friggeri

Moritz Klein is the reason I got into synths, his YouTube videos[0] are a really accessible intro. I’m super excited for this, we need more accessible prototyping platforms/tools. I ended up going down that rabbit hole myself a few months ago and built this[1] but it would be a pain to commercialize.

[0] https://youtube.com/@moritzklein0

[1] https://old.reddit.com/r/synthdiy/comments/14roaib/synth_bre...

Two very very deep rabbit holes in the last 6 months:

- Designed/built a small USB controlled pan/tilt camera head to control the mirrorless I use as a webcam (couple of servos, gears, belts), and then designed/built a custom ortholinear keyboard with a joystick to control the camera (custom PCB, CNC'd aluminum case, etc)

- I'm a pretty big runner, built my own web based calendar UI that integrates with Google Calendar where I can type in workouts like "1 mile warmup @z2 + 5x(30 seconds @ 6:00/mile + 0.5 miles recovery) + 1 mile cooldown" and this gets parsed/total weekly mileage gets tallied. The next step down this rabbit hole is building a small iOS app to automatically generate Apple Watch Workouts using WorkoutKit.

I've had a similar weather experience where my internet connection dropped when it was cold. Turns out some water had seeped into the optical fiber connector, when it froze it broke the connection, and it would recover when it thawed. This one was a nightmare to troubleshoot.

In my experience, the most effective way to minimize the odds of a company never responding is to identify someone who works at that company who either you know personally (best) or they know someone you know and you ask for an intro. This allows you to skip the online application black hole. As a corollary, invest in your network and keep good relationships with former coworkers.

> Most "AI"s would be better called "Machines that use tons of statistical learning to decide their next move". ChatGPT (and similar AI) were trained on several hundred gigabytes of data, so it has a lot of raw data to train on.

Whenever I read this argument I ask myself with a certain amount of dread: "what if I'm nothing more than a machine that uses tons of statistical learning to decide my own next move?".

Put differently, it's unclear to me whether we have compelling evidence that we humans, in fact, are "better" / "more intelligent" than those LLMs.

Candidly, a lot of the reasons why Clubhouse failed is because it did not follow any of those lessons.

Source: I was an early Clubhouse employee and left a few months ago.

These resonate a lot! Tiny projects are very rewarding, because you get something that works pretty quickly, this works well with my short attention span. Couple of ones I did in the last two weeks:

- A little mac app that reads my calendar, grabs the next 2-4 events and sends them over to my Vestaboard.

- Setting up Cloudflare as a failover load balancer to reach my home network (I have two ISPs for redundancy), this involved writing a small script to get the WAN IPs out of my Omada router and using that as a custom command in ddclient to update the DNS entries for both uplinks.

Over the past few years I’ve thought about building something similar to Home Assistant, but for human centered data and APIs vs home centered. Essentially a shared bus on which different data sources could be plugged into a standardized format, on which people could build dashboards and automations. Never got around to doing so, because sidetracked by other projects, but if others are interested in teaming up I’d love to work on something like that (email in profile if anyone wants to start this)

Building keyboards is such a rewarding activity. Once you go down that rabbit hole, you can go as deep as you want.

Here's one I did a couple of years ago, including a custom PCB: https://medium.com/@friggeri/the-new-blanck-keyboard-c7563c4...

Some related links if you’re interested in this kind of approach:

https://nbviewer.jupyter.org/github/friggeri/notebooks/blob/...

https://github.com/jupytercalpoly/reactivepy

https://dataflownb.github.io/

https://github.com/stitchfix/nodebook

I'm very uncomfortable with the current trend of trying to get big tech to self police. They'll never get it right because of moving goal posts and moreover I don't want a private company to have that power in shaping social discourse.

If we as a society deem that type of content to be improper and that it should not be allowed on such platforms, then let's make that explicit from a legal perspective and regulate them.

One of the things I've been struggling with is that once I've mastered a topic enough to feel comfortable writing about it, everything I write sounds obvious to me, and from there there is an easy leap to "it's probably obvious to everyone", and I start questioning whether I should write at all. Anyone has dealt with this? Tips would be appreciated.

Corollary: Any headline which ends in a question mark will lead to someone commenting with a reference to Betteridge's law of headlines.

Great feedback thanks, being able to untrack specific artists is coming in the next release!

Support for Spotify/rdio/last.fm is definitely something I'm thinking about for the future, yes!

Spotify/rdio support is one of the most requested feature so far, so it's definitely on my things-to-consider list.

Most probably in the next few months yes!

Two reasons mainly: 1. being able to scan the user's iTunes library, and 2. push notifications (could have done email, but I find it higher friction).

for now, it's iTunes focused, so I'm using that data, looking into expanding to other services/sources

Great read, reminds me a lot of Last and First Men. http://en.wikipedia.org/wiki/Last_and_First_Men

The data backs it up: http://research.microsoft.com/pubs/167719/whyfromnigeria.pdf

If you're into that kind of things, here is a similar project I did last year: http://friggeri.net/research/senate/

I might open source the app in the future.

I have been doing something somehow similar for the past ~6 months by tracking (per minute) the number of keystrokes, distance travelled by my mouse cursor, number of clicks, active app, active wifi network and free disk space. I can't wait to start visualizing these over time.

My personal machines are named after characters from Shakespeare's The Tempest: Caliban, Prospero, Miranda, Ariel, etc.

"Pie charts are great at comparing proportions within a single data set."

No they're not.

This is the partition problem[1]. So to answer your question: the problem is NP-complete, but there are pseudo-polynomial time dynamic programming solutions (see the wikipedia entry for details).

[1]: http://en.wikipedia.org/wiki/Partition_problem

I can't recommend enough A Guide to the Good Life: The Ancient Art of Stoic Joy[1]. It's a great and modern introduction to stoicism.

[1]: http://www.amazon.com/Guide-Good-Life-Ancient-Stoic/dp/01953...

This is the main reason why I went the other way round when developing Tehula[1]. The basic idea is that knowing the location of someone should be the exception and not the norm, and that the burden should be put on the one requesting the location and not the locatee (who should have the right to ignore a request too).

[1]:http://tehula.com

You might be onto something about the merge/split balance. Will definitely investigate in that direction.