HN user
@rrherr rrherr.github.io
"VexTab is a language that allows you to easily create, edit, and share standard notation and guitar tablature. Unlike ASCII tab, which is designed for readability, VexTab is designed for writeability."
Software engineers began to fracture into three groups. The first group is the “never AIs”.
The second group consists of ‘pragmatic AI adopters’. I identify the most with this second group.
And so it is the frame of a third group that has posed a challenge for me. Some call this the ‘software dark factory folks’. This group believes that it is possible to have AI coding agents write code with little to no human intervention.
If software dark factories are possible, then the entire practice of software engineering is going to change. But how can you take the software dark factory folks seriously when you encounter incredibly stupid coding agent behaviours in your own day-to-day work? Nothing they say lines up with your own frame; all their ‘data points’ are off-the-cuff remarks that may be rejected due to your own lived experiences.
Avoiding frame fixation _sounds_ easy when it’s about another person’s domain. It’s less easy when a) it’s about your own domain, b) when the new frame goes against everything you believe about your own hard-won expertise, and c) _when the new frame is fundamentally uncertain._
I want to talk about that last point. We do _not_ know if these ‘software dark factories’ are possible. I’m not saying that the folks writing field reports are lying, or that the benefits they’re already seeing are fake. I’m saying that we _can’t_ know what the tradeoffs are, and where the limits of this approach lies. Nobody can. This is a new technology with new affordances. Nobody can know what’s possible here. This is what uncertainty feels like.
But I think it’s also true that you need to take this ‘dark factory’ frame seriously. There are enough field reports now from enough unrelated people that indicate that _something_ is going on. More importantly, the potential impact on your career — if you are a software engineer — is too large to ignore.
Thankfully, the Data-Frame theory already offers us one way out: you _don’t_ have to believe their frame. You may hold on to your current frame, and elaborate a second frame in parallel. [This post explains how... ]
Audio can be a great way to capture ideas and thought processes ... This can work especially well for people who are distracted by form and "writing correctly" too early in the process, for people who are intimidated by blank pages, for non-neurotypical people, etc. Self-recording is a great way to set all of those artifacts of the medium aside and capture what you want to say.
Yes, this is my process:
Record yourself rambling out loud, and import the audio in NotebookLM.
Then use this system prompt in NotebookLM chat:
Write in my style, with my voice, in first person. Answer questions in my own words, using quotes from my recordings. You can combine multiple quotes. Edit the quotes for length and clarity. Fix speech disfluencies and remove filler words. Do not put quotation marks around the quotes. Do not use an ellipsis to indicate omitted words in quotes.
Then chat with "yourself." The replies will match your style and will be source-grounded. In fact, the replies automatically get footnotes pointing to specific quotes in your raw transcripts.
This workflow may not save me time, but it helps me get started, or get unstuck. It helps me stop procrastinating and manage my emotions. I consider it assistive technology for ADHD.
My current manager does not care about my personal growth or career goals whatsoever and he’s a bad manager for it. But he is good at delivering projects.
I've gone back and forth between Manager and IC. I'm the opposite: I don't care about delivering projects, but I do care about my team's personal growth and career goals.
> The elephant in the room is that we’re all using AI to write but none of us wants to feel like we’re reading AI generated content.
Reminds me of a quote from St. Augustine's autobiography, "Confessions":
"I have known many men who wished to deceive, but none who wished to be deceived."
Yes, I've had great results with a similar workflow.
I record myself rambling out loud, and import the audio in NotebookLM.
Then I use this system prompt in NotebookLM chat:
Write in my style, with my voice, in first person. Answer questions in my own words, using quotes from my recordings. You can combine multiple quotes. Edit the quotes for length and clarity. Fix speech disfluencies and remove fillers. Do not put quotation marks around the quotes. Do not use an ellipsis to indicate omitted words in quotes.
Then chat with "yourself." The replies will match your style and will be source-grounded. In fact, the replies automatically get footnotes pointing to specific quotes in your raw transcripts.
I also like brainstorming by generating Audio Overviews, Slide Decks, and Reports in NotebookLM. The Audio Overviews don't sound like AI writing. The Slide Decks and Reports do sound like AI writing, if you use the defaults, but you can use custom prompts.
This workflow may not save me time, but it helps me get started, or get unstuck. It helps me stop procrastinating and manage my emotions. I consider it assistive technology for ADHD.
The title is misleading. The message is really:
Stop using unlabeled icons in data tables.
It says, "Norman Nielson argues that text + icon has the highest cognitive recall and lowest error rate"
Here's what the Nielsen Norman Group says about Icon Usability: https://www.nngroup.com/articles/icon-usability/
The conclusion: "Always include a visible text label. As Bruce Tognazzini once said, 'a word is worth a thousand pictures.'"
Here's the quote in context: https://www.asktog.com/columns/038MacUITrends.html
"In 1985, after a year of finding that pretty but unlabeled icons confused customers, the Apple human interface group took on the motto 'A word is worth a thousand pictures.' This still holds true."
Thanks! Also, according to that LinkedIn post date extractor, this post by first author Giuseppe Soda was made on Thu, 08 May 2025 06:22:28 GMT:
https://www.linkedin.com/posts/giuseppe-beppe-soda-414749b0_...
Good question!
Leonardo Rizzo, one of the researchers, claimed on X.com that they published before the Pope was elected.
An X user commented:
“Guessed” after the fact. Interesting nonetheless and worth sharing before the event next time!
Rizzo replied:
Thanks a lot! We shared it the 8th morning on linkedin, the university website and few other sources (italian press). Next time I’ll also share it on X
https://x.com/LnrdRizzo/status/1920841806096343409
https://www.linkedin.com/posts/universita-bocconi_a-new-way-...
Check out "African Polyphony and Polyrhythm", a presentation by Chris Ford at Strange Loop 2016. He uses Clojure to model traditional central African drumming patterns with variations.
This Substack post is a summary of an essay by Joseph Brodsky about Dostoevsky — but the post does not link or name the essay.
The essay is named "The Power of the Elements" and it can be read here on Google Books:
https://www.google.com/books/edition/Less_Than_One/N5Nzm2uih...
Here's a plain language explanation of why uplift modeling is useful, written by the same author as the paper:
https://stochasticsolutions.com/uplift/
It is normally assumed that the worst outcome direct marketing activity can have is to waste money. In fact, some direct marketing provably drives away business within certain segments, and it is not unknown for it to drive away more business in total than it generates. This is especially true in retention activity.
[Non-Uplift] Churn and attrition models prioritize customers whose probability of leaving is highest. Such customers tend to be dissatisfied, so are usually hard to retain. To make matters worse, in many cases, the only thing currently keeping them is inertia, and interventions run a serious risk of back-firing, triggering the very defections they seek to avoid.
It is more profitable to focus retention activity on those people who ... will leave without an intervention, but who can be persuaded to stay. Uplift models allow you to target them, and them alone. At all costs, you want to avoid targeting the ... so-called Sleeping Dogs, whose defection you are likely to trigger by your intervention. Again, uplift models can direct you away from those customers.
This isn't exactly what you asked for, but there's a "drumsep" model, which takes a drum audio track and separates it into 6 stems: kick, snare, toms, hi-hat, ride, and crash.
Ctrl+F for "drumsep" in this doc:
Instrumental, vocal & other stems separation & mix/master guide - UVR/MDX/Demucs/GSEP & others - Google Docs https://docs.google.com/document/d/17fjNvJzj8ZGSer7c7OFe_CNf...
Here's the most impressive results I've seen for automated guitar transcription:
High-resolution guitar transcription via domain adaptation
Demo Videos: https://xavriley.github.io/HighResolutionGuitarTranscription... Paper: https://arxiv.org/abs/2402.15258
We propose the use of a high-resolution piano transcription model to train a new guitar transcription model. The resulting model obtains state-of-the-art transcription results on GuitarSet in a zero-shot context, improving on previously published methods.
Betteridge's law of headlines: "Any headline that ends in a question mark can be answered by the word no."
Here’s a better answer, in my opinion: https://www.ethanhein.com/wp/2020/so-what/
If you have never listened to jazz before, Miles Davis’ Kind Of Blue is a great place to start. The heart of the album is its first track, “So What.”
“So What” is famous for being one of the first modal jazz tunes. This just means that it doesn’t have a lot of chord changes compared to the fast harmonic rhythms of bebop. The A sections use the D Dorian mode. This scale is especially easy to play on the piano; just play the white keys. The B section is up a half step, on E-flat Dorian. If you play the black keys on the piano, you get five of the seven notes in this scale. I had a complete beginner pianist improvise a solo over “So What” in class. I called out when she needed to switch between the white and black keys. It worked!
“So What” occupies a similar place in jazz pedagogy to the blues: it’s simple enough for beginners to play, but you can devote a lifetime to practicing and never get to the bottom of it. If you want to learn how to improvise jazz, you should definitely learn Miles’ solo.
Black American music uses lots of call and response as a structuring element. “So What” has many call-and-response pairs at different scales. Here are all the layers I can detect, ranging from micro to macro … (7)
I would bet that this fractal-like self-similarity across different levels is a major reason for the tune’s appeal. Any tune this immediately catchy yet also structurally deep is going to attract a lot of imitation.
Anybody who’s been to music school can write complex and abstruse jazz tunes, and blow complicated solos over them. Not many musicians can write memorable hooks. And only the most profound artists can write a hook that conceals as much depth and possibility as “So What.” I wonder if that level of creativity is teachable, or learnable?
There’s only one other jazz standard I know of with comparable lyrics:
“Lush Life” by Billy Strayhorn. https://en.m.wikipedia.org/wiki/Lush_Life_(jazz_song)
Listen to Johnny Hartman singing in this recording with John Coltrane: https://youtu.be/sNIn1_RLkmc
"Data centres have turned Big Tech into big spenders: Companies need AI services revenues, not cost savings, to fuel data centre boom"
Adrian, I made a Google Colab notebook to try a different beat detection algorithm with your tune. The results sound pretty good to me!
You can listen here:
https://colab.research.google.com/drive/1Pqgc9s-nBKxU_3Ap6K0...
See also:
How 'The Karate Kid' Ruined The Modern World (2010)
https://www.cracked.com/article_18544_how-the-karate-kid-rui...
"I think The Karate Kid ruined the modern world. Not just that movie, but all of the movies like it (you certainly can't let the Rocky sequels escape blame). Basically any movie with a training montage.
You know what I'm talking about; the main character is very bad at something, then there is a sequence in the middle of the film set to upbeat music that shows him practicing. When it's done, he's an expert. ...
Every adult I know--or at least the ones who are depressed--continually suffers from something like sticker shock (that is, when you go shopping for something for the first time and are shocked to find it costs way, way more than you thought). Only it's with effort. It's Effort Shock.
We have a vague idea in our head of the "price" of certain accomplishments, how difficult it should be to get a degree, or succeed at a job, or stay in shape, or raise a kid, or build a house. And that vague idea is almost always catastrophically wrong.
Accomplishing worthwhile things isn't just a little harder than people think; it's 10 or 20 times harder."
---
Effort Shock and Reward Shock (2014)
https://www.ribbonfarm.com/2014/07/09/effort-shock-and-rewar...
"The good news is what I’ve started calling reward shock. In some (not all) domains, it is more than enough to offset effort shock.
When you overcome effort shock for a non-trivial learning project and get through it anyway, despite doubts about whether it is worth it, you can end up with very unexpected rewards that go far beyond what you initially thought you were earning. This is because so few people get through effort shock to somewhere worthwhile that when you do it, you end up in sparsely populated territory where further gains through continued application from the earned skill can be very high.
Programming, writing and math are among the skills where there you get both significant effort shock and significant reward shock."
What do you mean by candidate? Applicant, or interviewee?
OP was writing about interviewees, not applicants:
“In order for it to be ghosting, the ghosted party has to expect the conversation will continue. This means that if you apply and never hear back from a job, that’s not ghosting, a conversation never started. It only becomes ghosting when there is an expected next step that never happens.”
“Two Pizza Hut operators in California are eliminating their in-house delivery services at hundreds of stores, resulting in more than 1,200 driver layoffs”
Thanks. I’d love to try TuneNN! Are you releasing a pretrained model? How do I run it on a wav file?
How does the accuracy of this compare to CREPE?
https://github.com/maxrmorrison/torchcrepe
Does anyone know what the current state of the art is, within the Music Information Retrieval community?
Kind of, but the Forbes post says more about Microsoft than The Verge post
That you, ChatGPT? You're hired!
"We developed a variation on the Ramer-Douglas-Peucker (RDP) algorithm, which is a curve simplification algorithm that reduces the number of points in a curve while preserving its important details. It achieves this by recursively removing points that deviate insignificantly from the simplified version of the curve."
This reminded me of an old side project, which others may be interested in. I applied Douglas-Peucker to Picasso for a talk at Strange Loop 2018:
Picasso's Bulls: Deconstructing his design process with Python https://rrherr.github.io/picasso/
The full quote in context was: "You can name a thousand great instrumentalists or you can name a thousand great vocalists, but he’s the only person you could find who changed the way people played their instruments and the way people sang. Louis does that in a four-year period in the 1920s; by 1930, if you aren’t playing or singing like him, you’re out of work."