News for hackers. Stuff that matters.

markrages

Re:Previewing GPT‑5.6 Sol: a next-generation model on Friday June 26, 2026 @05:10PM (#48692004)
I can't be the only one whose memory is so bad that I am like this in my own code base.
Faster than some, slower than some.

PCBs come in all different levels of complexity.

Re:The Serial TTL connector we deserve on Saturday May 09, 2026 @10:37PM (#48080477)
RFCOMM is old Bluetooth, not BLE.

Nordic has a serial-over-ble profile, but it is not standard like RFCOMM.

Re:Why can't you tune your guitar? on Sunday March 08, 2026 @06:50PM (#47302460)
GP is talking about seeing the string oscillations alias against the 30 Hz camera frame rate.

I've never tried it.

Why not? It's been a decade away for the past 20 years.
"defining them right before the dependency is actually used"
You have always been able to do the same thing in Python. This PEP isn't needed for that functionality.
Re:The scariest "user support" email I've ever received on Monday October 20, 2025 @05:38PM (#45649730)
ChatGPT doesn't run commands, does it?
Re:Why does a fire truck cost $2m on Sunday July 27, 2025 @09:51PM (#44706452)
Vacuum people have been trying to normalize high prices for a long time... From the old Internet: https://www.cockeyed.com/citizen/kirby/kirby.html
I think a lot of real reviews are written by ChatGPT. People are lazy!
Re:HDR‑Infused Emoji on Friday April 18, 2025 @12:58AM (#43725155)
The Loudness War has come to Slack.

https://en.wikipedia.org/wiki/Loudness_war

...for now.
Re:Electron Band Structure in Germanium, My Ass on Tuesday April 01, 2025 @04:08PM (#43550832)
ULN2003 is not DIP8

And I would assume for pedalogical purposes a bare transistor would be preferred rather than the '2008 with its extra diodes and base resistor.

Re:War story: the hardest bug I ever debugged on Thursday March 27, 2025 @09:37AM (#43493505)
Of course the comments section is going to be full of war stories about everyone's hardest bug.

This is how humans work, and this is why I am reading the comments.

Re:The Lost Art of Logarithms on Thursday March 13, 2025 @05:49PM (#43357560)
Thanks. What a weird API.
Re:The Lost Art of Logarithms on Thursday March 13, 2025 @04:42PM (#43357048)
shouldn't it be

var actualSize = 1 << Integer.highestOneBit(approxSize - 1);

?

Re:Blocklist for AI Music on YouTube on Sunday February 16, 2025 @02:11PM (#43070729)
> The statistical likelihood is that other civilisations will arise. There will one day be lemon-soaked paper napkins.

> ‘Till then, there will be a short delay. Please return to your seats.

> Is there any precedent to car carriers dedicated to one manufacturer?

General Motors helped design the Vert-A-Pac. https://chevyvega.fandom.com/wiki/Vert-A-Pac

Re:The Begining of the End for ANT+ Wireless on Saturday January 04, 2025 @08:37PM (#42599024)
> I am sure is not a fundamental but is the reality for the currently available chipsets.

It is pretty fundamental. Ant has an inverted master/slave (or whatever we're calling it nowdays) relationship. In Ant, the sensor determines the timing, and can broadcast to many receivers. In Bluetooth, the central device (phone) determines the timing, and each sensor connects to one central with a one-to-one connection.

There are ways around this limitation of BLE:

1. A few bytes of data can be stuffed in BLE advertising, so the sensor can communicate without a connection in the Ant style. To my knowledge, none of the Ant+-replacing profiles support this.

2. The sensor can basically run multiple instances of bluetooth stack at the same time to connect to multiple central devices. This basically doubles the resource usage, and good luck determining if your sensor supports this without trying it.

#2 appears to be the path forward. A few sensors support it already, and the next generation of radio SOCs will make the resource requirements less onerous.

Re:The Begining of the End for ANT+ Wireless on Saturday January 04, 2025 @08:27PM (#42598979)
Ant also broadcasts a unique(-ish) identifier. This was enough to end Nike+. https://en.wikipedia.org/wiki/Nike%2BiPod#Wireless_protocol_...
Re:The Rules of Programming on Sunday December 08, 2024 @05:26PM (#42361248)
The Amazon preview included a page advocating for systems Hungarian in naming variables.

I am no longer interested in the book.

In the 90s I worked for a company in Kansas City designing pagers. We rented space underground for RF testing. There are no radio signals in the caves, at least back then. I'm sure that's different now, especially with manufacturing happening there. (Our cave was closer to downtown and was mostly used for warehousing.)
Re:Reformatting 100k Files at Google in 2011 on Saturday June 15, 2024 @10:57PM (#40694354)
Yes! Code is for people to read, not computers. Beautiful code uses the full expressiveness that the language allows.
Re:The Simple Beauty of XOR Floating Point Compression on Thursday April 11, 2024 @04:49PM (#40006666)
For integer values, just subtracting bytes is enough to allow much better compression ratios. See the table at https://opensource.quarq.com/zfit/zfit.pdf
Does the output work with patch(1)? Or does this use a different patch?
Re:What does Alan Kay think about LLMs? on Tuesday March 19, 2024 @03:14PM (#39759225)
Even with these in place, the output of the LLM is not trustworthy. In the sense that it doesn't care what it true, only what is plausible.
Re:I Accidentally Deleted a Game from MAME on Friday March 01, 2024 @09:26PM (#39569248)
For problem with mechanical interference, the old school trick was to stack a few DIP sockets together and plug into those. These would space the adapter up off of the reader. (We used to use extra sockets to save wear on the chip pins during the program/burn/crash/debug cycle.)

Older »