(I was interested in this because I was actually working on something similar recently: https://github.com/imbue-ai/latchkey. To avoid the certificates issue, this library uses a gateway approach instead of a proxy, i.e. clients call endpoints like "http(s)://gateway.url:port/gateway/https://api.github.com/..." which can be effectively hidden behind the "latchkey curl" invocation.)

Comments URL: https://news.ycombinator.com/item?id=47184538

Points: 9

# Comments: 1

With that said, a specialized tool like this will almost certainly work better if Slack is the only service you want your agents to interact with. I like that the auth is transparent.

At Imbue, we’ve been following the rapid developments in the agent landscape and noticed that the way agents interact with third-party services on users’ behalf often leaves a lot to be desired. Integrations are ad hoc, complicated, context-heavy, and either unfriendly to non-technical users or tied to a lock-in of some sort.

We‘re experimenting with a command-line tool, Latchkey, that could be used by local agents targeted at non-technical users while avoiding remote intermediaries. As far as we know, this is the only existing approach at the intersection of these two goals.

Core idea: Agents access APIs of third-party services by prepending ordinary `curl` calls with the `latchkey` command, like this:

    latchkey curl -X POST 'https://slack.com/api/conversations.create' \
      -H 'Content-Type: application/json' \
      -d '{"name":"something-urgent"}'

Benefits:

* A single skill to integrate with all services.

* Direct communication between the agent and the third-party service (no OAuth intermediary app needed).

* Agents usable by non-technical users.

* Secrets do not leak to logs or chat transcripts.

We believe this aligns with a vision of a decentralized future in which people don’t need to ask corporations for permission to use their own data. We imagine a lively ecosystem of local agents that people use freely, supported by a community helping each other keep these tools useful and functional.

We’re aware that this approach comes with some downsides, too, and would love your feedback.

P.S. Here’s also a link to Passepartout, a toy demo AI assistant app built using Latchkey: https://github.com/imbue-ai/passepartout

Comments URL: https://news.ycombinator.com/item?id=46876636

Points: 12

# Comments: 0

  - minimal (no chat, no history, no replies, not even an input field)
  - universal (available in all your terminal applications)
  - transparent (the only context sent to an LLM is what you see on the screen)

Clearly, it's quite constrained in what you can get out of the LLM but sometimes constraints can lead to better focus.

I'm still not sure if / how useful it is but I find myself using it from time to time.

Comments URL: https://news.ycombinator.com/item?id=46464327

Points: 1

# Comments: 0

I still don't see how exactly. E.g., even if it "modified itself in response to this genetic engineering" then it wouldn't really transfer the lethal gene into another species, would it?

> Your use of absolutes here (...)

Which absolutes exactly are you talking about? :) Nothing absolute in the "almost certainly" statement, in my opinion.

> I am just surprised how little is known before a public trial was executed.

What do you mean by "is known"?

Secondly, these genes are pretty much self-regulating; they disappear from the population in one generation. They almost certainly won't be in the ecosystem long enough in order to play a role in any improbable scenario.

Take physics, for example. It can tell you why some objects behave the way they do by telling you there are certain particles, interacting forces, etc.. In this way you can explain, say, the photoelectric effect.

But it isn't really an answer to the "why" question, is it? It just pushes the question one level lower. Why are there such and such particles and forces? Why the constants? The very nature of these answers is descriptive. It is a description of how the world works, not why it works that way.

Maybe asking "why" in this ultimate manner is an ill-posed question - but that's not the deal here. It just doesn't seem that science in its current form, unlike religion or philosophy, could ever even attempt to answer it.

Don't get me wrong, I'm strongly atheistic myself, but there are some inherent limitations of scientific exploration and clarification with respect to the answers it can provide.

Just as a matter of interest - the current Czech prime minister, who, at least in theory, has more power than the president, has graduated in physics, and has actually worked in research for some time. Angela Merkel (coming from Eastern Germany) has also a scientific background (in physics and chemistry).

But otherwise, it seems you are right about the fact that there is no such rule about post-totalitarian leaders being educated in engineering or sciences in general.

I almost didn't manage to read further. The same can be said about so many things (if everyone just took or borrowed what he needed in a responsible manner, we wouldn't need money at all... to name the most utopian example), yet this kind of reasoning simply doesn't work, as has been shown both practically and theoretically (from game theory to "selfish gene").

It makes me always kind of sad to see smart and enthusiastic people not to grasp this simple fact.