fred123

It’s just $20, that’s almost free compared to cost of human labor

With an LLM integrated into your IDE like Cursor or Copilot, oftentimes the LLM autocompletes the correct code faster than I can think about what must be done next. I’ve been coding for 15 years.

I feel like text detection is much better covered by the various ML models discussed elsewhere in the comments. Maybe you can combine those with Live Text. I found Tesseract pretty ok for text detection as well but I don’t know if any of the models are good for vertical text.

Fair enough. There are some new OCR APIs in the next macOS release. I wonder if the model has been improved.

You can get them through the Vision API (Swift/Objective-C/AppleScript)

It seems to struggle with German text a lot (umlauts etc)

Azure Vison OCR is supposed to be the best commercial OCR model right now and it’s really cheap (same price as Google‘s)

IIRC Tesseract is trained on 300 DPI

Something wrong with your setup. It should be less than 30 s per page with your hardware

macOS Live Text is incredible. Mac only though

https://en.m.wikipedia.org/wiki/Extraversion_and_introversio...

sharing state between threads is such a narow niche use case

It is the norm. „Kafka scale“ problems are not the norm.

Do you think Meta (Instagram) are pushing GIL removal and Cinder for no reason? They clearly have that scale and still benefit from faster single machine performance

Why? On AWS you can rent a 24 TB, 500 core machine. Almost all problems are smaller than that so don’t need to scale to more than one machine.

Building applications that run on multiple machines is at least one order of magnitude more complex and thus slower (in development velocity), so needlessly building an application to work distributedly is just bad engineering.

In most parts of Europe the standard room socket has >= 3.6 kW

More like 20x actually, eg Germany has a homicide rate of 0.8 (and 0.3 if you don’t take attempted homicides into account)

I’d consider that dangerous when you compare it to the ~10x lower rate in Western Europe

SIMD, multithreading, better handling of NaN, maybe other factors.

Or just use Conda/Mamba.

You should provide a way for people to reach out to you privately eg email. If you have any public profile/CV that helps as well.

Somewhat related: I built a library to speed up matching many regexes with mostly mismatches by adding non-regex pre matchers. https://github.com/Quantco/multiregex

Where are you based?

SSDs and other storage drives have two layers (or more). The last layer is stable storage (= when you disconnect power no data is lost or corrupted). When you write to such a device your writes are first made in an earlier layer that is more like your computer’s main memory than actual storage (when you lose power your data is gone or corrupted). Only after time or when the cache is full an actual persistent write is made.

What’s your expertise and technology stack?

100€ is easily possible, I think it is kind of the default rate in my space (Python Data Engineering, cloud stuff, …).

From my experience there are almost exclusively very professional people on those platforms so no low balling.

Typical contract will be 4-5 days a week for 3-18 months

Actually I think chapters 1–8 are more valuable than the chapters dealing with a concrete implementation because they teach you ideas and principles

Out of the tar pit is the best paper on this I’ve ever read: http://curtclifton.net/papers/MoseleyMarks06a.pdf

Bjoern author here.

The code looks very nice, looks like it’s inspired by bjoern but much cleaner. I’ve wanted to clean up the mess that the bjoern code is for years but never came around to actually doing it.

That being said I don’t think it provides a lot of value in practice. Same with bjoern. With a reasonably fast server implementation around 99% of time is spent in your Python application, not the server. So it doesn’t actually provide much value in practice to optimize the server. But it’s a nice project to learn about how to write a HTTP and WSGI server :)

Reported by the browser but not always accurate. Used in browser fingerprinting btw

Captcha submit requires two touches on iPhone 6S because the submit button will scroll out of view on first touch.

Do your devs or product people ever use your own product? That’s one of the first things I’d notice and fix.

Yes. Noise suppression is very similar to speech separation (separating multiple speaker voices that talk at the same time). For example you can use ConvTasNet for both speech separation and denoising; in the denoising case you set target track 1 = speech, track 2 = noise, hence you get a noise-only track.

I guess you can also simply subtract the clean speech from the original mixture to get the noise-only track.