all right this is take two of the last note I just made in which I stated that I’m taking another note that I’m just taking like in lieu of doing a text note somewhere just because I don’t have the proper system for taking text notes which is funny because that wouldn’t be hard to implement at least in an append-only kind of way, which would be really cool.
I want to be able to tweet at myself through the catalog, basically, and file away thoughts.
and each thought is taggable, and connectable, and groupable, and addable to a selection, all this.
and referenceable from other notes, et cetera.
but regardless, and has it marked in representation?
OK. Point is, because I don’t have that for text, I’m doing that here.
and this is the clear intent of this thing, which is that it’s not a task. It’s a note. It’s an understanding that I want to reconvey, re-express.
reframe
or just to express period in my outgoing work, which is that what I’m doing here, I’m actually publishing my podcast.
I’m not just publishing a website.
This is basically an asynchronous podcast.
It’s not sequential. It’s not like you have episode one, episode two.
It’s kind of like a digital garden, but it’s a podcast.
It’s something in between a digital garden and a podcast.
And it is more like a digital garden. It basically is just a digital garden.
But consuming it, I think, could be like a podcast.
You could have a podcast form of it, where you just assemble certain objects into groups.
And you just render the groups as audio, just like the snipping idea I was thinking about yesterday.
Or not snipping, stitching, actually, not snipping.
Basically, a podcast would be a very methodically, meticulously curated prune of the note tree, a curated selection for someone to take in like a piece of media
which we could represent in a standard podcast form
which we could represent in the form that I referenced well that I described in the note that I referenced in the previous message to this one
Or it can be taken raw, where these audio files are the equivalent of scrolling through someone’s Twitter feed.
You can do that.
And there’s actually a lot to gain from that, if you do that with someone like Janus, for example, or many others.
And it’s different for everyone.
So this website would be a very unique experience, I think, for both the
If we just put a gray anonymous person as the person who was accessing the site to understand something, my site would be unique in that sense,
but also it’s doubly unique because the person that’s parsing it is also unique.
I feel like everyone is going to interpret Janus’ account again, for example, in a very different way than someone else.
And Janus is a really good example of that, because I feel like we, as humans, tend to have like we fall into collectives, and we perceive things from or with the collective.
But with Janus, it’s like you almost can’t do that, because there’s not really a, like, I’m at a loss for the word.
There is no consensus, basically, by any one community as to what Janus’ work is, let alone what you should think about it.
There have been some attempts to form those consensus, or at least start to.
One example is the Holly Elmore thing.
But there have been others.
But mostly, there’s been people who are asking.
And the consensus that I find, or that I observe, is that people don’t really understand.
or they form incomplete conclusions and work from there,
or don’t, and basically just don’t, most people, but some do.
In fact, many do at this stage.
Anyway, so I think that my own media, right, because Twitter is a medium,
So the media that I’m producing, that I’m curating, to some extent, like, obviously, the first level is just, do I post this or not?
But then there’s the further levels where we actually create media, like, well, more standard media in our culture, like podcast episodes, sequences.
I think I prefer, like, a less wrong sequences type approach for presenting the podcasts.
But yeah, we probably should also have like a proper podcast on the podcast apps like Spotify, you know RSS
That would be interesting to start with RSS
Before we publish to Spotify and stuff like just have a feed that someone can just subscribe to
I mean, why the hell not?
Okay, that’s cool
realization
Anyway so It’s funny, because I had this simple thought that I wanted to note, which was just… It’s a reframe, that’s the thing.
It’s a reframe for myself, because I’ve been thinking, like, I wanted to publish a website last week, right?
Like, at this point, it’s over last week, like, last weekend.
And I didn’t, and I felt bad.
And I felt a bit stressed.
And I was like, especially as I realized, like, this is not as polished as I wanted it to be.
especially by now, considering that I’m now where I wanted to be a week ago, over a week ago, right?
I wanted to be this far two weeks ago, or whatever.
I mean, really, months ago.
But something that I’m realizing is, like, this does have value.
It’s okay that I’m taking months to produce this stuff, because it’s actually a lot of stuff that you can, like, someone could spend months pouring over this.
And especially working with what I have poured over and expressed in this stuff.
I can provide pointers.
I can share resources.
We could have a community to discuss topics that these point to.
I can build a platform.
I can open source the project and make it easy for someone else to publish their thoughts as well, to start doing so.
And we can build some kind of like, I mean, would it be local first? Or would it have a server? I don’t know.
new-idea
But there was this. I’m going to go into an idea now.
And this definitely should be a new section.
But I’m going to continue that thought, which is that.
I don’t think I’ve actually recorded this anywhere, but I had a thought with Shani that I may have recorded once or twice.
I don’t think so, though.
The thought was basically like every media object can be shareable.
And I kind of was getting at this, was that in this message?
No, it was in the previous one where I was saying the grouping task.
It would be possible to share a link with someone
And right now, I’m thinking very solo scale.
The whole system, it’s one single JSON file.
It’s the entire database, and it’s all my content.
But what if we had a social media site level database where these objects are not my objects, but it’s just objects on the platform that the owner is me, and the person with access to it is me.
And let’s think about Arena here.
When you put content on Arena in a private channel, the content is not stored in your server, the content is stored on Arena’s servers, but no one outside of your account or any organizations that have access to the channel that you’ve put your block into have access to the content that you upload onto their site, that you publish on their site.
We could follow the same model, where we can do local first.
And we definitely support local first and local only.
But you can also use the server.
And whenever you synchronize with the cloud and or share with someone, your object will be published in the way I just described where it’s still your object and only while someone has permission is your object accessible at all externally
and even then it won’t be accessible to the public unless you explicitly specify to do so
whether on the group that it’s part of kind of like Spotify playlists
or at the individual level like per block kind of like Notion or Google Docs
And what this would allow us to do is that we could have the ability to curate your own collections with playlists of anyone’s public objects.
And very interestingly, we could also start indexing, processing podcasts from YouTube, from Spotify. and so forth from RSS.
And we don’t necessarily have to host it, especially when IP starts to become a concern, which it would in this case, unless you’re doing purely local peer-to-peer.
But if we have a platform, it would have to be this way.
But I think this would be very good.
Like, when you add a YouTube object to your library or whatever, let’s say you have x processing minutes per month.
If you add an hour-long YouTube video, cool.
Transcribe, process with Haiku, great.
Now you have your speech data for that podcast or whatever.
And as we get 4.0, et cetera, we start to add further features
Like basically we just do the same thing like, you know, maybe you add a song, and it analyzes the sections and it creates like an equivalent of speech data for music.
And maybe that is actually itself dynamic, and we have UI components that can represent any kind of structure.
within the constraints that we build it to.
But we look at media and we’re like, okay, most media can be represented in a structured way, the same way that there’s these Twitter accounts that just run something through and be like, create a mermaid graph of this.
We could do the same freaking thing.
where we create structured representations of the information that’s in the input media.
But that doesn’t have to be verbal or linguistic at all.
It can be conceptual.
It can be vibe-based.
It can be visual.
It can be sonic, whatever.
But I think that would be, obviously, that would be well off.
But the point is, I can transcribe things.
I can process them.
I can create snips.
I can tag the snips.
I can highlight them.
I can create collections of them.
But by using this service, I’m not locking myself or my friends into the service.
It’s not something that feels annoying to use.
It’s something that feels delightful to use.
It is something that is very friendly, like the way that Arena is very friendly.
It lets anyone download public channels as a whole.
They do the same thing.
You can download a given group, a given…
A given group will have a download button where you can download the entire thing.
And when I say the entire thing, you get an ordered list of files or set of files that contains the Markdown representations of each object as Markdown.
And for whatever we are storing on the platform, you get the audio as well, like, for example, for voice notes.
And maybe that would be trimmed audio, like, for example, selections.
When you click download there, you have the base audio file that maybe the person who made the collection doesn’t even have access to the base audio file.
They only had access to the snip that someone created, the selection.
And the difference between this and a snip is that a selection can have multiple snips to form one coherent piece, one expression, a curated expression, a pruned expression.
Like if you use Descript for 20 minutes on a given video file, you might have several selections, which would be the sections of that, for example, or which might be.
And so the point is, for example, if I upload a Descript-published or Descript-processed video, no one has access to the B-roll or whatever, because I edited that out.
with this system, the content would still be on the server.
And it still could be published if the original owner either made new selections and published those, or if they published the entire thing and somebody else decided to create their selections.
But the point is, we would handle selection rendering from the root as available.
And anyway, that could be really interesting, I think.
I mean, I think all of this could be really interesting.
conclusion
And I’m also, wow, I look at the time, and I’m very frustrated that this was a 15-minute file.
But this was pretty high signal and high value and high intent.
So I want to follow through on as much of this as possible in the near term, especially the initial reframe, but also in creating and adding these tasks to the roadmap and maybe as a project or whatever.