zackify

You didn’t even mention how it’ll be on fire unless you use low power mode.

Yes all this has been known since the M4 came out. The memory bandwidth is too low.

Try using it with real tasks like cline or opencode and the context length is too long and slow to be practical

I know. I cannot believe lm studio. Ollama. Especially model providers, do not offer this yet.

The memory bandwidth is crap and you’ll never run anything close to Claude on that unfortunately. They should have shipped something 8x faster at least 2 tb/s bandwidth

It's such a pain when you want to do something that is easily possible but manifest v3 won't allow it... like having an internal extension that references remote content...

Almost all of this was true before they even announced the purchase. I was so shocked and now I’m not surprised it fell through

Everything is so dependent on the type of work you do. For me sonnet works pretty well, it’s hard for me to say. Even when I used opus nonstop last month it’s not like I was doing the exact same tasks on sonnet. However it did drop down to sonnet a lot and I didn’t have a problem

Since downgrading from max to pro.... i have been using sonnet 4 a TON and i havent even been limited yet. The usage allowance is awesome since gemini cli released.

It’s an awesome book and I’ve been dying to see how they’ll adapt it to a movie soon

If it helps anyone else. I downgraded from Claude max to pro for $20 and the usage limits are really good.

I think they’re trying to compete with Gemini cli and now I’m glad I’m paying less

It is new and exciting if you just learned to vibe code, and you don’t even know what a rest api is

As of 2 weeks back if I did this, it would reset back the moment cline made an api call. But lm studio would work correctly. I’ll have to try again. Even confirmed cline was not overriding num context

Because waiting forever for initial prompt processing with realistic number of MCP tools enabled on a prompt is going to suck without the most bandwidth possible

And you are never going to sit around waiting for anything larger than the 96+gb of ram that the RTX pro has.

If you’re using it for background tasks and not coding it’s a different story

Literally the only reason I would use them is price. Cannot get anything for the prices they offer at a normal place or through insurance

Ollama doesn’t even have a way to customize the context size per model and persist it. LM studio does :)

I love LM studio but I’d never waste 12k like that. The memory bandwidth is too low trust me.

Get the RTX Pro 6000 for 8.5k with double the bandwidth. It will be way better

What’s neat is we can proxy requests from Gemini or fork it with only replacing the api call layer so it can be used with local models!!!

When I was using it in cursor recently, I found it would break imports in large python files. Claude never did this. Do you have any weird issues using Gemini? I’m excited to try the cli today

The problem with that is it doesn’t work for people who are not technical. Remote mcp is pretty good even if I would have preferred a rest api helper endpoint to support existing apis

I’ve managed to do the same thing!

It’s actually surprising just how powerful 1-5 tools can be if you document it well and the llm knows how to pass arguments from other tool responses you had higher up in the thread

Will this end up getting an open ai compatible web server or is that out of scope.

Seems like this is the exact same extension that’s been in use but it’s just now publicly on the extension marketplace.

Just call them “web forms” for LLMs.

You can ask the client to fill in a dropdown or input.

Example they give is a restaurants table booking tool.

Imagine saying book a table at 5pm. But 5pm is taken.

You can “elicit” the user and have them select from a list of available times

How would you compare to a general meditation app like “Waking Up” curious if you’ve used it.

It has been moving fast.

One weird thing I found a few weeks ago, when I added my remote MCP to Claude's integration tab on the website, I was getting OAuth errors.

Turns out they are requiring a special "claudeai" scope. Once I added that to my server, I was able to use it remotely in claude desktop!

I couldn't find any docs or reasons online for them requesting this scope.

Also, I have been using remote mcps in claude code for weeks with the awesome mcp-remote proxy tool. It's nice to not need that any longer!

Then as I'm writing a book currently on MCP Servers with OAuth, Elicitations come out! I'm rushing to update this book and be the best source for every part of the latest spec, as I can already see lots of gaps in documentation on all these things.

Huge shout out to VS Code for being the best MCP Client, they have support for Elicitations in Insiders already and it works great from my testing.

Oh wow! This reminds me so much of what I was trying to do almost 10 years ago but I was too much of a noob to make it make more sense haha.

https://github.com/zackify/legible

I thought this simple setup is a great idea for internal MCP servers.

Essentially you can use stdio, without the hassle of a shared internal http server, and still have it auto update with changes your team makes.

I tried this idea with the new MCP prompt support in VS code and I’m pretty happy with it so far.

Yeah unfortunately doesn’t exist yet and it’s an open issue

I’m surprised this post is able to stay up on hackernews.

One time I simply made a post about how a page was simpler to code and understand in Remix.

You may not believe this but I had 3+ chat requests from multiple employees at vercel.

Across multiple social media accounts. All asking to take it down or meet and discuss with them.

It definitely works in bun just fine. I have a production built mcp server with auth running under bun.

Now if you convert the request / response types to native bun server, it can be finicky.

But it works fine using express under bun with the official protocol implementation for typescript.

Actually writing a book about this too and will be using bun for it https://leanpub.com/creatingmcpserverswithoauth

I came here to say the same thing.

Rather ship oven/bun through docker and have a 90mb container vs using node.