Thanks for flagging this! You're absolutely right about network security.
The spec does require TLS - from Section 4.1:
> "All HTTP endpoints MUST use TLS 1.2 or higher"
> "Clients MUST verify TLS certificates"
The `http://localhost:8080` example is just for local development/demo purposes.
In production, all discovery endpoints and AgentCard fetches must use HTTPS.
> What exactly is the difference between lms and llmsterm?
With lms, LM Studio's frontend GUI/desktop application and its backend LLM API server (for OpenAI compatibility API endpoints) are tightly coupled: stopping LM Studio's GUI/desktop application will trigger stopping of LM Studio's backend LLM API server.
With llmsterm, they've been decoupled now; it (llmsterm) enables one, as LM Studio announcement says, to "deploy on servers, deploy in CI, deploy anywhere" (where having a GUI/desktop application doesn't make sense).
By “based on” I don’t mean a shared codebase or features but rather Parker and I exchanged emails a decade ago to discuss business models and open source funding. He initially copied my Sidekiq OSS + Sidekiq Pro business model, with my blessing.
This is absolutely true (except we went OSS + Web initially, Pro came later). You were an inspiration, always helpful in discussion, and definitely paved the way for this business model.
Anybody else notice that half the video was just finding papers to decorate the bibliography with? Not like "find me more papers I should read and consider", but "find papers that are relevant that I should cite--okay, just add those".
Yes. That part of the video was straight-up "here's how to automate academic fraud". Those papers could just as easily negate one of your assumptions. What even is research if it's not using cited works?
"I know nothing but had an idea and did some work. I have no clue whether this question has been explored or settled one way or another. But here's my new paper claiming to be an incremental improvement on... whatever the previous state of understanding was. I wouldn't know, I haven't read up on it yet. Too many papers to write."
It's as if it's marketed to the students who have been using ChatGPT for the last few years to pass courses and now need to throw together a bachelor's thesis. Bibliography and proper citation requirements are a pain.
That is such a bummer. At the time, it was annoying and I groused and grumbled about it; but in hindsight my reviewers pointed me toward some good articles, and I am better for having read them.
I've noticed this pattern, and it really drives me nuts. You should really be doing a comprehensive literature review before starting any sort of review or research paper.
We removed the authorship of a a former co-author on a paper I'm on because his workflow was essentially this--with AI generated text--and a not-insignificant amount of straight-up plagiarism.
There is definitely a difference between how senior researchers and students go about making publications. To students, they get told basically what topic they should write a paper on or prepare data for, so they work backwards: try to write the paper (possibly some researching information to write the paper), then add references because they know they have to. For the actual researchers, it would be a complete waste of time/funding to start a project on a question that has already been answered before (and something that the grant reviewers are going to know has already been explored before), so in order to not waste their own time, they have to do what you said and actually conduct a comprehensive literature review before even starting the work.
Plus, this practice (just inserting AI-proposed citations/sources) is what has recently been the front-runner of some very embarrassing "editing" mistakes, notably in reports from public institutions. Now OpenAI lets us do pageantry even faster! <3
It's all performance over practice at this point. Look to the current US administration as the barometer by which many are measuring their public perceptions
The hand-drawn diagram to LaTeX is a little embarrassing. If you load up Prism and create your first blank project you can see the image. It looks like it's actually a LaTeX rendering of a diagram rendered with a hand-dawn style and then overlayed on a very clean image of a napkin. So you've proven that you can go from a rasterized LaTeX diagram back to equivalent LaTeX code. Interesting but probably will not hold up when it meets real world use cases.
A more apt example would have been to show finding a particular paper you want to cite, but you don’t want to be bothered searching your reference manager or Google Scholar.
E.g. “cite that paper from John Doe on lorem ipsum, but make sure it’s the 2022 update article that I cited in one of my other recent articles, not the original article”
You may notice that this is the way writing papers works in undergraduate courses. It's just another in a long line of examples of MBA tech bros gleaning an extremely surface-level understanding of a topic, then decided they're experts.
Authentic relating, every time I see that it comes up, continues to be a red flag warning to expect antisocial behavior by bad actors. This article does not break that trend.
I'm sure some people get utility out of it and improve their lives, but there seem to be quite a few people who seem to enjoy it for other reasons.
> In particular, the Valley incarnates the ability to “do things,” to iculate and achieve long-range projects, which has all but disappeared from other sectors of the economy and most glaringly from politics.
The author then goes on to observe that these folks are all right-leaning or working with the Republicans.
The reason for this--and the beatings will continue until the progressives, liberals, and Democrats figure this out--is that the right is not fundamentally opposed to economic and technological progress.
If you keep rigging the game, you lose good players, and they'll stop playing for you--and the fans will stop coming.
> Because no one can do everything. Telling devs to own their code is one thing. (Great.) Asking them to own their code and the entire technological iceberg beneath it is wholly another
A lot of people can do enough of everything as to make no practical difference--don't hire monkeys (and if you do, train them) and empower the people you do hire and you'll be amazed at how much they can do.
As for the rest--maybe the solution is not to build big icebergs!
Please, please, please require HTTPS. Dear god.
The network is not secure.
Your example:
```
# Start a discovery server
python -m server.lad_server --name "My Agent" --port 8080
# Discover agents (in another terminal)
python -m client.lad_client --url http://localhost:8080
```
reply