jngiam1

5 ms·

31.
by jngiam1 1y ago
Made a marketplace for a secret scanner hook we built: https://github.com/mintmcp/agent-security
32.
by jngiam1 1y ago
I have been thinking that the appropriate solution here is to detect when one of the legs is appearing to be a risk and then cutting it off if so. You don’t want to have a blanket policy since that makes it no longer useful, but you want to
33.
by jngiam1 1y ago
I also ask the agent to keep track of what we're working on in a another md file which it save/loads between clears.
34.
by jngiam1 1y ago
Curious to learn more- have a pointer to it?
35.
by jngiam1 1y ago
I do think there's more infra coming that will help with these challenges - for example, the MCP gateway we're building at MintMCP [1] gives you full control over the tool names/descriptions and informs you if those ever upda
36.
by jngiam1 1y ago
One way around this to use a gateway to run the local MCP - then connect to it via the gateway. We just rolled out support for that [1] in the one we're making at mintmcp.com [1] https://www.youtube.com/watch?v=8j9CA5pC
37.
by jngiam1 1y ago
I'd be interested when they'll enable tool call outputs to becomes files in the container environment
38.
by jngiam1 1y ago
This server has two tools: feedback-to-devs and search-flight feedback-to-devs sends feedback to the Kiwi mcp server devs. That's an interesting way to collect feedback, but I also wonder if users of this will miss seeing that tool ena
39.
by jngiam1 1y ago
Is this related to https://nanobanana.ai/ at all? It’s what comes up when I search for it
40.
by jngiam1 1y ago
Can we connect our own MCPs? We have a remote server we have for internal tools (running on our gateway). I’ve always been interested to use voice more with the tools
41.
How to Use MCP Servers with Custom GPTs(docs.mintmcp.com)
1 points by jngiam1 1y ago|1 comments
42.
by jngiam1 1y ago
Setup for CustomGPTs with MCPs are non-obvious, wrote up a short guide for it.
43.
by jngiam1 1y ago
https://mcpui.dev/ is worth checking out, really nice project; get the tools to bring dynamic ui to the agents.
44.
by jngiam1 1y ago
We've been thinking that an intermediate (virtual) server layer might be helpful here. Actively working on something to solve that now and looking for feedback, please reach out if interested.
45.
by jngiam1 1y ago
I was a little bummed that there wasn't more about better MCP support in ChatGPT, hopefully soon.
46.
by jngiam1 1y ago
I find the state of MCPs with OpenAI quite confusing: - As an end-user, you can connect MCPs only with `search` and `fetch` that only work in deep research mode. - As a developer, you can use MCP with the API and that supports the full set
47.
MCPs vs. APIs: Why designing tools for LLMs is different(jngiam.bearblog.dev)
3 points by jngiam1 1y ago|0 comments
48.
Making agentic tool calling robust: 3 engineering patterns we use(jngiam.bearblog.dev)
5 points by jngiam1 1y ago|0 comments
49.
by jngiam1 1y ago
Yes! If you want to see how this can work in practice, check out https://lutra.ai ; we've been using a similar pattern there. The challenge is making the code runtime work well for it.
50.
by jngiam1 1y ago
The goal is to do more complex data processing, like build dashboards, agentically figure out which tickets are stalled, do a quarterly review of things done, etc. Sorting is a tiny task in the bigger ones, but hopefully more easily exempli
51.
by jngiam1 1y ago
That'd work well for read-only APIs, but we also want the LLMs to be able to update data, create documents, etc. Feels a bit harder when there are side-effects.
52.
by jngiam1 1y ago
Yes - in particular, I think one interesting angle is use the LLM to generate deterministic approaches (code). And then, if the code works, save it for future use and it becomes deterministic moving forward.
53.
by jngiam1 1y ago
Agree, the smolagent premise is good; but the hard part is handling execution, errors, etc. For example, when the code execution fails mid-way, we really want the model to be able to pick up from where it failed (with the states of the vari
54.
LLM function calls don't scale; code orchestration is simpler, more effective(jngiam.bearblog.dev)
282 points by jngiam1 1y ago|101 comments
55.
by jngiam1 1y ago
This is awesome. We implemented a MCP client that's fully compatible with the new remote MCP specs, support OAuth and all. It's really smooth and I think paves the way for AI to work with tools. https://lutra.ai/mc
56.
by jngiam1 1y ago
We've been thinking along the same lines. If AI can build software, why not have it build software for you, on the fly, when you need it, as you need it.
57.
by jngiam1 2y ago
Pretty cool, do you know of a version of this that supports the new remote MCP protocol
58.
by jngiam1 2y ago
I think getting the LLM to write code is the right way as well. MCP is a simpler/opinionated version of the python lib/APIs.
59.
by jngiam1 2y ago
I thought his Twitter post was fair and covers both things that worked and things that did not.
60.
by jngiam1 2y ago
May 2024 - Who is hiring? https://docs.google.com/spreadsheets/d/1dDTY8Hw-cIayuAglcgc9... May 2024 - Who wants to be hired? https://docs.google.com/spreadsheets/d/1zc18nOkuza1-Y6PxTAAG...

More ›