tejassuds

8 ms·

1.
by tejassuds 2mo ago
There is so much weight in this one comment. I feel it now!
2.
by tejassuds 2mo ago
unless you can't, then wait, are u agi?
3.
by tejassuds 2mo ago
fr haha except you can forgive maybe? that's the DLC the model never shipped with
4.
by tejassuds 2mo ago
Honestly you've put your finger on the sharper version of it. Inside the context window it isn't forgetting, it's near total recall right up to saturation, and then the rockface. Which is why "catastrophic" fits and
5.
AI doesn't know how to forgive and cannot forget(tejassuds.com)
21 points by tejassuds 2mo ago|24 comments
6.
by tejassuds 4mo ago
We improved the runtime and made it better! Come check it!
7.
Show HN: Platos – like Claude Managed Agents but open-source and self-hosted(github.com)
10 points by tejassuds 4mo ago|1 comments
8.
Show HN: Neural DNA – 258 params that grow network topology(github.com)
1 points by tejassuds 6mo ago|0 comments
9.
by tejassuds 10mo ago
Ideally, you could. But that would mean fundamentally changing a lot in the framework you are porting it to. Maybe we could create a package that can be imported to other frameworks to enable skillcalling.. something to think about !
10.
by tejassuds 10mo ago
Yes! Largely inspired by Anthropic's work on code execution with MCP! The idea of more sophisticated, autonomous units instead of raw tool functions really clicked for me. Skills in SkillGraph can orchestrate multiple tools, manage sta
11.
Show HN: SkillGraph – Open-source agentic framework with skills instead of tools(github.com)
19 points by tejassuds 10mo ago|4 comments