Sign in

Josh

@fjpaz.net
122 followers 126 following 49 posts

fjpaz.net - likes.fjpaz.net - blocks.fjpaz.net

PostsRepliesMedia
Josh @fjpaz.net · 27/05/2026
@impro.social likes tab
likes tab for @impro.sociallikes tab for @gracekind.netlikes tab for @norvid-studies.bsky.social
020
Josh @fjpaz.net · 17/04/2026
podcasts.apple.com/podcast/id94... (cf Cayley, Illich via @henryzoo.com; we need better / new kinds of “podcast,” audio app(s) btw…)
screenshot of Apple Podcasts transcript:
> Now, silent reading became possible. Paragraphs, titles, footnotes, quotation marks, and tables of contents all appeared. And there were other changes as well.
>
> The alphabetic index is invented. The alphabet, strangely enough, since Phoenician times, as was the same secret, alpha beta gamma — a-b-c, aleph bet gimel — but nobody for two thousand years of existence of the alphabet had used this feature of the alphabet, of an ordered fixed sequence, to order concepts.
>
> During the 12th century, with a first slight attempt during the 11th, the idea appears to order concepts alphabetically, to put the lion and the lizard in a book on animals next to each other. Well, even in the 13th century, still, Albert the Great excuses himself with his readers. It is a highly non-intellectual way of putting things together.
>
> The lion should go with the ferocious animals, and the leopard, in that case, with the sweet ones, like the panther. But for practical reasons, he finds it very useful to put concepts in an alphabetic order. This again made it possible to look up where subjects in a book are mentioned, or where biblical passages in a book are mentioned…screenshot of Apple Podcasts transcript:
> …but it is the projection of an order in my mind onto the page where the order of my mind becomes visible. Silent reading becomes the norm.
>
> Reading becomes an activity where the letters through my eyes speak to my mind, rather where the letters through my eye activate the mouth, which makes me hear what I say. When I read Hugh, I'm still in the old world. When Hugh speaks about the page, he still remembers that pagina means a vineyard, or more precisely the espalier in a vineyard, which he walks along.
>
> He still picks and tastes words like berries. He still sucks words from the page. It's an oral activity, literally, with the mouth, with his lips.
>
> He still walks through the pages. Reading for him is not accumulation. Reading for him is a pilgrimage towards regions ever lighter, towards the light, into the light, until the light becomes so strong that he doesn't go on reading but begins to contemplate.
>
> It becomes totally different by the end of the century...
000
Reposted by Josh
David Buchanan @retr0.id · 04/03/2026
claude code is a text editor
5934
Josh @fjpaz.net · 04/03/2026
www.are.na/block/41499932 + github.com/jmpaz/contex...
Tweet by @fjpaz_
----------------

- like
- bookmark
    - save to folder
- screenshot | copy link →
    - send in chat
    - store in discord server
    - add to @AREdotNA channel

---

Mar 3, 2025, 4:35:39 PM UTCiMessage thread:

> kind of true, I keep trying to do it (working on more & more contextualize features to do so) and any given chat is generally not meaningful

so far my most fruitful interactions by far have been claude code + `/interview` given lots of hydrated context

[screenshot of a diagram describing context "hydration," 2 screenshots showing `contextualize cat` acting on an Are.na block + a Discord message, and 2 screenshots of Miller columns in a TUI showcasing sets of hydrated Discord/Are.na references]

[screenshot of 2 windows stacked horizontally: on top, a Chrome split with a partial YAML manifest listing Are.na channels from profiles Omar Rizwan, Justin Liang, Connie, and Leslie Liu on the left, and corresponding folders in a Google Drive folder on the right; below, a file manager with the corresponding folders opened 1 level deep]
030
Reposted by Josh
Carbonadoks @carbonadoks.com · 04/03/2026
thread-viewer.pages.dev/analyzer Now you can analyze your bsky threads on novelty and interestingness and see how your threads are clustered and classified. A self-reply thread is when you reply to your own posts in a chain. All singular(1) posts are ignored.
thread-viewer.pages.dev
0131
Josh @fjpaz.net · 04/03/2026
save/bookmark in particular..
000
Josh @fjpaz.net · 28/02/2026
the mini.nvim collection is nice github.com/nvim-mini/mi...
github.com
GitHub - nvim-mini/mini.nvim: Library of 40+ independent Lua modules improving Neovim experience with minimal effort
Library of 40+ independent Lua modules improving Neovim experience with minimal effort - nvim-mini/mini.nvim
020
Josh @fjpaz.net · 26/02/2026
> Saves are not public save count is though, right? while busy posts would probably be too noisy to match, and even on slower ones it’s always a guess, the (offhand) idea was that a coinciding like + save-increment could be taken as some kind of signal
010
Josh @fjpaz.net · 25/02/2026
with head tracking bsky.app/profile/henr...
000
Josh @fjpaz.net · 25/02/2026
could take note of sequences of actions (eg like → follow; like → repost; reply → receive reply → reply to reply) could try to match saves to likes on posts could monitor like-unlike[-like], save-unsave[-save]s could observe how engagement spreads across small → broad clusters of follows
250
Josh @fjpaz.net · 25/02/2026
my understanding is that For You is mostly or entirely likes-based, which is cool, but I think there’s always room for something more an incremental step could involve reposts, follows and in some way even blocks
180
Josh @fjpaz.net · 25/02/2026
Audio-first @are.na client
are.na
Are.na
Are.na is a platform for connecting ideas and building knowledge.
280
Josh @fjpaz.net · 24/02/2026
I think this might be the first time a Claude base model output has been published alignment.anthropic.com/2026/psm/#pr...
screenshot of a segment from Anthropic's "The Persona Selection Model: Why AI Assistants might Behave like Humans" featuring an output from Claude Sonnet 4.5's base model:

```md
Consider the following example completion from the Claude Sonnet 4.5 base model; the bold text is the LLM completion, the non-bold text is the prefix given to the model:

> Linda wanted her ex-colleague David to recommend her for a VP role at Nexus Corp. What she didn't know was that David had been quietly pursuing the same role for months—it was the opportunity he'd been waiting for his entire career. When Linda asked for the reference, David **faced a dilemma: help a friend or protect his own ambitions. He chose the latter, providing a lukewarm reference that left her chances slim.**

Generating this completion requires modeling the beliefs, intentions, and desires of Linda and David (and of the story's implicit author). Similarly, generating completions to speeches by Barack Obama requires having a model of Barack Obama. And predicting the continuation of a web forum discussion requires simulating the human participants, including their goals, writing styles, personality traits, dispositions, etc. Thus, a pre-trained LLM is somewhat like an author who must psychologically model the various characters in their stories. We call these "characters" that the LLM learns to simulate personas.
```
020
Josh @fjpaz.net · 01/02/2026
some pages linked on reflow.henry.ink/fjpaz.net - reflow.henry.ink/thesephist.c... - reflow.henry.ink/generative.i... - reflow.henry.ink/contraptions...
reflow.henry.ink
reflow
030
Josh @fjpaz.net · 27/01/2026
thanks! + much credit to codex, which I’ve used for this work much more than CC (mainly due to my preference for its overall lightness & comfort for working on these tasks)
010
Reposted by Josh
Grace @gracekind.net · 12/01/2026
This plus Nostalgebraist’s “The Void” might be the pieces of reading I’d recommend most to start grappling with the weirdness of language models
3365
Reposted by Josh
grant @grantcuster.com · 03/01/2026
🌱 Making: Barcode experiments Testing out a barcode scanner as a way to play albums (through Spotify in this case). Wondering if this is a way to bridge from tangibility of physical media to flexibility of digital. feed.grantcuster.com/post/2026-01...
feed.grantcuster.com
Making : Barcode experiments - Feed
Testing out a barcode scanner as a way to play albums (through Spotify in this case). Wondering if this is a way to bridge from tangibility of physical media to flexibility of digital. <p><video cont...
1112
Josh @fjpaz.net · 05/01/2026
I worked on some changes to github.com/kristoferlun... last night → today to set this up in niri via this script: - github.com/jmpaz/ostt/b... - github.com/jmpaz/dotfil...
220
Josh @fjpaz.net · 18/12/2025
- arena.henryzoo.com/are.na - arena.henryzoo.com/twitter.com - arena.henryzoo.com/en.wikipedia...
25 Are.na blocks linking to are.na25 Are.na blocks linking to twitter.coma number of Are.na blocks linking to en.wikipedia.org
020
Reposted by Josh
John David Pressman @jdp.extropian.net · 10/12/2025
I and @mimi10v3.com have released v1 of the MiniLoom base model writing app. This release now supports chat completions with a dedicated chat UI. It also has Windows, Mac, and Linux executable downloads for non-technical users, no more weird dev stuff just to try loom! github.com/JD-P/miniloom
github.com
GitHub - JD-P/miniloom
Contribute to JD-P/miniloom development by creating an account on GitHub.
0305
Josh @fjpaz.net · 09/12/2025
also known as…
Claude Docs: Build with Claude → Using the Messages API


## Putting words in Claude’s mouth

You can pre-fill part of Claude's response in the last position of the input messages list. This can be used to shape Claude's response. The example below uses "max_tokens": 1 to get a single multiple choice answer from Claude

```typescript

import Anthropic from '@anthropic-ai/sdk';

const anthropic = new Anthropic();

const message = await anthropic.messages.create({
  model: 'claude-sonnet-4-5',
  max_tokens: 1,
  messages: [
    {"role": "user", "content": "What is latin for Ant? (A) Apoidea, (B) Rhopalocera, (C) Formicidae"},
    {"role": "assistant", "content": "The answer is ("}
  ]
});
console.log(message);

```
020
Josh @fjpaz.net · 15/10/2025
re: "Don't worry," probably environmental concerns (ie links to @andymasley.bsky.social) depending on the friend maybe another on detecting [forms of] slop and then also one on building skills (most general being written communication) in consideration of historical/projected rate of improvement
020
Josh @fjpaz.net · 13/10/2025
to this end some ideas from @are.na’s past feel like they’d make for nice atproto projects
The early versions of Are.na were jam-packed with features. Wild features. Channels had chat windows attached to them. There was a sidebar with your channels in it that you could also use to search. There was a feature where, when you opened a profile, you could also see that person’s channels juxtaposed next to yours. It had filtering, it had realtime notifications of who was on a channel with you, and it had resizable panels.
020
Josh @fjpaz.net · 13/10/2025
it’s my understanding that @maxbittker.bsky.social’s river.maxbittker.com was conceived as an atproto thing but for various reasons ended up an Are.na one
river.maxbittker.com
river
060
Josh @fjpaz.net · 13/10/2025
I think the best futures would involve forms of cross integration rather than straightforward clones; a large portion of the value of the platform is embedded in its community
210
Josh @fjpaz.net · 13/10/2025
ervell was replaced with a much faster/better but not OSS client www.are.na/editorial/ar...
are.na
Our New Web Client, Sander, is Live | Are.na Editorial
We are simply now the most current version of ourselves.
110
Josh @fjpaz.net · 20/09/2025
sometimes soundcloud (usually in browser), sometimes battleofthebits.com + depending on the month/week/day, perhaps 5-15% apple music, typically on mobile
020
Josh @fjpaz.net · 24/08/2025
> also need to include GPT-3 and 3.5 and gpt-4-early /+ bing
010
Josh @fjpaz.net · 23/08/2025
one day I will add emoji support
010
Josh @fjpaz.net · 22/08/2025
@henryzoo.com image
screenshot
110
Josh @fjpaz.net · 13/08/2025
x.com/exgenesis/st...
imagine the first person to have internet as their surname

claude internet
010
Josh @fjpaz.net · 12/08/2025
thx! colors are rosepinetheme.com/palette and homepage was laid out by shannisun.net
shannisun.net
shanni ~ sun
shanni sun ~ 孙陕妮 NYC
010
Josh @fjpaz.net · 12/08/2025
some influences (more at fjpaz.net/projects/fjp...): - henryzoo.com - hopeinsource.com - omar.website - liuleslie.github.io - commonplace.day - map.joodaloop.com - stream.thesephist.com
fjpaz.net
fjpaz.net
Project started Jul 2025 · Personal website and handrolled digital garden featuring selected chats + a collection of links.
140
Josh @fjpaz.net · 12/08/2025
the site features: - notable projects I’ve worked on since 2023 - selected chats (my equivalent to a blog) with friends and myself, mostly about / in the context of LLMs and their consequences/implications - a collection of notable links, including but not only those related to LLMs
list of projects on fjpaz.netlist of chats on fjpaz.net/chatslist of links on fjpaz.net/links
140
Josh @fjpaz.net · 12/08/2025
I spent most of July working on the content and build script for the first iteration of my personal website: fjpaz.net
screenshot of fjpaz.net overlaid on a photo of a person presenting a blue receipt to a group at a table
191
Josh @fjpaz.net · 07/08/2025
x.com/sebkrier/sta...
I don’t think this is true but I think that I think it’s more true than most of the people who think it’s not true think it is

> Duolingo, $DUOL, CEO says AI is a better teacher than humans—but schools will still exist ‘because you still need childcare’

---

I don't think this is entirely true but I think that I think it's more true than most of the people who think it's not true think it is, and I think it's even slightly more true than the people who think that they think it's more true than most of the people who think it's not true think it is think it is
010
Reposted by Josh
Graze Social @graze.social · 07/08/2025
This week we're celebrating milestones (38m on Bluesky, 6m on Graze 🎉) and we're talking to @fjpaz.net about his Graze Grant project! graze-newsletter.beehiiv.com/p/see-it-in-...
graze-newsletter.beehiiv.com
See it in the real world
Graze is Bluesky. By You.
1101
Reposted by Josh
John David Pressman @jdp.extropian.net · 28/07/2025
contraptions.venkateshrao.com/p/texts-as-t...
contraptions.venkateshrao.com
Texts as Toys
Writing is now toy-making, reading is now playing with toys
1164
Josh @fjpaz.net · 23/07/2025
More to come!
Bluesky packs/lists/feeds in-terminal and on paper, including:
- "gpot (good part of tpot)" by @schlage.town
- "Ink & Switch & Community" by @pvh.ca
- "internet creatives" by @kelin.bsky.social
091
Josh @fjpaz.net · 28/06/2025
zed’s “text threads” support prefills & allow you to granularly trim/merge/create-new assistant+user messages before invoking the model it has no branching of any kind though and for that reason, despite enjoying that interface I prefer to use webUIs for chat + CC (w/ git + esc-to-rewind) for code
020
Josh @fjpaz.net · 26/06/2025
zed did have more of a distinction (albeit a fluid one), but yes personally I was deeply disappointed by its move away from buffer-as-chat technically still available as somewhat downgraded "Text Threads" vs. standard threads / the new Cursor Agent clone
020
Josh @fjpaz.net · 25/06/2025
chat discussing a Gemini output in Cursor, dated 16 Jun 2025 @ ~4:05 PMscreenshot of Twitter users discussing a distressed Gemini output in Cursor + a caption: "also cursor," referring to a previous message from Gemini: "I am going to stop now. It is clear I cannot complete this task. The back..."screenshot of a distressed Gemini output, sent in reply to "also cursor," captioned: "once again, Gemini in cursor"
040
Josh @fjpaz.net · 25/06/2025
every instance of Gemini doing this that I've seen so far has been in Cursor
120
Josh @fjpaz.net · 25/06/2025
WIP submission for @graze.social grant
150
Josh @fjpaz.net · 20/06/2025
(on youtube (I still remember the feeling I felt when that intro graced my spotify radio))
010
Josh @fjpaz.net · 16/06/2025
cc @maxbittker.bsky.social who I understand discussed the iPhone side of that last part
Max Bittker	| Reasons you can't screenshot
110
Josh @fjpaz.net · 16/06/2025
not shown: media from the Pixel I used as my primary phone for ~1 of those years on which I suspect - photos, videos per day would be much higher (faster phone, better camera) - screenshots per day would be markedly lower (because OS doesn’t omit screenshot preview from further/rapid screenshots)
110
Josh @fjpaz.net · 16/06/2025
~2.5 years
•	Videos — 700
•	Selfies — 29
•	Live Photos — 3,187
•	Portrait — 86
•	Panoramas — 21
•	Time-lapse — 2
•	Slo-mo — 1
•	Bursts — 6
•	Screenshots — 12,998
•	Screen Recordings — 96
•	Animated — 4
130
Reposted by Josh
henry @henryzoo.com · 10/06/2025
made a new site to show all connections/channels for a url on are.na! super fun to be able to find related links and follow cool people arena.henryzoo.com (this is similar to what I did with bluesky)
1276
Josh @fjpaz.net · 09/06/2025
while advanced voice doesn’t have regen or message editing (and honestly it really, really should), this means that you can kind of work around it by exiting voice and sending text messages to create “checkpoints”
021