dllu

6 ms·

1.
by dllu 6d ago
I found it interesting that Thorsten mentions the importance of attaching screenshots. In my experience, it varies widely by workload. Vision is great for some UI stuff but extremely terrible for my work, which involves visualizing connecti
2.
by dllu 7d ago
I stand corrected. However, fundamentally, the idea of "tiny CPU-only neural network" combined with traditional alpha-beta search is substantially different from "big GPU network" combined with Monte Carlo Tree Search. A
3.
by dllu 7d ago
The Stockfish NNUE is completely unrelated to AlphaZero.
4.
by dllu 7d ago
I recently scored an amazing deal on the Samsung Odyssey G8 G80HS, which is 32", 6144 x 3456 at 165 Hz for only $700 via perks at work [1]!!! Even if computers typically cannot game at 6144 x 3456 px, it is incredibly nice to simultane
5.
by dllu 24d ago
The windshield wiper controls are a physical button plus physical scroll wheel on the steering wheel. You don't have to touch the touchscreen.
6.
by dllu 26d ago
I feel like taking pictures is a fun experience in itself. For me, it isn't really "missing the experience". If anything, it makes me pay more attention to the amazing view and details that I'd have otherwise missed as I
7.
by dllu 26d ago
I travel to tourist sites to take the exact same photo that is taken thousands of times every day because I have an above-average camera so I can take better quality photos than most other people lol
8.
by dllu 1mo ago
As someone who also did some line scan photography, I also ran into a lot of similar challenges! > Displaying and sharing the pictures I've taken has been a pain throughout the project. Most software on my computer doesn't like
9.
by dllu 1mo ago
I've further improved my long trains scanning setup with: * portable battery-powered enclosure with display: https://daniel.lawrence.lu/blog/2026-06-12-standalone-cnc-li... * fine-grained distortion correction by
10.
by dllu 1mo ago
Neat. It links to some of mine from an older HN comment but here's another one of some houses from Caltrain: https://i.dllu.net/2017-11-23-14-30-54_stack_cdb771693eea9fe...
11.
by dllu 1mo ago
No pre-prompting, although I can't be sure it didn't use memories. "No tools" should theoretically have prevented it from looking up memories. FWIW, Grok and Gemini both failed in a similar way. With Python, it was able
12.
by dllu 1mo ago
Vision is still embarrassingly bad. ChatGPT Pro with GPT 5.6-sol: https://chatgpt.com/share/6a834217-ca8c-83e8-a8e8-45d5b8797b... The puzzle: https://activityvillage-files.s3.eu-west-2.amazonaws.com/s3f
13.
by dllu 1mo ago
In terms of using experience, I found Grok 4.5 to be way more pleasant to use than GPT 5.6 Sol and Claude 4.8/5. It just gets to the point, and is super fast and concise, no yapping. That's how AI agents should be imo. None of the
14.
by dllu 1mo ago
Talking about being "tired", the phrase お疲れ様でした (otsukaresama deshita) is commonly used to express "thanks for the hard work" and it's a catch-all greeting for coworkers now, but it literally means something like &q
15.
by dllu 1mo ago
Japanese also has a word you say to the cook, ご馳走様でした (gochisousama deshita) but that's said after the meal as thanks.
16.
by dllu 1mo ago
The Smithsonian has a great book called "Wheels and Wheeling" that covers the early history of bicycles. Fascinating stuff. https://repository.si.edu/server/api/core/bitstreams/02b8e94...
17.
by dllu 2mo ago
https://simonwillison.net/2026/Jun/9/claude-fable-5/ The bike is generally okay, apart from medium which derped hard. Max has correct diamond, correct head tube, and so on. Only nitpicks are that the fro
18.
by dllu 2mo ago
Agreed. It's far from solved. Modern LLMs still generate pelican bike SVGs with obvious errors: * some omitted the bottom of the diamond which connects from the pedals to the rear wheel * some added an extra connection from the pedals
19.
by dllu 2mo ago
Original founder of darktable, hanatos, went and made vkdt which fixes a bunch of my pet peeves with darktable: https://github.com/hanatos/vkdt https://jo.dreggn.org/vkdt/ * Vulkan instead of Open
20.
by dllu 2mo ago
I have a pet peeve that darktable-cli requires the GUI darktable to not be running. It would be nice to export photos using my script without having to disrupt my editing workflow...
21.
by dllu 2mo ago
I use Darktable a lot for 10 years. It has improved vastly over the years. I love the new additions like AgX! However I tried out the new AI denoise and it still isn't great. I also wish there was a way to manually correct barrel disto
22.
by dllu 2mo ago
iirc the bug that the R8127 patch was addressing was fixed in the R8169 driver upstream in kernel 6.18 and later. Not super sure what other patches there are and what the status of those are...
23.
by dllu 2mo ago
The Ampere cpu in that post has a "lack of single core CPU speed" despite ripping through compilation tasks with its 80 cores. But the DGX Spark's CPU has fairly decent single core speed and is more like Apple Silicon in this
24.
by dllu 2mo ago
It's an Orico 9948C3 with four Seagate Barracuda 24TB drives. They were on sale last year [1]. Unfortunately, the enclosure doesn't work super well on Linux. There is a weird bug where the drives don't enumerate when I boot u
25.
by dllu 2mo ago
Sorry, my comment was confusingly worded, but I am precisely thinking about the head-space thing you are talking about. My point is: the fact that image generation models can generate a perfect pelican riding on a topologically correct and
26.
by dllu 2mo ago
The DGX Spark hits a sweet spot for me where it can simultaneously work for general computing and run local LLM inference fast enough for some hobbyist dabbling. The RTX 6000 Pro Blackwell is more than twice the price (it has increased quit
27.
by dllu 2mo ago
Many other distros almost work out of the box (as in they boot and run without any modifications). The custom kernel patches you mentioned address mainly non-critical bugs such as a bug where the Realtek r8127 stops working after a reboot (
28.
by dllu 2mo ago
I know. I'm just thinking about how to make AI create SVGs better... in theory, a sufficiently smart AI could "generate an image in its head", think about it, and then output the SVG paths to produce said image. Intuitively t
29.
by dllu 2mo ago
I disagree with this in the blog post: > Every single one is a pelican, on a bicycle, first try. When every student gets an A, the exam has stopped grading. Numerous pelicans and their bikes are clearly horribly malformed. In fact none o
30.
by dllu 2mo ago
I feel like getting LLMs to spit out an SVG is akin to getting a human artist to draw something by just reciting a list of coordinates. It's insanely hard and unnatural. Image generation models nowadays can easily generate a photoreali

More ›