The best Hacker News stories from All from the past day
Latest posts:
Why can't my mom email me?
Aider: AI pair programming in your terminal
End of the Line? Saudi Arabia to scale back plans for desert megacity
Mental health in software engineering
Apple alerts users in 92 nations to mercenary spyware attacks
Show HN: I made a new sensor out of 3D printer filament for my PhD
Here's a "behind-the-scenes" look at my development of a cool sensor during my PhD (electrical engineering). This sensor is only about 1/3 of my total research for my degree and took about a year.<p>I've been on HN for a while now and I've seen my fair share of posts about the woes of pursuing a PhD. Now that I'm done with mine I wanna share some anecdotal evidence that doing a PhD can actually be enjoyable (not necessarily easy) and also be doable in 3 years.<p>When I started I knew I didn't want to work on something that would never leave the lab or languish in a dissertation PDF no one will ever read. Thanks to an awesome advisor I think I managed to thread the needle between simplicity and functionality.<p>Looking back, the ideas and methods behind it are pretty straightforward, but getting there took some doing. It’s funny how things seem obvious once you've figured them out!<p>Oh, I love creating GUIs for sensor data and visualizations as you'll see -- it's such a game changer! pyqtgraph is my go-to at the moment - such a great library.
Show HN: I made a new sensor out of 3D printer filament for my PhD
Here's a "behind-the-scenes" look at my development of a cool sensor during my PhD (electrical engineering). This sensor is only about 1/3 of my total research for my degree and took about a year.<p>I've been on HN for a while now and I've seen my fair share of posts about the woes of pursuing a PhD. Now that I'm done with mine I wanna share some anecdotal evidence that doing a PhD can actually be enjoyable (not necessarily easy) and also be doable in 3 years.<p>When I started I knew I didn't want to work on something that would never leave the lab or languish in a dissertation PDF no one will ever read. Thanks to an awesome advisor I think I managed to thread the needle between simplicity and functionality.<p>Looking back, the ideas and methods behind it are pretty straightforward, but getting there took some doing. It’s funny how things seem obvious once you've figured them out!<p>Oh, I love creating GUIs for sensor data and visualizations as you'll see -- it's such a game changer! pyqtgraph is my go-to at the moment - such a great library.
AI-generated sad girl with piano performs the text of the MIT License
Anyone got a contact at OpenAI. They have a spider problem
Why does part of the Windows 98 Setup program look older than the rest? (2020)
Kobo announces color e-readers
Show HN: Sonauto – A more controllable AI music creator
Hey HN,<p>My cofounder and I trained an AI music generation model and after a month of testing we're launching 1.0 today. Ours is interesting because it's a latent diffusion model instead of a language model, which makes it more controllable: <a href="https://sonauto.ai/">https://sonauto.ai/</a><p>Others do music generation by training a Vector Quantized Variational Autoencoder like Descript Audio Codec (<a href="https://github.com/descriptinc/descript-audio-codec">https://github.com/descriptinc/descript-audio-codec</a>) to turn music into tokens, then training an LLM on those tokens. Instead, we ripped the tokenization part off and replaced it with a normal variational autoencoder bottleneck (along with some other important changes to enable insane compression ratios). This gave us a nice, normally distributed latent space on which to train a diffusion transformer (like Sora). Our diffusion model is also particularly interesting because it is the first audio diffusion model to generate coherent lyrics!<p>We like diffusion models for music generation because they have some interesting properties that make controlling them easier (so you can make <i>your own</i> music instead of just taking what the machine gives you). For example, we have a rhythm control mode where you can upload your own percussion line or set a BPM. Very soon you'll also be able to generate proper variations of an uploaded or previously generated song (e.g., you could even sing into Voice Memos for a minute and upload that!). @Musicians of HN, try uploading your songs and using Rhythm Control/let us know what you think! Our goal is to enable more of you, not replace you.<p>For example, we turned this drum line (<a href="https://sonauto.ai/songs/uoTKycBghUBv7wA2YfNz">https://sonauto.ai/songs/uoTKycBghUBv7wA2YfNz</a>) into this full song (<a href="https://sonauto.ai/songs/KSK7WM1PJuz1euhq6lS7">https://sonauto.ai/songs/KSK7WM1PJuz1euhq6lS7</a> skip to 1:05 if impatient) or this other song I like better (<a href="https://sonauto.ai/songs/qkn3KYv0ICT9kjWTmins">https://sonauto.ai/songs/qkn3KYv0ICT9kjWTmins</a> - we accidentally compressed it with AAC instead of Opus which hurt quality, though)<p>We also like diffusion models because while they're expensive to train, they're cheap to serve. We built our own efficient inference infrastructure instead of using those expensive inference as a service startups that are all the rage. That's why we're making generations on our site free and unlimited for as long as possible.<p>We'd love to answer your questions. Let us know what you think of our first model! <a href="https://sonauto.ai/">https://sonauto.ai/</a>
Show HN: Sonauto – A more controllable AI music creator
Hey HN,<p>My cofounder and I trained an AI music generation model and after a month of testing we're launching 1.0 today. Ours is interesting because it's a latent diffusion model instead of a language model, which makes it more controllable: <a href="https://sonauto.ai/">https://sonauto.ai/</a><p>Others do music generation by training a Vector Quantized Variational Autoencoder like Descript Audio Codec (<a href="https://github.com/descriptinc/descript-audio-codec">https://github.com/descriptinc/descript-audio-codec</a>) to turn music into tokens, then training an LLM on those tokens. Instead, we ripped the tokenization part off and replaced it with a normal variational autoencoder bottleneck (along with some other important changes to enable insane compression ratios). This gave us a nice, normally distributed latent space on which to train a diffusion transformer (like Sora). Our diffusion model is also particularly interesting because it is the first audio diffusion model to generate coherent lyrics!<p>We like diffusion models for music generation because they have some interesting properties that make controlling them easier (so you can make <i>your own</i> music instead of just taking what the machine gives you). For example, we have a rhythm control mode where you can upload your own percussion line or set a BPM. Very soon you'll also be able to generate proper variations of an uploaded or previously generated song (e.g., you could even sing into Voice Memos for a minute and upload that!). @Musicians of HN, try uploading your songs and using Rhythm Control/let us know what you think! Our goal is to enable more of you, not replace you.<p>For example, we turned this drum line (<a href="https://sonauto.ai/songs/uoTKycBghUBv7wA2YfNz">https://sonauto.ai/songs/uoTKycBghUBv7wA2YfNz</a>) into this full song (<a href="https://sonauto.ai/songs/KSK7WM1PJuz1euhq6lS7">https://sonauto.ai/songs/KSK7WM1PJuz1euhq6lS7</a> skip to 1:05 if impatient) or this other song I like better (<a href="https://sonauto.ai/songs/qkn3KYv0ICT9kjWTmins">https://sonauto.ai/songs/qkn3KYv0ICT9kjWTmins</a> - we accidentally compressed it with AAC instead of Opus which hurt quality, though)<p>We also like diffusion models because while they're expensive to train, they're cheap to serve. We built our own efficient inference infrastructure instead of using those expensive inference as a service startups that are all the rage. That's why we're making generations on our site free and unlimited for as long as possible.<p>We'd love to answer your questions. Let us know what you think of our first model! <a href="https://sonauto.ai/">https://sonauto.ai/</a>
Twitter's pivot to x.com is a gift to phishers
Double-entry bookkeeping as a directed graph
Yes, social media is a cause of the epidemic of teenage mental illness
Re-creating Disney's sodium vapor process [video]
Cow Magnets
Fairbuds: In-ear with replaceable batteries
Show HN: We built PriceLevel to find out what companies pay for SaaS
Hey HN! Christine and Steven here. As a PM and engineer, we’ve both evaluated and purchased a lot of software. One of the biggest frustrations was figuring out how much it would cost us without having to go through the sales process. When we did have a quote, we had no idea if we were getting a good deal or ripped off.<p>We built a site where you can see what other companies are actually paying for SaaS and enterprise software. Buyers contribute prices via quotes, pricing proposals, and other documentation to ensure quality.<p>We unlocked Talkdesk for Show HN users so that you can use the product without needing to sign in or upgrade. Check it out at <a href="https://www.pricelevel.com/showhn" rel="nofollow">https://www.pricelevel.com/showhn</a>. Would love to hear any feedback, thank you!