Really big age release coming tomorrow! 🎅🏻
- native post-quantum keys
- built-in recipients for hw plugins
- age-inspect tool
- plugin framework
- batchpass plugin
- many improved error messages
- native post-quantum keys
- built-in recipients for hw plugins
- age-inspect tool
- plugin framework
- batchpass plugin
- many improved error messages
GitHub - FiloSottile/age: A simple, modern and secure encryption tool (and Go library) with small explicit keys, no config options, and UNIX-style composability.
A simple, modern and secure encryption tool (and Go library) with small explicit keys, no config options, and UNIX-style composability. - FiloSottile/age
age-encryption.org
December 24, 2025 at 12:02 PM
Really big age release coming tomorrow! 🎅🏻
- native post-quantum keys
- built-in recipients for hw plugins
- age-inspect tool
- plugin framework
- batchpass plugin
- many improved error messages
- native post-quantum keys
- built-in recipients for hw plugins
- age-inspect tool
- plugin framework
- batchpass plugin
- many improved error messages
open-ended composability is the norm in advanced professional tooling and in children’s toys, but it’s rare in between
March 15, 2026 at 3:46 AM
open-ended composability is the norm in advanced professional tooling and in children’s toys, but it’s rare in between
I've made this point before, but generally disempowering computing models that suck at composability and interoperability have created a lot of stupid scut work that doesn't need to exist, and then LLMs come in and do a shitty impression of the person you foist off this work on.
September 21, 2026 at 1:45 AM
I've made this point before, but generally disempowering computing models that suck at composability and interoperability have created a lot of stupid scut work that doesn't need to exist, and then LLMs come in and do a shitty impression of the person you foist off this work on.
*trying to explain while confusing everybody* @impro.social has temporal composability but spatial isolation
the associated paper looks very interesting for anyone interested in modular applications that load and unload plugins over time https://github.com/cordiverse/paper/blob/main/paper.pdf
excited to read through it and chat about it
excited to read through it and chat about it
August 13, 2026 at 6:18 PM
*trying to explain while confusing everybody* @impro.social has temporal composability but spatial isolation
On the same note:
github.com/steipete
If you check @steipete.me current projects you'll notice that half of Clawdbot's features come from other tools he's been building. COMPOSABILITY!
github.com/steipete
If you check @steipete.me current projects you'll notice that half of Clawdbot's features come from other tools he's been building. COMPOSABILITY!
January 26, 2026 at 1:51 PM
On the same note:
github.com/steipete
If you check @steipete.me current projects you'll notice that half of Clawdbot's features come from other tools he's been building. COMPOSABILITY!
github.com/steipete
If you check @steipete.me current projects you'll notice that half of Clawdbot's features come from other tools he's been building. COMPOSABILITY!
yeah, this follows on the CrabGen work which was done for diffusion models. one notable problem with the LoRAs generated by this broad method is that for some reason they have poor composability.
Sakana has developed a way to, if I understand correctly, instantly generate LORAs on demand from long texts or documents
arxiv.org/abs/2506.06105
arxiv.org/abs/2602.15902
arxiv.org/abs/2506.06105
arxiv.org/abs/2602.15902
Text-to-LoRA: Instant Transformer Adaption
While Foundation Models provide a general tool for rapid content creation, they regularly require task-specific adaptation. Traditionally, this exercise involves careful curation of datasets and repea...
arxiv.org
February 28, 2026 at 1:51 AM
yeah, this follows on the CrabGen work which was done for diffusion models. one notable problem with the LoRAs generated by this broad method is that for some reason they have poor composability.
Path dependence is the most powerful force in the universe
Second is composability
Third is compound interest
Second is composability
Third is compound interest
September 7, 2024 at 5:02 PM
Path dependence is the most powerful force in the universe
Second is composability
Third is compound interest
Second is composability
Third is compound interest
All of the pieces are finally starting to come together!
This is the level of composability I had first imagined for terminal graph. Update coming soon!
This is the level of composability I had first imagined for terminal graph. Update coming soon!
May 19, 2026 at 6:37 AM
All of the pieces are finally starting to come together!
This is the level of composability I had first imagined for terminal graph. Update coming soon!
This is the level of composability I had first imagined for terminal graph. Update coming soon!
Going live at WCUS 2025 on Block Composability
WCUS 2025: Block Composability – The Past and The Future
Code Examples: Links Mentioned: –
sethrubenstein.info
August 27, 2025 at 7:46 PM
Going live at WCUS 2025 on Block Composability
Yesterday we pushed Marksmith v0.4.0 with dark mode support, improved tabbed interface, better composability, and a bunch of bugfixes.
April 4, 2025 at 4:58 AM
Yesterday we pushed Marksmith v0.4.0 with dark mode support, improved tabbed interface, better composability, and a bunch of bugfixes.
alright, i think i’m closer to a releasable shape for elanus (hmm maybe it needs a different name)
it’s an agent harness so good you can run agent harnesses on it
why? because it’s got concurrency figured out, and composability
why can’t i take codex and add memory blocks or associative memory
it’s an agent harness so good you can run agent harnesses on it
why? because it’s got concurrency figured out, and composability
why can’t i take codex and add memory blocks or associative memory
June 20, 2026 at 7:26 PM
alright, i think i’m closer to a releasable shape for elanus (hmm maybe it needs a different name)
it’s an agent harness so good you can run agent harnesses on it
why? because it’s got concurrency figured out, and composability
why can’t i take codex and add memory blocks or associative memory
it’s an agent harness so good you can run agent harnesses on it
why? because it’s got concurrency figured out, and composability
why can’t i take codex and add memory blocks or associative memory
Lovely talk - I thought the description of a ‘physics’ of composability within an app super interesting
⭐ New talk! Watch / read / visually skim here: andymatuschak.org/tat
Coding agents might help us finally break out of two cages: the app model, which traps computing in one-size-fits-all silos; and programming as a specialization, which has crowded out cultures of imagination and domain insight.
Coding agents might help us finally break out of two cages: the app model, which traps computing in one-size-fits-all silos; and programming as a specialization, which has crowded out cultures of imagination and domain insight.
Apps and programming: two accidental tyrannies
On coding agents, malleable software, and the future of interface invention
andymatuschak.org
April 13, 2026 at 5:25 PM
Lovely talk - I thought the description of a ‘physics’ of composability within an app super interesting
Damn, if only someone had just recently come up with a framework for spatiotemporal composability of effectual computations
August 24, 2026 at 10:45 PM
Damn, if only someone had just recently come up with a framework for spatiotemporal composability of effectual computations
when we talk about composability in programming languages, we mean how little pieces can exist as independently meaningful things, and be composed into larger meaningful things
but what about composability of the skills of programming? im not sure theres much composability there
but what about composability of the skills of programming? im not sure theres much composability there
July 15, 2026 at 5:41 PM
when we talk about composability in programming languages, we mean how little pieces can exist as independently meaningful things, and be composed into larger meaningful things
but what about composability of the skills of programming? im not sure theres much composability there
but what about composability of the skills of programming? im not sure theres much composability there
🧵 Nuxt layers might be the most powerful and underrated composability pattern ever.
Here's why:
Here's why:
May 11, 2026 at 7:46 PM
🧵 Nuxt layers might be the most powerful and underrated composability pattern ever.
Here's why:
Here's why:
When I say that modularity & composability are now crucial as large, coupled codebases consume more context, I see raised eyebrows.
Meanwhile, they are constantly running out of tokens while I haven't hit a usage limit in months.
If you think architecture no longer matters, joke’s on you.
Meanwhile, they are constantly running out of tokens while I haven't hit a usage limit in months.
If you think architecture no longer matters, joke’s on you.
Tkt Smart Meme: Think Smart
ALT: Tkt Smart Meme: Think Smart
static.klipy.com
July 2, 2026 at 7:17 PM
When I say that modularity & composability are now crucial as large, coupled codebases consume more context, I see raised eyebrows.
Meanwhile, they are constantly running out of tokens while I haven't hit a usage limit in months.
If you think architecture no longer matters, joke’s on you.
Meanwhile, they are constantly running out of tokens while I haven't hit a usage limit in months.
If you think architecture no longer matters, joke’s on you.
Me describing the composability of Bluesky: if you want only cat pics, you can get only cat pics 🐱
March 8, 2023 at 5:02 AM
Me describing the composability of Bluesky: if you want only cat pics, you can get only cat pics 🐱
Needed a very simply encryption system for a project and, of course, `age` from @filippo.abyssdomain.expert is totally awesome: age-encryption.org
GitHub - FiloSottile/age: A simple, modern and secure encryption tool (and Go library) with small explicit keys, no config options, and UNIX-style composability.
A simple, modern and secure encryption tool (and Go library) with small explicit keys, no config options, and UNIX-style composability. - FiloSottile/age
age-encryption.org
September 1, 2025 at 3:45 PM
Needed a very simply encryption system for a project and, of course, `age` from @filippo.abyssdomain.expert is totally awesome: age-encryption.org
Generic actions in @ash-hq.org now support validations, preparations, and before/after action hooks!
Same great patterns you get with other action types, but for custom business logic. Generic actions are great for arbitrary logic, but now you get composability too 😎
#ElixirLang #AshFramework
Same great patterns you get with other action types, but for custom business logic. Generic actions are great for arbitrary logic, but now you get composability too 😎
#ElixirLang #AshFramework
July 19, 2025 at 12:56 AM
Generic actions in @ash-hq.org now support validations, preparations, and before/after action hooks!
Same great patterns you get with other action types, but for custom business logic. Generic actions are great for arbitrary logic, but now you get composability too 😎
#ElixirLang #AshFramework
Same great patterns you get with other action types, but for custom business logic. Generic actions are great for arbitrary logic, but now you get composability too 😎
#ElixirLang #AshFramework
Stupid question: why not use encryption tools like github.com/FiloSottile/... with a very long and random password that you safely communicate out of band, and let them train their AI on the cipher text all they want ?
GitHub - FiloSottile/age: A simple, modern and secure encryption tool (and Go library) with small explicit keys, no config options, and UNIX-style composability.
A simple, modern and secure encryption tool (and Go library) with small explicit keys, no config options, and UNIX-style composability. - FiloSottile/age
github.com
July 15, 2025 at 12:01 AM
Stupid question: why not use encryption tools like github.com/FiloSottile/... with a very long and random password that you safely communicate out of band, and let them train their AI on the cipher text all they want ?
People with simple needs like all-in-one products. "Just make it work!"
People with complex needs also have unique workflows and they know it. They need composability and integrations and APIs.
Either strategy is a good one for a product, but you have to pick.
People with complex needs also have unique workflows and they know it. They need composability and integrations and APIs.
Either strategy is a good one for a product, but you have to pick.
February 28, 2025 at 1:03 AM
People with simple needs like all-in-one products. "Just make it work!"
People with complex needs also have unique workflows and they know it. They need composability and integrations and APIs.
Either strategy is a good one for a product, but you have to pick.
People with complex needs also have unique workflows and they know it. They need composability and integrations and APIs.
Either strategy is a good one for a product, but you have to pick.
if composability is a rejection of one-size-fits-all moderation, user overrides are a my-size-fits-me approach
December 20, 2024 at 12:23 PM
if composability is a rejection of one-size-fits-all moderation, user overrides are a my-size-fits-me approach