HackerTrans
TopNewTrendsCommentsPastAskShowJobs

sakuraiben

no profile record

Submissions

Show HN: TakoVM – open-source sandboxing for your agent's code

github.com
6 points·by sakuraiben·11 dagen geleden·0 comments

Show HN: TakoQA – A harness to get a swarm of agents to break your application

github.com
2 points·by sakuraiben·17 dagen geleden·0 comments

Show HN: Reachpad – open-source .md sharing platform for companies and agents

github.com
3 points·by sakuraiben·18 dagen geleden·0 comments

Show HN: Reachpad – document and knowledge sharing for your agents

reachpad.dev
3 points·by sakuraiben·21 dagen geleden·0 comments

Show HN: Memharness – Bi-temporal memory for AI agents, in one SQLite file

github.com
1 points·by sakuraiben·24 dagen geleden·1 comments

Show HN: Guarden – Authorization for AI agent actions powered by OPA

github.com
1 points·by sakuraiben·vorige maand·0 comments

Show HN: TakoVM – Isolated model and tool execution used by enterprises

github.com
23 points·by sakuraiben·vorige maand·6 comments

Show HN: Intencion – Product analytics that improves your AI agents continuously

intencion.io
9 points·by sakuraiben·vorige maand·2 comments

Show HN: Weight-inspect, examine popular ML models formats without the weights

github.com
1 points·by sakuraiben·5 maanden geleden·0 comments

comments

sakuraiben
·11 dagen geleden·discuss
Take a look at https://github.com/Tako-Research/TakoVM!
sakuraiben
·11 dagen geleden·discuss
so cool! big fan of barley tea growing up with it
sakuraiben
·24 dagen geleden·discuss
Been using this everyday to dog food myself, would love to hear what everyone thinks
sakuraiben
·vorige maand·discuss
Downstream ;)
sakuraiben
·vorige maand·discuss
Thanks for the comment - and to answer your question: AI use dynamic workflows which require dynamic code generation at runtime. In most critical deployment scenarios you need to isolate the code execution and network controls ahead of time. That's the main use case I had in mind while building this. Hope it helps!
sakuraiben
·vorige maand·discuss
Probably when it was able to reverse engineer a bunch of firmware from scratch and it was unable to explain how and why it was able to reason through it.
sakuraiben
·vorige maand·discuss
One thing I've noticed using agents for coding is that they really like to write thousands of unit tests but not dynamically test.
sakuraiben
·vorige maand·discuss
Great name choice
sakuraiben
·vorige maand·discuss
Would be interested to see what kind of eval results you get from this
sakuraiben
·vorige maand·discuss
would love to see more lines on my face