nonoesp

One step closer to Black Mirror's 'Common People'

https://en.wikipedia.org/wiki/Common_People_(Black_Mirror)

git update-index --assume-unchanged path/to/file

https://git-scm.com/docs/git-update-index

Which company will win the AI race is anybody’s guess. Regardless, OpenAI’s embrace of the Anthropic business model makes one thing abundantly clear: For all the wonder and change that generative AI brings as a technology, there hasn’t been any real innovation in the business models of Silicon Valley. For decades, most tech companies have succeeded by either selling ads or selling enterprise tools. One day OpenAI or Anthropic might cure cancer and remake the world, but for now they still have to pay the bills.

First thing that came to mind was git butler.

What's moon plus sun?

"Monsoon," says ChatGPT.

I create a playlist for each month (e.g., 2026.01) inside of a folder of the corresponding year (e.g., 2026) and add new tracks there (instead of using Spotify likes). This allows me to jump back to a specific year/month. You can also hit play on the folder to play all songs from a specific year.

I used to have a single Inbox playlist to save music and keep the date when I saved each song.

WTFPL (Do What The F*k You Want To Public License) https://en.wikipedia.org/wiki/WTFPL

Sharing a short post I posted a year ago with some thoughts on password managers.

## The password-management promise

I don't buy the promise behind 1Password or LastPass.

You only need to remember one password. The last password you'd need to remember.

They don't tell you that you're also building a one-stop shop for hackers to steal it all at once.

The solution?

Store hints, not passwords.

Don't reuse passwords. Use algorithmic passwords instead.

Use passkeys and security keys.

https://sketch.nono.ma/the-password-management-promise

---

I've always preferred 1Password and Bitwarden to LastPass.

I still prefer to encrypt sensitive data and "secure notes" with custom workflows (GPG keys, for instance) instead of relying on third parties, and even more when the data would be store in the cloud, in a centralized location.

I can't imagine the nightmare of having all your secrets exposed, not just for the risk of it but for having to reset all your exposed accounts.

(+1 to GPG encryption.)

https://archive.is/aRldC

August 16, 2023

The only thing WordPress and Laravel have in common is that both are written in PHP.

But if you've used Laravel, you know it's pleasant to write apps with it.

"If it’s finished, the applause, the thanks, the gratitude are something else. Something extra and not part of what you created. If you play a beautiful song for two people or a thousand, it’s the same song, and the amount of thanks you receive isn’t part of that song." (Seth Godin, The Icarus Deception: How High Will You Fly?)

You can write and run code in the browser with p5.js's playground, and even create an account to store your code sketches.

https://editor.p5js.org/nonoesp/sketches/ruSYeEUEH

To get started, simply SSH to our CMS.

pico.sh is a content management system (CMS) for hackers.

That's what I infer from the site, but it would be helpful if they added a one-liner to the home page and repo.

Descript acquired Lyrebird years ago and already have a product that can synthesize your voice.

They require training data longer than 15 seconds, which could lead the out out to resemble more the actual voice.

I've seen weird behaviors where the AI voice forces a British accent to pronounce certain words which I don't have.

Descript also uses voice synthesis to regenerate edited portions of conversations with a noticeable cut to smoothen the transition, which is pretty useful.

When the product officially rolls out there will be a “no-go voice list” that detects and prevents AI-generated speakers that are too similar to prominent figures.

Should they then let anyone register their "voice id" to prevent others frok generating similar voices?

What if your voice happens to resemble the voice of a "prominent figure?"

There are different "problems" as you're saying that could tackle your needs.

- Gesture recognition to detect different hand gestures

- Skeleton detection to infer people's positions and whether they have a raised hand

- Object detection (e.g., YOLO) to detect hands

I believe in this case you could go for skeleton detection as you get information from which you could infer raised hands.

In theory, switching from diesel or electricity to PV pumping should eliminate greenhouse gas emissions. But in practice, farmers often use their solar pumps to supplement existing pumps, rather than replacing them. And, however it is pumped, the extra water available will also encourage farmers to adopt more intensive farming methods, using more fertilizer and machinery to grow thirstier cash crops, increasing the carbon footprint of the farm.

I've been streaming and recording locally with OBS in 2K (1440p) with an M3 Max MacBook Pro and it works like a charm—with a Sony ZV-E10 being capture by Elgato HD60X.

I can even develop and run Apple MLX code while I'm streaming. (I lose a few frames when generating images with Stable Diffusion or load big LLMs like Gemma 7B.)

My MacBook Pro M1 wasn't there for streaming and recording at the same time. But even an M1 Max could do the job as well.

Amazing write-up. Thanks for sharing.

How reliable is Porkbun?

What about CloudFlare's "at-cost pricing for registration and renewal," they don't make a profit? [1]

[1] https://www.cloudflare.com/products/registrar/

I can't believe I guessed that was the coast of Zakynthos from a photo.

Can't count how many times I've played The Grid on loop.

This piece resonates so much with what the movie and soundtrack evoked in me.

I discovered Glass' works from it and love his discography.

I wrote a live video/audio marker creator with OpenAI's Whisper.

A Stream Deck XL button runs a Python script that creates a timestamp for ongoing video and audio recordings which I use for live stream and podcast chapters while I'm streaming or recording.

Different buttons on the Stream Deck create markers with different labels, e.g., Introduction, Break, Marker, etc.

But as it's hard to name markers for every section and you have to go back to the recordings to infer what each section was about, I added a dynamic marker script that trims audio around a marker with ffmpeg, locally transcribes it with Whisper, and appends the transcript to the marker, so I can easily guess what the marker title should be without having to scroll and watch the video.

Here's the Product Hunt page → https://www.producthunt.com/posts/gen2-by-runwayml

Agreed! I thought the same.

If you use the large-v2 model they expose via the API, the more accurate, in your local machine, you'll see that even though it works great it's slow and won't work for long audio files because of memory limitations.

It's fairly easy and quick to run Whisper for free either locally in an Anaconda environment with Python or the command-line interface or, even better, in a Google Colab notebook.

Here's a sample notebook that builds on a notebook by Pete Warden.

https://colab.research.google.com/drive/1sxsey3n0jd09MjUd9Ky...

Learn how to encrypt and decrypt sensitive data

I was expecting someone to have mentioned Descript's Sound Studio audio enhancement filter.

I use Typinator expansions with the input video path (from the clipboard) as a variable to perform common ffmpeg operations, such as downsize, speed up, extract audio, strip audio, export frames to create GIF animations, or encoding as MP4.

Typinator lets you use what's on the clipboard but also add scripting (say, Python or JavaScript) when certain parameters need to be calculated, like sizes or names.