HNHacker News
TopNewBestAskShowJobs

riemannzeta

447 karma · joined May 29, 2008

submissionscomments
riemannzeta··on What I Learned Writing an Album in Just Intonation
I will admit that I already had some familiarity with the topic, but I feel you did an amazing job at explaining specifically the constraints you're under as a composer, and in elucidating how having digital tools doesn't magically solve the problems you face in harmonizing in just intonation over the course of a piece. I had not really understood that before reading your article, and at least suspected that we might transition to hearing mostly Just Intonation eventually as digital tools became more widely used.

Of course you could make it more accessible by starting from some even more basic principles and elaborating further on those foundations, but then you risk losing some of our audience that's already familiar with those. For me at least you struck a good balance. (No pun intended!)

riemannzeta··on What I Learned Writing an Album in Just Intonation
Readers who enjoyed this essay might also enjoy this one, which goes into further depth on the mathematics of musical intervals, the connection between the "blue note" in Equal Tempered jazz music, and more:

https://iopscience.iop.org/article/10.1088/0034-4885/72/7/07...

As an added bonus, the author draws an extended analogy between musical intervals and energy levels in quantum mechanics toward the end of the article, and indeed the study of the patterns in intervals between and among energy levels in the simple harmonic osciallator remains an active area of reasearch!

riemannzeta··on What I Learned Writing an Album in Just Intonation
I feel like I can hear beat frequencies in the 12TET chords that I don't hear in the JI chord examples.

Anybody else hearing that? Or is that my imagination?

riemannzeta··on Poor Richard's Almanack
Agreed. If you consider his habits of:

1) seeding his former apprentices with funding in exchange for a small amount of equity in their new printing ventures, 2) forming non-profit organizations to address public needs (such as subscription libraries, fire departments, and street sweepers) 3) forming groups of like minded tradespeople who were interested in improving themselves and their communities through ventures like 1) and 2)

...then one might even say that Benjamin Franklin was a kind of 18th century YCombinator

riemannzeta··on Artificial consciousness: a perspective from the free energy principle
I found the paper very interesting and thought provoking. But I wonder, as a practical matter, whether the heavy reliance placed upon the von Neumann architecture as providing a distinction between a conscious and non-conscious intelligence is meaningful. It seems like most modern computer architecture (and GPU architecture, in particular) doesn't easily fit clean definitions of von Neumann architecture. Does this mean that a machine learning model trained on GPUs might be conscious? The paper's explanation of what exactly is meant by unmediated causal flow is quite murky here.

But I appreciated the author's careful definition of the FEP and its use in his framework.

riemannzeta··on What would it take to recreate Bell Labs?
It became a legend.
riemannzeta··on What would it take to recreate Bell Labs?
Very cool. What did you use it for?
riemannzeta··on Silicon Valley's best kept secret: Founder liquidity
This is a wonderful article and kudos to the author for his moral sensibility here. The lack of liquidity and anti-dilution rights for any except a handful of key persons is a dirty secret of Silicon Valley. Most startup employees do not end up better compensated than they would at a larger company on a net present value basis even when their startup is successful -- and they don't as easily get liquidity along the way although there are more private secondary market brokers than there used to be.

The other angle worth observing here is the tax angle. In many cases, the founders are taking liquidity at valuations that are only loosely tied to tax valuations of the company. These valuations are fine for the founders and preferred by the buyers/investors, but undercut the premise of the tax system that was redesigned after the options backdating scandals of the early 2000s, in part, to ensure that taxes were getting paid in accordance with the actual capital gains.

riemannzeta··on How Does GPT-4o Encode Images?
Love this curious and open-minded exploration of how this stuff works.

The pyramid strategy loosely tracks with renormalization group theory, which has been formally studied for years as a method of interpreting machine learning models:

https://arxiv.org/abs/1410.3831

I love the convergence we're seeing in the use of models from different fields to understand machine learning, fundamental physics, and human consciousness. What a time to be alive.

riemannzeta··on Chaotic Progress
How vigorous competition among public political parties and private for-profit companies has promoted progress by balancing the personal interests of new entrants and incumbent elites
riemannzeta··on Weather.gov 2.0
A vision of what the future could be for all of our federal agencies. Kudos to the National Weather Service for leading.
riemannzeta··on Where is Noether's principle in machine learning?
Your work here is so beautiful, but perhaps one lesson is that growth and learning result where symmetries are broken. :-D
riemannzeta··on Where is Noether's principle in machine learning?
Not quite what you're looking for, but worth pointing out that Grant Sanderson of 3Blue1Brown has published the "framework" he uses for his math videos on GitHub.

https://github.com/3b1b/manim

riemannzeta··on YC: Requests for Startups
Anybody interested in talking to a potential customer for:

* Developer tools inspired by existing internal tools

* LLMs for manual back office processes in legacy enterprises

Feel free to message me. Having worked in-house for the last 15 years, I've seen a variety of these, have built more than a few of them myself, and am in the middle of building some new ones using LLMs right now — all within large enterprise legal departments.

I won't be able to share any confidential data, but I'm happy to answer questions about patterns I've seen across a few different organizations and where the unmet needs seem to be.

riemannzeta··on Lessons from history's greatest R&D labs
That's different from how things worked for Edison. But similar to how they worked for Luther Burbank.

I'll be rooting for you to succeed in this endeavor. I used to visit IBM ARC when I was a kid, and to my mind the academic environment is not quite the same. But as you know we really have not seen corporate R and D recover from the peaks it attained over fifty years ago.

riemannzeta··on Lessons from history's greatest R&D labs
Eric's depiction of Edison and his rapid experimentation sounds very similar to the descriptions of one of Edison's contemporaries, Luther Burbank, as told in The Garden of Invention by Jane S. Smith. I recommend this book to the author (if he hasn't already read it himself) as I do to anybody else who enjoyed his essay.

There's a lot to be said for just exploring the combinatorial space of possibility in even a naive or brute force manner.

riemannzeta··on Lessons from history's greatest R&D labs
Makes sense. We have to work within the system as we find it. Would be great to see the system evolve in a way that would support efforts like yours though!
riemannzeta··on Lessons from history's greatest R&D labs
I'm not sure who's better positioned to answer, you or Jeremy and I asked him already:

How are you thinking about the role of patenting at Answer.AI? How about handling tech transfer issues with universities?

riemannzeta··on Lessons from history's greatest R&D labs
How are you thinking about the role of patenting at Answer.AI? How about handling tech transfer issues with universities?
riemannzeta··on Mr. Rogers' music was intricate (2020) [video]
Custom GPT trained on some of Fred Rogers work

https://chat.openai.com/g/g-oW7MpoDDn-mister-rogers-gpt

riemannzeta··on Mr. Rogers' music was intricate (2020) [video]
https://open.spotify.com/album/5rorTQbNjzQGwlfmv1ecsK

You are Special

Really great energy. Nothing quite like it.

riemannzeta··on Patrick McKenzie (patio11) on navigating complex systems
I love your writing. I believe the reason some people do not is because you are very precise, which means using more words. As you know, people are impatient and prefer less words unless they can see clearly how more words were necessary.

In case you needed any encouragement, here's some: Keep going, you're amazing!

riemannzeta··on Open-Source AI Is Uniquely Dangerous
Are the guns sold at Walmart uniquely dangerous?
riemannzeta··on LLM Visualization
Are you referring specifically to line 141, which sets the number of embedding elements for gpt-nano to 48? That also seems to correspond to the Channel size C referenced in the explanation text?

https://github.com/karpathy/minGPT/blob/master/mingpt/model....

riemannzeta··on Broadcom lays off many VMware employees after closing acquisition
I have heard this colorfully and memorably referred to as "the need for a single throat to choke."
riemannzeta··on The Lack of Compensation in Open Source Software Is Unsustainable
There have been steps forward in the direction of making donation easier:

https://github.com/sponsors

GitHub Sponsors directs individuals to Open Collective https://opencollective.com/ , which can serve as a "fiscal host." The advantage here is that the default rule at law for how a group of developers working together will be treated is partnership, which means joint and several liability. Working with a fiscal host partitions individual liability from group liability.

But there are still open questions. I don't know all the details of how Open Collective works from a corporate law perspective. How do they partition the liability of different collectives that are hosted by the same fiscal host? That is important to understand because otherwise, the collectives that share a fiscal host are partners. This is better than all of the individuals who contribute to each and all of the collectives being partners (because there's still a partition between individual and group liability), but worse than if each collective's liability was partitioned from the other collectives' liability (and the fiscal host itself's liability).

This seems like an active area for legal innovation. No jurisdiction I know of is optimizing to maximize speed and minimize cost of setting up corporations. Yet without the ability to shield the contributors of capital and labor from individual liability, there isn't much a group of individuals can do in our modern economy. The problems that many DAOs have had is a case study in this.

People who care about open source should also care about scaling up the speed and minimizing the cost of incorporating. Stripe Atlas and similar services are underrated sources of economic growth for this reason, IMHO.

riemannzeta··on Matplotlib
Exactly! Typing "import matplotlib" at the python REPL instantly gave you the capability of Matlab without the requirement of a commercial license. If I recall correctly, matplotlib was built after Matlab increased the prices for its commercial license for academics.

Between matplotlib and pandas I believe we have a sufficient explanation as to how python became the language of choice for data analysis.

riemannzeta··on The pivot table, the spreadsheet's most powerful tool (2020)
As somebody who's been at this for a couple of decades, this strikes me as a rather well balanced view. What I can't understand is how no patent reform discussions have anything like this kind of balance. It's all "stronger patents" or "weaker patents." Nobody seems to understand how and when they work vs. don't work.
riemannzeta··on The pivot table, the spreadsheet's most powerful tool (2020)
A similar story could be told about Steve Jobs and Apple. Anecdotally at least, Steve hired the patent litigators (who had represented Apple in a patent infringement lawsuit with Taiwanese MP3 maker Creative Technologies over the iPod) to patent the iPhone. But that was more targeted than the spray and pray approach.
riemannzeta··on The pivot table, the spreadsheet's most powerful tool (2020)
The patent in question appears to have been filed in 1996

https://patents.google.com/patent/US5915257A/en?oq=5915257

A quick search suggests that Excel pivot tables launched in 1993. So the Brio patent does indeed appear to have come later in time.

Note that although the patent indicates ownership by Oracle, the assignment history shows that it originated with Brio.

← PreviousPage 2 of 6Next →