Hacker News new | past | comments | ask | show | jobs | submit | nonrandomstring's comments login

Essay touches on an important point that I also discovered along my research journey; Much of the knowledge of experts is ineffable [0]. And so here's the problem with present "AI". It contains no ineffable knowledge, because that knowledge is never symbolically expressed and cannot be inferred from what is expressed. In fact the "official" narrative often disguises the real knowledge.

This is what the analytical elicitation stage of expert systems [1] was supposed to address (a now mature but presently unfashionable branch of "AI").

[0] https://en.wikipedia.org/wiki/Ineffability

[1] https://en.wikipedia.org/wiki/Expert_system


Notable line:

"framing the narrative around the technology is far more important than the technology itself"

We're on the umpteenth cycle of this and we all seem very slow to learn the pattern. I tried to express it in "Chindogu" [0] but since then "AI" hype has taken the fake technological narrative to a whole new level.

[0] https://cybershow.uk/blog/posts/chindogu/


You claim the exact opposite of what the author wrote.

   " the reason I wrote this book was that I became frustrated with
   myself...  You write books to convince people of certain opinions
   and then you hope that some other people do the actual work of
   making the world better... I experienced this emotion that I
   describe as moral envy: You're standing on the sidelines and
   wishing, gosh, wouldn't it be awesome to be in the arena? To
   actually have skin in the game? "
"Moral ambition" isn't for the faint-hearted or superficially "successful". You actually have to do stuff. Take a road less travelled. Eat your own dogfood. Make sacrifices and live by principles you espouse. Very few are authentic, courageous and determined in this regard and "successful" within our culture which actively rewards moral delinquency... bar a very few rare diamonds; for example Anita Roddick [0] who led the first wave of environmental, fair-trade ethics in beauty retail.

Besides, I think these are foundational personality traits ... very difficult to "learn/add-on/fake" later in life. So I think the author wastes time appealing to "elites" already saddled by their shameless immorality. Those panged by deathbed "philanthropic" regrets, fretting on their "moral legacy" or place in eternity - having spent their whole lives shitting on the world to get ahead - are a tough, niche audience. Better to speak to younger people who are not yet soured, who have not yet become extremely fearful of taking social risk or jeapordising their "career". If you're under 25 and questioning what "successful" maens, this could be a life-changing book.

[0] https://en.wikipedia.org/wiki/Anita_Roddick


This takes me back. 1990, building Boltzman machines and Perceptrons from arrays of void pointers to "neurons" in plain C. What did we use "AI" for back then? To guess the next note in a MIDI melody, and to recognise the shape of a scored note, minim, crotchet, quaver on a 5 x 9 dot grid. 85% accuracy was "good enough" then.

> recognise the shape of a scored note, minim, crotchet, quaver on a 5 x 9 dot grid

Reading music off a lined page sounds like a fun project, particularly to do it from scratch like 3Blue1Brown's number NN example[1].

Mix with something like Chuck[2] and you can write a completely clientside application with today's tech.

[1] - https://www.3blue1brown.com/lessons/neural-networks

[2] - https://chuck.stanford.edu/


Thanks for these links. You're right, I think computer-vision "sight reading" is now a fairly done deal. Very impressive progress in the past 30 years.

Did the output sound musical?

For small values of "music"? Really, no. But tbh, neither have more advanced "AI" composition experiments I've encountered over the years, Markov models, linear predictive coding, genetic/evolutionary algs, rule based systems, and now modern diffusion and transormers... they all lack the "spirit of jazz" [0]

[0] https://i.pinimg.com/originals/e4/84/79/e484792971cc77ddff8f...


Spoiler: Security Awareness Special Interest Group. We spoke to the founder. Interesting inights about sales and community based trust networks.

You have "epistemic integrity'.

I heard someone say "epistemic humility" the other day to mean fallibilism [0] and the conversation got interesting when we moved on to the subject of "what one can and should reasonably claim to know". For example: should cops know the law?

Not every programmer needs to be a computer science PhD with deep knowledge about obscure data-structures... but when you encounter them it's a decision whether to find out more.

Integrity is discomfort with "hand-waving and magical" explanations of things that we gloss over. Sure, it's sometimes expedient to just accept face-value and get the job done. Other times it's kinda psychologically impossible to move forward without satisfying that need to know more.

Frighteningly, the world/society puts ever more pressure on us to just nod along to get along, and to accept magic. This is where so much goes wrong with correctness and security imho.

[0] https://iep.utm.edu/fallibil/


What happened when DOGE gained access to US government servers? An exclusive interview with Alt US Digital Service (AKA "We The Builders") which HN readers might find interesting. Audio. 46 min.

Never had the pleasure of a Sennheiser but when working in radio I got my hands on a lot of rack vocoders for doing branding, stings and idents. Funny how the number 9000 comes up a lot, like Roland VP9000 and Eventide H9000. 80 and 90s vintage ones like Korg VC-10 or Elektronik EM-26 had unique sounds, but tbh the modern digital recreations are amazing models. There's not a world of difference between vocoding, autotune, shifting, harmonising etc once you realise how all the fx are now based in FFT, convolution etc - just different variations on processing and control graphs - and so it's fun to create your own vocal effects in things like Max/MSP/PureData. Technically there's a distinction between "effects" and "processing" in terms of how much of the direct (parallel) signal is put through. Chers Believe is a yardstick for "effect", whereas a lot of what I hear with Daft Punk (and Air, Kraftwerk) is quite heavily processed as to disguise the original voice entirely - just letting a bit of top/sibilant through to define the stops and fricatives.

Analog vocoders are only nominally like analog FFTs.

The shape of the filters, the smoothing between the filters and the synthesis section, and (on some models) the patchability all create a very different result.

The reason the best analog vocoders are so expensive is because the filter for each band is much more complex than a plain old bandpass filter, with a much higher component count. Typically there's a flatter passband and a steeper slope than you'd expect.

You can do digital convolution with thousands of bins and it sounds nothing like analog vocoding. It's much cleaner, doesn't have those lovely harmonically spaced filter resonances, and creates sounds that can feel more acoustic than electronic.


Did you listen to the example audio in the video? Soft synths and digital emulation can be absolutely amazing these days, but the VSM201 and Ultimate VoIS are in their own league. It’d be pretty easy to pick them out from a blind test with other vocoders.

Oh, it also might be of interest that the IVL algorithm isn’t FFT-based. I think their harmonizers sound better than the rest, so maybe FFT isn’t the best way to go.


Yes exactly, I was really excited when I found out that you do not need a FFT to do speech processing.

If you look at the code of (phone/voice) codecs GSM/Speex/Opus you can see that you can estimate the spectral envelope (or the configuration of a physical tube model for the vocal tract) in time domain with linear prediction coefficients (LPC).

And it is simple, e.g. the often used Levinson-Durbin algorithm is just 22 lines of C code. It is an interesting exercise to build your own vocoder from scratch that fits in a single screen page.

Many of the code snippets I have seen (which likely have already processed your voice) are just translations of the Fortran code of the book "Linear Prediction of Speech" by Markel and Gray (1976).


Ah yes, ladder or lattice filters. If you don't mind old fashioned mailing lists there's still a few of hanging around in MUSIC-DSP@LISTS.COLUMBIA.EDU where code gets shared.

I thought it was phase synchronous overlap add, but I just checked and now I'm not so sure.

Has anyone got more details?


That gives me something to research! I’ve only scratched the surface of IVL’s algorithm, but intend to look into it further.

...and billions of taxpayer dollars, hundreds of years of European science, standing on the shoulders of giants, thousands of years of Greek, Arabic and Far Eastern mathematics.... The "self-made industrialist" sketch is funny when it's Monty Python, but when I hear whining SV bros claiming they built an empire from a rolled-up-newspaper, it's so avoidantly ungrateful. Like some kid "divorces" their own parents, disowns their lineage and declares themselves a unique and special self-creation. The US would do well to reconnect its Native American culture and have more respect for what got everyone here.


I associate May 1 with getting mashed in Helsinki as for many years I spent it in Finland, with amazing parties in the park for Vappu [0] the Spring Festival. It's a celebration of Spring, labour day, and also "education and industry" since people proudly wear their school colours, company badges and graduation caps. Quite an atmosphere!

[0] https://en.biginfinland.com/vappu-spring-fest-finland/


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: