Mythra was created out of the question "how can solving problems be better than searching on a browser?" I entered the competition AI + UX Reflecting on 2023, Anticipating 2024 hosted by StratMinds and Design Buddies where we explored this prior question. I designed the concept, and my developer built the research preview, and we both crafted the white paper under the demo.
This is the next logical step to bridge this gap
What about ChatGPT plugins?
Even with plugins installed the best ChatGPT can do is generate an image or text of the output from the plugin, which isn't a compelling experience.
Modern internet communication is done through HTTP requests.
As LLMs grow more sophisticated, it is important for the user to autonomously output an interface.
Ultimately, Mythra offers more intuitive and dynamic solution to the public in solving problems than existing LLMs.
interface Protocol {
description: string,
endpoints: {
[key: string]: {
description: string,
action: 'GET' | 'POST' | 'PUT' | 'DELETE',
params: {
[key: string]: {
description: string,
type: 'string' | 'number' | 'boolean'
}
},
request: {
headers: string[],
description: string
body: {
[key: string]: string
}
},
response: {
headers: {
[key: string]: string
},
description: string;
body: {
[key: string]: string
}
}
}
},
baseUrl: string,
}
We believe the key to low cross-entropy loss results is for developers to provide as thorough and accurate a description of each endpoint and its parameters as possible.
By ingesting these descriptions along with the HTTP interface, an LLM can be made to learn how to communicate with it purely through HTTP.
Granted the honor of winning the challenge and will be competing in a stealth AI hackathon for 2024, where we will be further polishing this idea into viable world wide web integration.
For next steps, we plan to consider features such as suggestions, working on a blueprint for an AI protocol or mapping out a protocol beyond HTTP. The ideal milestone would be taking this to W3C.
Have a project or opportunity in mind?
Mythra was created out of the question "how can solving problems be better than searching on a browser?" I entered the competition AI + UX Reflecting on 2023, Anticipating 2024 hosted by StratMinds and Design Buddies where we explored this prior question. I designed the concept, and my developer built the research preview, and we both crafted the white paper under the demo.
This is the next logical step to bridge this gap
What about ChatGPT plugins?
Even with plugins installed the best ChatGPT can do is generate an image or text of the output from the plugin, which isn't a compelling experience.
Modern internet communication is done through HTTP requests.
As LLMs grow more sophisticated, it is important for the user to autonomously output an interface.
Ultimately, Mythra offers more intuitive and dynamic solution to the public in solving problems than existing LLMs.
interface Protocol {
description: string,
endpoints: {
[key: string]: {
description: string,
action: 'GET' | 'POST' | 'PUT' | 'DELETE',
params: {
[key: string]: {
description: string,
type: 'string' | 'number' | 'boolean'
}
},
request: {
headers: string[],
description: string
body: {
[key: string]: string
}
},
response: {
headers: {
[key: string]: string
},
description: string;
body: {
[key: string]: string
}
}
}
},
baseUrl: string,
}We believe the key to low cross-entropy loss results is for developers to provide as thorough and accurate a description of each endpoint and its parameters as possible.
By ingesting these descriptions along with the HTTP interface, an LLM can be made to learn how to communicate with it purely through HTTP.
Granted the honor of winning the challenge and will be competing in a stealth AI hackathon for 2024, where we will be further polishing this idea into viable world wide web integration.
For next steps, we plan to consider features such as suggestions, working on a blueprint for an AI protocol or mapping out a protocol beyond HTTP. The ideal milestone would be taking this to W3C.
Have a project or opportunity in mind?
Mythra was created out of the question "how can solving problems be better than searching on a browser?" I entered the competition AI + UX Reflecting on 2023, Anticipating 2024 hosted by StratMinds and Design Buddies where we explored this prior question. I designed the concept, and my developer built the research preview, and we both crafted the white paper under the demo.
This is the next logical step to bridge this gap
What about ChatGPT plugins?
Even with plugins installed the best ChatGPT can do is generate an image or text of the output from the plugin, which isn't a compelling experience.
Modern internet communication is done through HTTP requests.
As LLMs grow more sophisticated, it is important for the user to autonomously output an interface.
Ultimately, Mythra offers more intuitive and dynamic solution to the public in solving problems than existing LLMs.
interface Protocol {
description: string,
endpoints: {
[key: string]: {
description: string,
action: 'GET' | 'POST' | 'PUT' | 'DELETE',
params: {
[key: string]: {
description: string,
type: 'string' | 'number' | 'boolean'
}
},
request: {
headers: string[],
description: string
body: {
[key: string]: string
}
},
response: {
headers: {
[key: string]: string
},
description: string;
body: {
[key: string]: string
}
}
}
},
baseUrl: string,
}We believe the key to low cross-entropy loss results is for developers to provide as thorough and accurate a description of each endpoint and its parameters as possible.
By ingesting these descriptions along with the HTTP interface, an LLM can be made to learn how to communicate with it purely through HTTP.
Granted the honor of winning the challenge and will be competing in a stealth AI hackathon for 2024, where we will be further polishing this idea into viable world wide web integration.
For next steps, we plan to consider features such as suggestions, working on a blueprint for an AI protocol or mapping out a protocol beyond HTTP. The ideal milestone would be taking this to W3C.
Have a project or opportunity in mind?