get it spinning.

it's not bad advice but it's worth noting that petroleum jelly is pretty thick compared to a machinists' oil or dry graphite.

stavros•1h ago
I have a personal rule that I only follow instructions when I understand more than 10% of the words in them, so I'm buying a new fan.
phanimahesh•1m ago
That's a good personal rule to copy
cmurf•1h ago
Check your credit card agreement. Some cards have a "purchase return protection" for when a retailer won't take back a product.

It's kindof a gimmick because it's easy to forget you have it, but if you have it, use it.

ziml77•1h ago
I see they have the exact same frustration with the results of trying to research how to safely work on a PSU as I've had. I know there's plenty of info out there about working on CRTs; are the caps in PSUs really more dangerous than that?
duskwuff•33m ago
Modern computer power supplies are far less dangerous than CRTs were.

CRTs were dangerous because they used extremely high voltages (often as high as 25 kV), and that voltage was often present on the CRT itself, which acted as a capacitor. The capacitance of the CRT was relatively low, but it still stored enough energy to be very hazardous. There's ways to safely discharge them, but you absolutely need to know what you're doing and work with one hand behind your back.

The voltages in a modern switching power supply are much lower (typically no more than ~1.5x line voltage ≈ 180V), and the capacitors usually have "bleeder" resistors which will discharge them to safe voltages within a few seconds. I'd still give one a few minutes before touching it - and maybe check one or two of the bigger capacitors with a multimeter - but it's still much less dangerous.

userbinator•2m ago
The main danger with CRT HV is not the HV itself but what the shock will make you do, such as jump and cut your hand on something sharp in the chassis, or even crack the neck of the tube, or fall and hit yourself. There is very little current available, especially if the set has been off, but it's comparable to a very painful static shock. Instead, the "low" voltage B+ supply is far more dangerous when the set is on, as that is basically rectified mains.