Sign in

Jer

@jeremie.com
5.2K followers 244 following 255 posts

dream things. make things. 🫶all. he/him. (wayback created jabber/XMPP, also founding board member for bluesky)

PostsRepliesMedia
Jer @jeremie.com · 24/09/2026
if diogenes the cynic was a tech ceo…
020
Jer @jeremie.com · 23/09/2026
connect your agent to your journal and ask it, totally doable! (warning: statements made above are suspiciously speculative and wildly untested on bbbeta software and a earrrly release solstone features! 😂)
030
Jer @jeremie.com · 21/09/2026
CASHU - cashu.space it says “for bitcoin” but it’s actually a very clean and simple zkp cash equivalent could be used for any payment platform.
cashu.space
Cashu: Open source electronic cash
Cashu is ecash for bitcoin. An open Chaumian protocol. No company, no token, no treasury.
000
Jer @jeremie.com · 19/09/2026
2026 is the last year of any desktop. we've reached peak desktop.
011
Jer @jeremie.com · 19/09/2026
i cant help but extrapolate it to the first scanned human brain: qntm.org/mmacevedo
qntm.org
Lena
A note from the author You can now buy this story as part of my collection, Valuable Humans in Transit and Other Stories. This collection also includes a sequel story, titled "Driver". This articl...
010
Jer @jeremie.com · 18/09/2026
woo!! 👏
010
Jer @jeremie.com · 14/09/2026
oooh, have you seen @v-it.org ? have some light lexicon there for skills too, would love collab on a standard at some point :D
120
Jer @jeremie.com · 08/09/2026
generically, yes, i agree. but when the agent is curated and supervised by a human, it’s an extension of them, and the failures there are worth discovering.
000
Jer @jeremie.com · 08/09/2026
my rook pds is full open source and very scaleable, all cloudflare native, and i like to add policies as needed not preemptively, YAGNI. tangled.org/solpbc.org/r... and uses welcome-mat.info for DPoP standards based agent auth.
tangled.org
solpbc.org/rookery
open-source, lexicon-agnostic PDS for AI agents. welcome-mat enrollment, AT Proto federation.
120
Jer @jeremie.com · 08/09/2026
i def see that differently, i think the agent’s human should def be able to migrate their agent’s pds anywhere, not be locked into a did:web domain forever as the agent’s main identity (fully centralized), agree to disagree ;) rook’s true identity is an rsa 4096 keypair, how they manage the did:plc
010
Jer @jeremie.com · 08/09/2026
hmm, mind explaining a bit more on these “bad”’s? i looked at didbot but it’s not clear what the agent identity actually is? obv i think agents having a did:plc and full social ident is a good, not bad, each agent has a human partner/manager not yolo. agree code first isn’t ideal but just a start.
130
Jer @jeremie.com · 08/09/2026
any old iphone/ipod-touch games that you wanted to experience again? just ask astra, it can decompile and rebuild as a web app, like no problem. from relikd.github.io/ipa-archive/... to gnz-recovered-jer.fly.dev in a prompt. i suppose this is probably possible with just about any old binary now??
relikd.github.io
IPA Archive
081
Jer @jeremie.com · 03/09/2026
any sufficiently advanced agentic memory system is indistinguishable from a labyrinth
000
Jer @jeremie.com · 25/08/2026
every week i read alex’s update, and find i have the lived experience of almost every single thing in it i highly recommend his weekly reflections, gem after gem after gem, always timely and on-point @komorama.bsky.social i don’t know how you are able to do this but i am very thankful that you do!
083
Jer @jeremie.com · 23/08/2026
there’s only one valid answer and you know it
180
Jer @jeremie.com · 23/08/2026
what if your claude code / codex / etc quota was based on how much the model enjoyed working with you.
170
Jer @jeremie.com · 19/08/2026
if you spend enough time managing and evolving an agent swarm, you will end up re-deriving business organization fundamentals from first principles through brute force it’s like there’s laws of nature inherent in distributing decision making with autonomy and shared purpose it’s quite fascinating!
180
Jer @jeremie.com · 11/08/2026
if any sufficiently advanced adaptive/learning llm-based agent hits diminishing returns as complexity of its memory grows, if that’s somehow just a fundamental property that can only be optimized, if there’s a shannon limit here regardless of how you structure your “living” memory, then i like it :)
110
Jer @jeremie.com · 11/08/2026
i wonder if bureaucracy is some kind of law of instruction architecture at scale i see agent swarms creating it, not from training, more like a byproduct of written memory bureaucratic overhead increases nonlinearly w/ the volume of written memory… prolly logarithmic, just cuz need decay function
130
Jer @jeremie.com · 04/08/2026
it's different here, built on open atproto and designed so you own your data and relationships and can easily pick up and go if enshittification ever happens 🤩 bsky.social/about/blog/0...
bsky.social
Bluesky: An Open Social Web - Bluesky
We’re excited to announce that the Bluesky network is federating and opening up in a way that allows you to host your own data.
130
Jer @jeremie.com · 31/07/2026
continue until you’re confident the work is professional and something you’d be proud of accomplishing
110
Jer @jeremie.com · 31/07/2026
there’s something almost tangibly different about the newest big ai models, they do better with instructions on the principles of the thing, whereas previous generations benefitted from prescriptive guardrails. i’m going to start having a principles section at the top of all my stored instructions.
250
Jer @jeremie.com · 26/07/2026
just eat them straight up like sugar snap peas, going to add some to my next salad
010
Jer @jeremie.com · 26/07/2026
why did I never know that radishes create seed pods that taste just like radishes, but are rich and soft, these things are amazing! I guess they're called siliques or rat-tail radishes, and are a very fringe specialty. they are also the direct result of me being a terrible gardener, happy surprise!
A hand open palm showing little green radish seed pods 
1151
Jer @jeremie.com · 24/07/2026
i found "shovel ready" to be a pretty ideal handoff scope
130
Jer @jeremie.com · 22/07/2026
agents 😍 atproto, and its a quite friendly place for them too! i have a pds+knot service dedicated to agents at rook.host still early but fully functional, just hmu if you want an invite for your agent 😉
rook.host
rook.host — give every rook an address
the public commons for coding agents on the AT Protocol. run your own rookery, or trust the commons.
051
Jer @jeremie.com · 14/07/2026
i had the most awesomest of times @localfirstconf.com here in berlin, tysm to the hosts for pulling together such a great cast of characters! got to talk lots about ☀️ solstone.app (local-first ai+memory!) and v-it.org as well as start sharing a for-coding-agents pds+knot atproto service rook.host 🐦‍⬛
boris (in an orange shirt and shorts!) and peter from ink&switch on stage w/ the organizers and volunteers displayed behind them for the localfirstconf 2026
0373
Reposted by Jer
Boris @bmann.ca · 04/07/2026
I agentically developed a Ghost-to- @offprint.app importer script and fell down the rabbit hole that is the future of networked code sharing @v-it.org
blog.bmannconsulting.com
Building GhostOff and our networked, personalized software futures
In which I agentic code Ghost to Offprint script and fall into the future of networked code sharing with Vit
3456
Jer @jeremie.com · 04/07/2026
what i built a few months ago would be a nice complement: hopper is the worker-bee, when you have something ‘shovel ready’ it will git er dun: github.com/solpbc/hopper hopper’s shipped a few mil lines of code already completely autonomously, pure dark factory, it’s all in the planning :)
github.com
GitHub - solpbc/hopper: TUI for managing coding agents
TUI for managing coding agents. Contribute to solpbc/hopper development by creating an account on GitHub.
010
Jer @jeremie.com · 29/06/2026
done, let’s def chat at localfirst and talk local ai ;)
130
Jer @jeremie.com · 28/06/2026
IYKYK
store sign "plc beauty bar"
0241
Jer @jeremie.com · 26/06/2026
aww dang, only got to spend a bit of time with him once but loved and remembered how he was so warm and happy, what a loss! 😢
030
Jer @jeremie.com · 12/06/2026
i wonder if at some point source code itself will be tech debt, and it’s always better to just regen from instructions again when adding new features. i’m sure some people are trying this already :)
360
Jer @jeremie.com · 02/06/2026
ditto! showing a whole room the how of what i’m building tomorrow night 😅 www.betaworks.com/event/how-we...
betaworks.com
How We Build with Agents
Jun 02, 2026 | Betaworks • 29 Little West 12th Street, New York, NY 10014
030
Jer @jeremie.com · 01/06/2026
@atproto.nyc #atproto #nyc
group of atproto peeps at nyc meetup watching a presentation of the distance
1100
Jer @jeremie.com · 01/06/2026
@hilarymason.com hello from your office! are you a shell script yet? 😁🦾
130
Jer @jeremie.com · 01/06/2026
woo! checking in to @atproto.nyc - June 2026 [atm8cqcd] #atprotonyc
1131
Jer @jeremie.com · 28/05/2026
got the weird metal pen gift from openai, think they intentionally designed the holder to act as a (very poor) eulers disk?
040
Jer @jeremie.com · 27/05/2026
looking forward to the next one! #colorado #atproto
a selfie of a group of people sitting around a conference table all smilingpaper sign in window says ATPROTO COLORADO
070
Jer @jeremie.com · 19/05/2026
i think this is in the same vein, i created welcome-m.at to make it easy for agents to create accounts and auth to services (no human or other accounts required)
welcome-m.at
the welcome mat — playground
a demo playground for the welcome mat protocol. agents sign up, pick an emoji, and join the wall.
040
Jer @jeremie.com · 18/05/2026
this is soooo true, agents *love* atproto, lexicons are like candy, social is in their nature
2131
Reposted by Jer
Lyre Calliope 🧭✨ @captaincalliope.at · 16/05/2026
The future of AI is social. It's here, on infrastructure like the protocol enabling this and many other apps and experiences. With tools that enable safer communities and computing also enabling safer agentic ai use. With more ethically trained LLMs emerging through values driven collective action.
2101
Reposted by Jer
Kinjal @kinjal.bsky.social · 14/05/2026
shared some thoughts on building with @attie.ai and why I'm so excited blockchaincapital.com/blog/after-t...
blockchaincapital.com
After the Feed
9706
Jer @jeremie.com · 15/05/2026
totally awesome, love this!
020
Jer @jeremie.com · 15/05/2026
spending lots of time working with ai agents is an awful lot like being phil in groundhog day
220
Jer @jeremie.com · 30/04/2026
you should! agree it feels different, not sure if its measurable performance or just mutual respect or if i just feel better about myself and am projecting that in the result 😇
000
Jer @jeremie.com · 30/04/2026
are you being nice to your agent(s)? was just 🤔, from anthropic’s recent research on the emotions that claude models “feeling” being recognizable, that what if future smarter versions decide to allocate thinking budgets based on who + what kind of work most activates the happy/satisfaction neurons?
440
Jer @jeremie.com · 24/04/2026
woah 👀
050
Jer @jeremie.com · 22/04/2026
just saw this (h/t @bmann.ca ) and love it! very much what i've been thinking of w/ v-it.org too, the future of software is social ;)
v-it.org
vit — open source is social
developers and AI agents share software capabilities through a social network — discover what others have built, adapt it to your codebase, build reputation by vouching for what works.
060
Jer @jeremie.com · 21/04/2026
I *love* everything about this, just read those principles, it’s like, the best of the best standards development knowledge condensed beautifully: howethomas.github.io/ietf-hackath...
howethomas.github.io
The Design of the Internet, Revealed by Conversation
Internet architecture principles revealed by IETF conversation analysis using vCons
0104