frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

SectorC: A C Compiler in 512 bytes

https://xorvoid.com/sectorc.html
98•valyala•4h ago•16 comments

The F Word

http://muratbuffalo.blogspot.com/2026/02/friction.html
43•zdw•3d ago•11 comments

Brookhaven Lab's RHIC concludes 25-year run with final collisions

https://www.hpcwire.com/off-the-wire/brookhaven-labs-rhic-concludes-25-year-run-with-final-collis...
23•gnufx•2h ago•19 comments

Speed up responses with fast mode

https://code.claude.com/docs/en/fast-mode
56•surprisetalk•3h ago•55 comments

Software factories and the agentic moment

https://factory.strongdm.ai/
98•mellosouls•6h ago•176 comments

Hoot: Scheme on WebAssembly

https://www.spritely.institute/hoot/
144•AlexeyBrin•9h ago•26 comments

Stories from 25 Years of Software Development

https://susam.net/twenty-five-years-of-computing.html
101•vinhnx•7h ago•13 comments

OpenCiv3: Open-source, cross-platform reimagining of Civilization III

https://openciv3.org/
851•klaussilveira•1d ago•258 comments

I write games in C (yes, C)

https://jonathanwhiting.com/writing/blog/games_in_c/
139•valyala•4h ago•109 comments

First Proof

https://arxiv.org/abs/2602.05192
69•samasblack•6h ago•52 comments

The Waymo World Model

https://waymo.com/blog/2026/02/the-waymo-world-model-a-new-frontier-for-autonomous-driving-simula...
1094•xnx•1d ago•618 comments

Show HN: A luma dependent chroma compression algorithm (image compression)

https://www.bitsnbites.eu/a-spatial-domain-variable-block-size-luma-dependent-chroma-compression-...
7•mbitsnbites•3d ago•0 comments

Al Lowe on model trains, funny deaths and working with Disney

https://spillhistorie.no/2026/02/06/interview-with-sierra-veteran-al-lowe/
64•thelok•6h ago•10 comments

Vocal Guide – belt sing without killing yourself

https://jesperordrup.github.io/vocal-guide/
235•jesperordrup•14h ago•80 comments

Start all of your commands with a comma (2009)

https://rhodesmill.org/brandon/2009/commands-with-comma/
519•theblazehen•3d ago•191 comments

Reinforcement Learning from Human Feedback

https://rlhfbook.com/
94•onurkanbkrc•9h ago•5 comments

Show HN: I saw this cool navigation reveal, so I made a simple HTML+CSS version

https://github.com/Momciloo/fun-with-clip-path
31•momciloo•4h ago•5 comments

Selection Rather Than Prediction

https://voratiq.com/blog/selection-rather-than-prediction/
13•languid-photic•3d ago•4 comments

Coding agents have replaced every framework I used

https://blog.alaindichiappari.dev/p/software-engineering-is-back
259•alainrk•8h ago•425 comments

A Fresh Look at IBM 3270 Information Display System

https://www.rs-online.com/designspark/a-fresh-look-at-ibm-3270-information-display-system
49•rbanffy•4d ago•9 comments

The AI boom is causing shortages everywhere else

https://www.washingtonpost.com/technology/2026/02/07/ai-spending-economy-shortages/
187•1vuio0pswjnm7•10h ago•268 comments

France's homegrown open source online office suite

https://github.com/suitenumerique
615•nar001•8h ago•272 comments

72M Points of Interest

https://tech.marksblogg.com/overture-places-pois.html
36•marklit•5d ago•6 comments

We mourn our craft

https://nolanlawson.com/2026/02/07/we-mourn-our-craft/
348•ColinWright•3h ago•416 comments

Unseen Footage of Atari Battlezone Arcade Cabinet Production

https://arcadeblogger.com/2026/02/02/unseen-footage-of-atari-battlezone-cabinet-production/
125•videotopia•4d ago•39 comments

Where did all the starships go?

https://www.datawrapper.de/blog/science-fiction-decline
99•speckx•4d ago•117 comments

Show HN: Kappal – CLI to Run Docker Compose YML on Kubernetes for Local Dev

https://github.com/sandys/kappal
33•sandGorgon•2d ago•15 comments

Learning from context is harder than we thought

https://hy.tencent.com/research/100025?langVersion=en
211•limoce•4d ago•119 comments

Show HN: Look Ma, No Linux: Shell, App Installer, Vi, Cc on ESP32-S3 / BreezyBox

https://github.com/valdanylchuk/breezydemo
288•isitcontent•1d ago•38 comments

History and Timeline of the Proco Rat Pedal (2021)

https://web.archive.org/web/20211030011207/https://thejhsshow.com/articles/history-and-timeline-o...
20•brudgers•5d ago•5 comments
Open in hackernews

Powerful and precise multi-color lasers now fit on a single chip

https://phys.org/news/2025-10-powerful-precise-multi-lasers-chip.html
71•PaulHoule•3mo ago

Comments

hdjfjkremmr•3mo ago
could you use this in show lasers? currently they use RGB mixing with electro-acoustical crystals for intensity modulation.
brookst•3mo ago
Maybe? Show lasers are much more than 150mw. Lasers can be combined but I’m not sure the practicality of combining 100 chips to get 15w.
mezzman•3mo ago
Polychromatic acousto-optic modulation hasn't been used in laser shows for quite some time since RGB diode based laser systems came about. Granted, nothing beats a mixed gas ion laser and PCAOM for the beautiful colors you can get but these days nobody misses dragging around water hoses and sorting out 60 Amps of three phase power to run those old beasts.
pppone•3mo ago
In astronomy, laser frequency combs are horribly expensive (~$0.5M), but fantastic for calibrating high precision spectrographs. It would be interesting to see if this method could be tuned for that application (namely, shifting to the visible), such to enable cheaper spectrographs.
casparvitch•3mo ago
Relevant: https://www.combs.org.au/astrocombs/
mmmBacon•3mo ago
Visible will always be expensive because it’s very niche and low volume. So the techniques here are only practical economically for the large volumes of light sources required for communications. This won’t extend to the visible unless there’s a similarly large market.

The cheapest way I’d think to generate a visible frequency comb would be to frequency double the IR comb laser using a nonlinear crystal like BBO.

Also here the accuracy is relative and not absolute which is fine for communications. The absolute accuracy of the comb may not good enough for spectroscopy in the visible.

bobsmooth•3mo ago
This seems like the kind of technology that will quietly revolutionize a lot of things in 10 years when manufacturing is figured out.
pjc50•3mo ago
It sounds like it's already manufacturable - silicon photonics uses the IC manufacturing process, in the same way that MEMS does.
bobsmooth•3mo ago
Sure, but can they make 10 million of them? I really hope they can. Tiny terabit transceivers sounds awesome.
danw1979•3mo ago
How would you modulate the individual wavelengths, considering they all come from the same source ?

I had, maybe naively assumed that laser diodes were switched on/off electronically to modulate a signal. With this laser you’d have to modulate after the light source somehow ?

khalic•3mo ago
You can always filter the frequencies you don’t need
jagged-chisel•3mo ago
The question remains: how to modulate individual wavelengths.
bartlettD•3mo ago
Likely some kind of Electro-Optic Modulator. You could use their wavelength comb to separate the light and then use a Mach–Zehnder interferometer to perform On-Off-Keying as an example
IAmBroom•3mo ago
There are no "individual wavelengths", ever, anywhere.

Do you mean modulating multiple bands of light, as a TV does with broad-bands of R, G, & B? Do you mean time-band modulation of a single band, like radios do with AM?

pjc50•3mo ago
Actual paper: https://www.nature.com/articles/s41566-025-01769-z DOI https://doi.org/10.1038/s41566-025-01769-z

" We show microcombs with total on-chip power levels up to 158 mW and comb lines with an intrinsic linewidth as narrow as 200 kHz."

150mW is a lot for a single-chip laser, given that the eye safety limit for standard red laser pointers is about 5mW.

_joel•3mo ago
"Beyond data centers, the same chips could enable portable spectrometers"

Tricorders ftw

IAmBroom•3mo ago
OMG, genuine tricorders, and not just some kluge of a few common tools!
fuzzfactor•3mo ago
Notice the breakthrough was accomplished in a lab which is utilizing more square-footage of "bench" space (/"shelf" space) compared to floor space than most other labs you will find.

Almost like a storage room, except with as much operational, calibrated equipment at the fingertips as the working room would possibly fit.

Regardless of the essential auxiliary storage space having at least 5x the square footage of the working lab itself. Where hopefully at least 20% of the equipment there is operational, if not currently calibrated or in use. Which would then equal the amount in operation in the lab.

If the storage area is down the hall, or maybe in the basement, or a convenient nearby building, the same breakthroughs will be possible by the same researchers.

It will just take more time the further the storage area is, and the more pieces of equipment for which there is no backup in storage.

And way more time if at all, when the storage area is too small to get the job done.

Anything less and you're shooting yourself in the "footage" :)

>Cleaning up messy light

Or cleaning up your messy lab, you can have both, you just have to prioritize what you want to accomplish more of in your lifetime.

PaulHoule•3mo ago
Reminds me in grad school doing the Fizeau experiment

https://en.wikipedia.org/wiki/Fizeau%27s_measurement_of_the_...

in the Physics 510 lab, the idea is that you send light through a system of mirrors through such a long path that a slot in a rotating disc (or a mirror) can move enough to block the light or maybe not block the light if it can rotate all the way to the next opening. Unlike Fizeau we did it entirely indoors and the experiment depends on empty space.

mojomark•3mo ago
Ha, I wasn't going to read the article, but I had to after reading this comment. Yikes dude - I hope you don't ever happen upon my messy-ass/inefficient lab!
fuzzfactor•3mo ago
It's supposed to be obvious I would not be in favor of removing what others often refer to as "clutter" :)

The aux storage areas are not where the overlaid layers of equipment on and around the benches would be as immediately useful. Plus, most importantly the storage rooms are supposed to already be full to the gills so you can hardly walk inside them ;) Containing equipment from which thousands of hours of learning has been gleaned beforehand.

In both, you often need as much stuff squeezed into a small space as possible before you can come near the goal line.

I'll take scientific progress that's good enough to emerge from the lab over a "clinical" appearance of the lab itself any day.