@zerosnacks

Building systems for collaborative intelligence @selemis_com | Previously Foundry & Alloy at @tempo, @ithacaxyz, @paradigm.

github.com/zerosnacks
Joined October 2021
  • Tweets2.2K
  • Following828
  • Followers1.6K
137Media

Introducing Kival

Article · @zerosnacks
2
1
16
2,933
zerosnacks retweeted
prof-g
@prof_gOct 23
increasingly concerned that ai is revealing the mathematical world as negatively curved. instead of everyone wandering about in idea-space and clustering around the same famous conjectures (local positive curvature), ai is pushing out the boundaries rapidly and revealing that the space of mathematical ideas is ~hyperbolic (negative curvature). though this means there's lots left to discover, it also means that small separations between people/agents will lead to wandering alone. we may be emerging into an idea-space where there's so much to see and do but everyone is alone in a private heaven, unable to share their subtle joys with others...
201
171
72
2,442
395,748
If you are not using agents for personalized daily briefings you are missing out! I have one monitoring for interesting meetups and events in my area, another monitoring for commits on projects I like, another for specialized news feeds and one for a daily exercise and food plan.
4
207
recommended reading by @mdarcemont manuel.darcemont.fr/posts/th…

The last time my family was replaced by technology

My great-great-grandfather thought he’d be replaced by technology too. He was a farrier in Mandres-en-Barrois, the small village in rural France where I grew up.

manuel.darcemont.fr
5
12
119
12,938
zerosnacks retweeted
zerosnacks
@zerosnacksSep 14

Introducing Kival

Article · @zerosnacks
2
1
16
2,933
zerosnacks retweeted
zerosnacks
@zerosnacksSep 16
Kival is up on Show HN 😀 news.ycombinator.com/item?id…
2
4
239
Kival is up on Show HN 😀 news.ycombinator.com/item?id…
2
4
239
Did you already try out the demo for Kival? Would love to hear what you think 🙏
zerosnacks
@zerosnacksSep 14
Replying to @zerosnacks
Try out the live ACME demo in < 1 minute here: github.com/selemis-com/kival…
1
3
353
zerosnacks retweeted
zerosnacks
@zerosnacksSep 14
Try out the live ACME demo in < 1 minute here: github.com/selemis-com/kival…
2
1
6
559
Introducing Kival, a self-hosted collaborative knowledge system for organizations.
zerosnacks
@zerosnacksSep 14

Introducing Kival

Article · @zerosnacks
1
1
14
819
Try out the live ACME demo in < 1 minute here: github.com/selemis-com/kival…
2
1
6
559
zerosnacks retweeted
zerosnacks
@zerosnacksAug 31
What if CLIs were actually optimized for agent use? Argx is an expressive, derive-first CLI and configuration framework for Rust, designed for humans and agents. Instead of parsing --help, agents get commands, arguments, results, errors, and schemas as structured data.
Rust example showing Argx deriving a typed CLI command with input, output and error types, and associating them with a command handler.

ALT Rust example showing Argx deriving a typed CLI command with input, output and error types, and associating them with a command handler.

1
3
1
20
1,242
zerosnacks retweeted
zerosnacks
@zerosnacksAug 31
Argx exposes the command surface as a machine-readable interface. Agents can inspect the command structure and determine what can be invoked, which inputs are required, and what results or errors to expect. github.com/selemis-com/argx

GitHub - selemis-com/argx: Expressive command-line parsing and configuration for Rust.

Expressive command-line parsing and configuration for Rust. - selemis-com/argx

github.com
1
2
7
290
Introducing `clap_schema`: machine-readable contracts for Clap CLIs. Give agents JSON Schema for commands, arguments, and the typed Rust results they return, all derived from the types you already use and checked at compile time. github.com/selemis-com/clap_…

GitHub - selemis-com/clap_schema: JSON Schema generation for Clap with typed Rust results.

JSON Schema generation for Clap with typed Rust results. - selemis-com/clap_schema

github.com
4
1
10
1,377
Update: I’ve archived clap_schema. Within days, I hit two limits: I was iterating much faster than upstream could support, and Clap’s reflection API was not intended as a stable long-term foundation. That work now continues in Argx: github.com/selemis-com/argx

GitHub - selemis-com/argx: Expressive command-line parsing and configuration for Rust.

Expressive command-line parsing and configuration for Rust. - selemis-com/argx

github.com
1
96
See: xitter.cf/zerosnacks/status/2094…
zerosnacks
@zerosnacksAug 31
What if CLIs were actually optimized for agent use? Argx is an expressive, derive-first CLI and configuration framework for Rust, designed for humans and agents. Instead of parsing --help, agents get commands, arguments, results, errors, and schemas as structured data.
Rust example showing Argx deriving a typed CLI command with input, output and error types, and associating them with a command handler.

ALT Rust example showing Argx deriving a typed CLI command with input, output and error types, and associating them with a command handler.

117
What if CLIs were actually optimized for agent use? Argx is an expressive, derive-first CLI and configuration framework for Rust, designed for humans and agents. Instead of parsing --help, agents get commands, arguments, results, errors, and schemas as structured data.
Rust example showing Argx deriving a typed CLI command with input, output and error types, and associating them with a command handler.

ALT Rust example showing Argx deriving a typed CLI command with input, output and error types, and associating them with a command handler.

1
3
1
20
1,242
more replies
Argx represents that interface as JSON Schema, covering invocation, results, and errors. Agents already understand the format extremely well, making the CLI immediately familiar and straightforward to consume.
JSON Schema generated by Argx for a get command, describing its object ID input, GetOutput result and NotFound error.

ALT JSON Schema generated by Argx for a get command, describing its object ID input, GetOutput result and NotFound error.

1
1
2
130
Argx also includes layered configuration. When both the CLI and config framework read environment variables, precedence gets messy. Argx keeps defaults, config files, dotenv files, environment variables, and CLI args in one ordered, typed stack.
Rust example showing Argx layered configuration. A typed `Config` is resolved in order from defaults, global and local TOML files, a dotenv file, environment variables, and CLI arguments, with later layers taking precedence.

ALT Rust example showing Argx layered configuration. A typed `Config` is resolved in order from defaults, global and local TOML files, a dotenv file, environment variables, and CLI arguments, with later layers taking precedence.

1
3
109
v0.2.0 is out: github.com/selemis-com/clap_…

Release v0.2.0 · selemis-com/clap_schema

clap_schema 0.2.0 is a breaking redesign of its machine-readable CLI contract. The 0.1 model exposed useful Clap metadata, but still mixed parser representation, rendered-help concepts, and machine...

github.com
1
1
43
Load more