sorenjan

7 ms·

61.
by sorenjan 6mo ago
Can you expand on the Gemini tagging part? What did you do with the tags, import them into Jellyfin after cutting the videos into parts?
62.
by sorenjan 6mo ago
They're using the truncated SVD, not the full variant, that's computationally cheaper.
63.
TinyLoRA – Learning to Reason in 13 Parameters(arxiv.org)
234 points by sorenjan 7mo ago|45 comments
64.
by sorenjan 7mo ago
The cropdetect example made me wonder if they're thinking about including support for yolo or similar models. They're including Whisper for text to speech already, I think yolo would enable things like automatic face censoring and
65.
by sorenjan 7mo ago
I've never tried doing frame perfect clips like that, that does sound annoying. But from a cursory read of the source, I don't think this program will solve that issue either? Because the time stamps in your examples are all corre
66.
by sorenjan 7mo ago
Missed opportunity to reference the famous Dropbox hn comment. I just think there are other closely related use cases where a separate program can add more value, especially in the terminal. I wouldn't suggest most people should use ff
67.
by sorenjan 7mo ago
I disagree, I don't want another ffmpeg binary, I already have one. Winget works well, especially since this is already a terminal program.
68.
by sorenjan 7mo ago
I don't find trimming videos with ffmpeg particularly difficult, is just-ss xx -to xx -c copy basically. Sure, you need to get those time stamps using a media player, but you probably already have one so that isn't really an issue
69.
by sorenjan 7mo ago
Both Russia and Ukraine build millions of drones per year, most of them fpv drones that are basically remote controlled flying grenades. There's plenty of electronic warfare with radio jamming, so in some places they use drone mounted
70.
by sorenjan 7mo ago
> uvx --with pillow --with okmain python -c "from PIL import Image; import okmain; print(okmain.colors(Image.open('bluemarble.jpg')))" [RGB(r=79, g=87, b=120), RGB(r=27, g=33, b=66), RGB(r=152, g=155, b=175), R
71.
by sorenjan 7mo ago
I wonder if one of the LLMs could generate code from a screenshot of a layout designed by this.
72.
by sorenjan 7mo ago
You can right click on it and choose "Show controls", at least in Firefox.
73.
by sorenjan 7mo ago
Planet labs has a solution specifically for ships. https://www.planet.com/pulse/illuminate-the-dark-fleet-with-...
74.
by sorenjan 7mo ago
A flathead screwdriver is not a valid analogy, because LLMs are big complicated and opaque machines. And while other ML methods are non-deterministic as well, gaussian process, decision trees or even CNNs are easier to try to make sense of
75.
by sorenjan 7mo ago
Seems like a way to use a sledgehammer to hammer in screws, and inviting nondeterminism in important systems. Besides being way larger and more complex than what most specialized industrial processes need, they are also vulnerable to advers
76.
by sorenjan 7mo ago
The discussion was about fine-tuned Qwen models, not industrial inspection in general. I would also find it interesting to learn about what kind of edge AI industrial inspection task you could do with fine-tuned llms, not some handwavy answ
77.
by sorenjan 7mo ago
But that's not something you'd use an LLM for. There have been computer vision systems sorting bad peas for more than a decade[0], of course there are plenty of use cases for very fast inspection systems. But when would you use an
78.
by sorenjan 7mo ago
Right, but where I live sunrise is in the middle of the night in the summer (around 03:30). Using standard time in the summer gives me one less hour of useful sunlight in the evening, and while it doesn't technically disappear it gets
79.
by sorenjan 7mo ago
They worked best when everybody were farmers and had to get up early and go to bed early. Now most people don't live their lives centered around noon, our free time comes after our work is done at around 17:00, so having more light in
80.
by sorenjan 8mo ago
The maintainer is the first comment, agreeing with the post author: > I concur, but there are two issues with this: Building the foundation of an editor is way harder than building features later on top of that, and currently it is unlik
81.
by sorenjan 8mo ago
Is this a distillation of Nano Banana Pro?
82.
by sorenjan 8mo ago
It hasn't gotten to that point yet, but the maintainer of edit wants to add a lot of features to it. Make Microsoft Edit a Lightweight 'VS Code for the Terminal' - https://github.com/microsoft/edit/d
83.
by sorenjan 8mo ago
Julia Turc recently did a video about diffusion LLMs as a paid collaboration with Inception: https://www.youtube.com/watch?v=-VGeHZqOk_s
84.
by sorenjan 8mo ago
Wood and drywall is how most houses in Sweden are built as well. Here's a timelapse of a Swedish house being built: https://www.youtube.com/watch?v=KbSm0Zw00Cs
85.
by sorenjan 8mo ago
If your main gripe about the Rust GUI ecosystem is that it's not mature then rewinding 20 years and using Qt/WPF/etc sounds like an excellent alternative. Old and mature versus modern and immature.
86.
by sorenjan 8mo ago
> I also live in Europe and tip, specially if I know that the salary of the staff heavily depends on tips. > If I ever find the system too unfair for the workers, then I won’t go to those restaurants anymore. Sounds like you only tip
87.
by sorenjan 8mo ago
Maybe you can log all the traffic to and from the proprietary models and fine tune a local model each weekend? It's probably against their terms of service, but it's not like they care where their training data comes from anyway.
88.
by sorenjan 8mo ago
It's always the next big thing. It used to be self driving, now it's AI and robots.
89.
by sorenjan 8mo ago
This is not true, 5G has multiple positioning improvements that are not related to higher frequencies. 5G has something called LMF (Location Management Function) that handles positioning of user clients through multiple means, like round tr
90.
by sorenjan 9mo ago
Norway has a slightly higher GDP (PPP) per capita than Switzerland, so saying it's just because they're rich enough to buy gas-guzzling G-wagens doesn't seem to be the answer.

More ›