Hacker Newsnew | past | comments | ask | show | jobs | submit | tb1989's commentslogin

I prefer this interpretation:

MCP Isn’t the USB-C of AI — It’s Just a USB-C(laude) Dongle

https://dev.to/internationale/mcp-is-not-ai-usb-c-its-usb-cl...

The illustrations are on their official website.

I find it hard to respect a company that invented dongle and claims to have invented USB


> MCP, despite its open-source claim, heavily relies on proprietary interfaces from Anthropic's Claude model. Thus, the proclaimed openness is surface-level at best, creating subtle but significant barriers within the ecosystem.

This strikes me as fundamentally untrue, but I'd love to see an example of this in the source code or documentation.


I think this article explains it all:

MCP Isn’t the USB-C of AI — It’s Just a USB-Claude Dongle

https://dev.to/internationale/mcp-is-not-ai-usb-c-its-usb-cl...

Nonetheless, I think your work is very good and it looks like a very useful dongle


I'm not even sure comparing it to USB is helpful, since USB is just a transport protocol, not communication protocol. Just because two devices have usb ports doesn't mean they can be connected together, otherwise we wouldn't need device drivers. So in this instance it's more of a unified device driver, rather than a port.


I totally agree. As someone with an EE background, this metaphor makes me a little physically uncomfortable. Considering that the developers and users of mcp are almost all programmers who need to use cli, I really don't understand why they don't tell the truth to programmers.


That article makes very little sense.

The protocol itself sits on top of JSON-RPC, and the specifications are there for anyone to implement. There's nothing specific to Claude about it.

There are various MCP client and server implementations available that are also unrelated to Anthropic.


- The official example strongly promotes the Anthropic API, which is on GitHub. This is clear evidence.

- There is no clear explanation of the coupling between the system prompt and the tool call. Even if it mentions the open source Gemma or Deepseek, it would be much better.

The official attitude makes it difficult to trust this project.

The point you made is exactly the cunning part. Anyone can copy it, but without official support, it is simply impossible: this is pure community exploitation


If you want an LLM to use a tool, you just need to implement a parser in your LLM client that extracts the tool call from the LLM's response, then give the LLM a syntax it can use to make the tool call.

For example, in Roo Code:

``` TOOL USE

You have access to a set of tools that are executed upon the user's approval. You can use one tool per message, and will receive the result of that tool use in the user's response. You use tools step-by-step to accomplish a given task, with each tool use informed by the result of the previous tool use.

# Tool Use Formatting

Tool use is formatted using XML-style tags. The tool name is enclosed in opening and closing tags, and each parameter is similarly enclosed within its own set of tags. Here's the structure:

<tool_name> <parameter1_name>value1</parameter1_name> <parameter2_name>value2</parameter2_name> ... </tool_name>

For example:

<read_file> <path>src/main.js</path> </read_file>

Always adhere to this format for the tool use to ensure proper parsing and execution. ```


Frankly speaking, labeling MCP as "USB-C for AI" is an exaggerated metaphor. USB-C is a universal standard driven by IEEE and the open-source community, while MCP is fundamentally an ecosystem tightly controlled by Anthropic itself. Despite adopting the MIT license, MCP is far from the spirit of genuine openness. For instance, the official examples exclusively utilize Anthropic’s own API, and the system prompt of LLM requests is hidden within their SDK—precisely the part that's crucial for openness.

Anthropic has historically been cautious, even conservative, regarding open source. They refuse to open-source Claude's CLI and the underlying model code, which makes their sudden enthusiasm for launching MCP, presented as an "open platform," rather puzzling. Is this openness genuinely for the community, or merely an attempt to benefit their own ecosystem and interests?

In reality, MCP is more like a clunky dongle that awkwardly connects HDMI or TF card readers specifically to their own models, rather than establishing a genuinely universal standard. It's nearly impossible for MCP to change the conventional methods of using APIs and authentication into an "AI-native" approach. Due to security considerations, this design inherently caps the potential of both agents and MCP itself.

For developers, MCP imposes high environmental prerequisites, like Python, Node, or Docker, virtually excluding non-technical users. The installation process is the opposite of elegant.

In the long run, this superficially open yet fundamentally closed approach is unlikely to win sustained community trust or long-term success. Security concerns, the allocation of community benefits, and the inherently closed mindset of the corporation are critical issues that will inevitably haunt MCP’s future development.

Calling MCP "USB-C for AI" feels more like marketing hype than technological reality.


I feel that calling it the AI tooling of the future is an exaggeration:

https://dev.to/internationale/the-mcp-trap-is-the-open-sourc...


Bro, this way of thinking of presuming others guilty is very primitive...Disclosure is very important! If you want to sell products overseas outside of mainland China, you must know this.


Apologies, I did overreact.


this 'grifter' is on TE's board, so things are out of control


Tell me one thing, not all (100%) AI companies are scammers, but all Internet scammers will definitely say that they are engaged in AI.


Doesn't surprise me. In my opinion, many companies try to achieve outrageous premiums by taking the route of teen engineering (the originator of hype). In addition to rabbit, there is also a series of nothing products.

This doesn't convince the tech fetishists. In fact, I think te's contribution to music is very limited, even harmful, especially when I see the latest Yamaha even imitating te, which feels a bit funny and ridiculous. We need real innovation and democratic pricing.

By the way, if you care, you can learn about the history of rabbit’s founder. Let's just say, in certain circles, this is a recognized liar. So I’m not surprised at all when it was said a few days ago that Rabbit stole everyone’s passwords.


Twitter thread on rabbit's founder and his previous involvement with NFTs that he scrubbed:

https://twitter.com/Andyparackal/status/1785676408280498655


Pikachu's everywhere were shocked.

This is just the same guy, over and over. If he isn't too old, I bet he was into dropshipping too. Always on whatever new grift, always ready to start selling shovels to gold miners. I wish people were smarter about this stuff.


building a hardware company and shipping thousands of devices doesn't seem like a get-rich-quick grift to me. There are a lot of things rabbit can do that you couldn't do on an app since they own the hardware. Just shipped it too soon.

The product is definitely not super appealing to me yet in the current state.


> There are a lot of things rabbit can do that you couldn't do on an app since they own the hardware. Just shipped it too soon.

I vehemently disagree with both of these statements.

For one, as far as I can tell every single hardware feature on the Rabbit (sans rotary encoder) has been present on every phone I've owned since 2011. Forward and rear-facing camera - check, nice bright touchscreen - check, speaker and mic - check... it's all there. Rabbit's "ecosystem" excuse is just as hollow as when literally any other company does it. They're trying to project a halo-effect, and too many people fell for it. Are there any software features that are exclusively enabled by Rabbit's form factor?

For two, waiting to ship it probably wouldn't have solved it either. Humane got absolutely humiliated last week with their own offering, it was a now-or-never opportunity to stage them up or let them set the tone forever. I don't think interest in AI or AI-focused hardware is liable to skyrocket any more than cryptocurrency-focused hardware was. Both of these things are software solutions; using it to sell you physical hardware is a 100% pure marketing gimmeck.


I will slightly disagree with your first point. I also partially agree though. I agree that every smart phone has most of the hardware of the Rabbit R1. But, just because the hardware is there, that doesn't mean the hardware manufacturer provides APIs/callbacks/hooks to third party developers for every capability of that hardware and operating system. The software developer has to build within the confines of the OEM. Building your own hardware, or at least building your own firmware and OS for open hardware, lets you do more with that same hardware.

Now, if we're talking about the OEMs building Rabbit-like features into their phones then that I agree with, and I hope (and assume) that's what will happen now. Rabbit has already shown some features that our smartphones should be able to do quite easily, such as sending a selected photo by text or email. My hope is that Rabbit pushes Apple and Google to build some of the low hanging fruit into their phones/operating systems soon.


I agree with you on hardware but I'm not talking about hardware. I'm talking about UX. There is a huge difference between being iOS and being an app on iOS as far as access to the user of the device. There is no way to innovate on novel notification mechanisms, for example.


I think at the very least it's a kind of trend-chasing. If the AI hype dies down a bit will Rabbit AI continue supporting their product and customers or are they going to wipe their online presence and move on to the next trend like with the NFT thing?


I don't even think in this case it's a matter of the AI hype dying down (though it is, and will continue to do so as products like these fail to get actually useful) I think this is just a complete nonstarter. Like, there's just nothing here that isn't better executed with existing tech. If I want to ask my phone to order a pizza, I can set that up in a lot of apps with Siri shortcuts, or, and far more likely, just open the damn app and pick what I want. Why the shit do I want to carry around a Rabbit, and make sure it has wifi access, and keep that also on my person next to my phone that does all that shit already?

It's literally just another smartphone, that you only control with your voice which is the worst way to control a smart phone, and it'll offer suggestions I guess? And I guess if you don't want the suggestions you need to have a whole fuckin conversation with it.

Like, we've already been down this road. It's tremendously easy as a customer to just be presented a list of things to buy, pick what you want, and swipe your credit card. Rabbit is a regression to ordering things with a phone, except instead of talking to a person, you're talking to a robot. But that's not an improvement, and in fact in many ways, it's a step back.

I just, I cannot fathom who wants this.


> Rabbit is a regression to ordering things with a phone, except instead of talking to a person, you're talking to a robot.

> I just, I cannot fathom who wants this.

If that doesn't make a difference to you, you're not the target market. It truely doesn't for some, but for others, it does!


Show me someone.



I'm a millennial and I hate talking on the phone, but I don't want to talk to a robot, either. Just give me the damn app with menus and stuff and I'll browse them at my convenience and then pick what I want.

And even if I did want to talk to a robot, I'd rather just use the gadget that I already have for talking into, not yet another thing to lug around, keep charged etc.

It really is completely pointless.


Well the unit economics don't make any sense. You get unlimited free LLM calls forever? I don't really see the argument for it shipping too soon (although it is too soon for what they're claiming). There's a basic unit economics problem that can't be solved regardless of their future roadmap. They're promising people a no subscription model, claiming it eventually will have a Large Action Model and LLM for question and answering, and somehow those will be available free forever? It's either a get-rich-quick scheme, which sounds likely given their background as crypto grifters, or complete incompetence (somehow no one there ever thought what it might cost to do inference). Neither seems like a legit company that just shipped too soon.


name one thing the rabbit can do that a smartphone can't.


While the rabbit, based on reviews I've seen, doesn't live up to its hype yet, AI assistants won't be able to do anything you can't do yourself on a smartphone or regular laptop.

Instead, the idea is they will be able to do these things for you, similar to a human assistant. Currently only rich people can afford assistants to manage their lives for them. This could be about to change.


But again: why do you want that? The pizza example is downright stupid, in no world would anyone prefer to tell an assistant, robot or otherwise, to order them a pizza. It's such a trivial task that outsourcing it to another entity, irrespective of what kind, is just less efficient, full stop.

And, even going for their big example: why would you want to outsource the planning of an overseas trip?! Planning a trip is the FUN part for goodness sakes! Choosing your destinations and building an itinerary. The only way I could see this is if you're so drowned in money and so utterly dead inside from years of casual overseas vacations that you just don't give a shit where you go anymore, and like, fair enough but a virtual assistant is not going to solve the yawning chasm of meaninglessness in your soul. You should probably see a therapist about that.

All of these things feel like faint imitations of something like JARVIS from the Iron Man movies, and look, if it worked like that as just a disembodied voice built into my home, and my cars, and my armored battle suit and my cell phones, hell yeah I'd have that. But the only way I would accept any technology that invasive is if I was running it on my own hardware, I will never, ever, in a thousand years give over that much hardware and privileged access to another pop-up company from silicon valley.


> You should probably see a therapist about that.

You're projecting so hard you might hurt yourself.

> in no world would anyone prefer to tell an assistant, robot or otherwise, to order them a pizza.

Sure they would, I'm one of them, I know many more.

> It's such a trivial task that outsourcing it to another entity, irrespective of what kind, is just less efficient, full stop.

Perhaps for you. As someone with a spinal cord injury whom at many times can not work a keyboard or a phone, a far simpler ability to get things done with my voice is something i've been looking forward too for some time and I hope continues to progress, especially as I age and things become more difficult.

> Planning a trip is the FUN part for goodness sakes!

What you enjoy and others enjoy is not the same. Your view of the world and what does and does not constitute a good time is not the benchmark or standard for anyone other than yourself. Again, you're projecting.

> I will never, ever, in a thousand years give over that much hardware and privileged access to another pop-up company from silicon valley.

How delightful for you.


I'd love for the iPod to come back. I have a small DAP, the Shanling M0, but it has not the iPod ergonomics. A phone can do what the DAP does, but not the way it does it. A better phone ergonomics would probably be a PDA.

(I'm not defending Rabbit R1 and I think it's useless).


I am not sure how active it is still but there was a pretty healthy ipod modding community a few years back if you are eager for those ergonomics with modern functionality.


not have a screen.

Sometimes the thing that defines something is not what it has, but what it lacks.


… or, you could do the same with a web app… and OAuth.

But a web app doesn’t get you the hype, VC funding and Galaxy brained Twitter tech bro adoration.


Care to elaborate, for the curious? A quick google on Jesse Lyu doesn't turn up much besides his own hype. Although he does appear to at least be obfuscating the truth, failing to admit that the only reason the "bootleg APK" isn't working is because of an IMEI whitelist.

Update: They've barely tightened up, now the only missing piece is the OS build fingerprint. https://twitter.com/uwukko/status/1785626783900930447/


I don't why it should surprise, or bother, anyone that it is running android. Totally reasonable choice. What does it say that it is treated like some sort of gotcha? Were they supposed to build their own AIOS?


because if it's just an android app, you have to wonder why it's not just an app


One reason was mentioned in the article:

After all, the Rabbit R1’s launcher app is intended to be preinstalled in the firmware and be granted several privileged, system-level permissions — only some of which we were able to grant — so some of the functions would likely fail if we tried

And the statement from Rabbit in the article says essentially the same:

rabbit OS and LAM run on the cloud with very bespoke AOSP and lower level firmware modifications, therefore a local bootleg APK without the proper OS and Cloud endpoints won’t be able to access our service


Are these privileged system-level permissions in the room with us now? What specifically are they?


why are you so incredulous that android might have some annoying privacy restrictions that a custom AOSP can sidestep? I would google rabbit's reasoning for this but I don't care enough


Their official reasoning: "rabbit OS and LAM run on the cloud with very bespoke AOSP and lower level firmware modifications". https://twitter.com/rabbit_hmi/status/1785498453097009473 This reads like obfuscation to me. Just tell us in plain english!


> why are you so incredulous that android might have some annoying privacy restrictions that a custom AOSP can sidestep?

I don't doubt that such restrictions exist. But I'm also curious as to what, specifically, they would be? Apps can access all sensors, cameras, microphone, network. So what's left?


You can CTRL-F here https://developer.android.com/reference/android/Manifest.per... for "signature" and view permissions that are only granted to apps signed with the platform key, i.e. built into the system image as part of the AOSP build process.

There's a good number that might be useful for the R1.


Because people were able to launch the app on their Android phones.


Because no one wants to pay for apps?


This one actually makes a lot of sense to me.

Just from a pure marketing perspective, look at all the insane hype they were able to generate due to possibility of a new physical form factor.

There’s no way in hell they could have generated $10M in pre-orders and potentially hundreds of millions in earned media coverage for the 700th app that’s basically a chatgpt api wrapper.


You’re absolutely right, it would “just be an app” and have a much harder time getting noticed.

But I really meant what I said too. It doesn’t matter how great your thing is, people complain to high health about spending $.99 on it. And God forbid you want to subscription even if it’s only $2 a year. And don’t ask for an IAP of any kind.

The app market is a disaster of ads because almost no one will accept anything else.

Even if everything ran fully on the device so they didn’t have any server costs of any kind I think they’d have a very very hard time making money off an app unless it was incredibly extraordinary.


> why it's not just an app

If it's just an app...their hardware has no reason to exist AND they are competing directly with Google/Microsoft.


I don’t get why people would pay for a bright orange box, or a snazzy little pin, when you can do the exact same thing with a phone app. Siri and Google’s offering on Android phones will eliminate this market when they start in on this stuff. I don’t see why any company making an AI widget would be valued high, unless they’re trying to get bought by Apple.


No argument from me. I don't see the end of smartphones any time soon. And if pulling a smartphone out of your pocket is too inconvenient, the smartwatch almost certainly has your use-case covered.


Because they want to develop their own idependent ecosystem? Similarly, why does Amazon Fire OS exists?

I mean the question is kind of vaguely legitimate, but regardeless of how it is implemented one could ask why it is not an Android app instead. They thought about it and made other choices, and might not want to publicly list all the reasons that made them choose what they did...


I thought that before I knew it was running android.


I don't like most of TE's output but the OP-1 is definitely innovative.


Apparently, TE knows this Rabbit person's background as we are talking about millions dollar business. TE's values here are questionable when they decide to go with this guy.

As for OP1, packaging a tape effect when it can't produce good sound quality is a fraud. Sampling from FM? Check Klaus Schulze.. Music startups can certainly innovate, such as Roli (Seaboard and Rise), moog. make noise. Open source monome. and, to my sorrow, Émilie Gillet (Mutable Instruments). TE's innovation in music itself is far inferior to them. If you praise te, I think it is unfair to other people who are obsessed with technological innovation.

I have always felt that TE is a electronics company with no innovation in hardcore technology. The appearance and packaging are different from previous ones, but there is not enough practical integration. Of course, this is just my personal opinion. If you disagree, you can write down your opinion. Thank you.


I would love to know about any other device that ticks the boxes the OP-1 does:

* Battery powered * Easily backpack portable

* Standalone DAW

* Piano layout

* Physical knobs and keys

* Sequencer

* Screen (and accompanying UI) large enough to not feel incredibly cramped

That's plenty of stuff that beats the OP-1 on plenty of these, but there just isn't much out there that hits all of them at once.


A Synthstrom Deluge arguably ticks those boxes, except perhaps for the "piano layout" (not that I think that makes much difference when you can't play either device like an actual piano, and at least the Deluge has a luscious isomorphic keyboard) and the large screen (though the screen isn't really the primary interface on the Deluge, taking a more supporting role of dynamic visual feedback to the real interface, which is pushing the physical buttons - you can get quite far on the Deluge without even looking at the screen.)


The Deluge came out a decade after the OP-1


The Deluge was definitely a close runner up. Ultimately the OP-1 won me over on being smaller and having an integrated battery.


Smaller, granted, but the Deluge also has an integrated battery? Lasts ages too.


And it's still a joke. Spend $200 less and get and Osmose[1] and have a keyboard that you can actually use, a better synth engine, and something that is actually used for music production. They even off free shipping. Portability is overrated when your keyboard tech is stuck in the early 90's, without velocity sensitivity let alone aftertouch and MPE.

1: https://www.expressivee.com/2-osmose


The original OP-1 was $600 less than that when I got it for noodling around with music during my commute to work. So yes, if you choose to ignore some criteria, there's plenty of options.


The OP1-F has velocity sensitivity now.


And it costs what, $600-800 more than the original OP-1?

The "field" reboot is such a sad joke. All that fans had been asking for, quite literally for decades, was for Teenage Engineering to fix the supply chain for OP-1 and produce them in-volume. Instead they upped the price and made it even more rare.


an iphone + korg nano series 2 :)


That's not standalone, that's two pieces.


try the latest Ableton Push, or NI Maschine+


OP-1 is under half the size.

The whole kitsch with that device is being able to make music anywhere.


Which is cute, but every single YouTuber you've seen fiddle with it bounces their stems to Ableton Live and mixes on their computer.


Neat! Those weren't available last time I looked around. Thanks!


> If you praise te, I think it is unfair to other people who are obsessed with technological innovation.

Lol this is music hardware. It's value is the joy people get from using it. The OP-1 has a huge fanbase that praise the workflow and the interface specifically. If "hardcore technology" is important to you then fine but I don't think that applies to most people buying synthesizers - right before the OP-1 the entire industry was deep in a trend of analog reissues.

FWIW I don't think an OP-1 is worth the asking price but after a short time with one it's clear where the money went.


I do some hobby music production and have no idea where I would incorporate any TE equipment into my life, but I can't explain it, I have this visceral reaction that I want to buy their stuff. It just looks so cool. Thankfully it's not at my "won't use more than 3 days then throw in my music closet" price level so I've never even come close to ordering anything.

I've been tempted to get that cheaper Pocket Operator to see if I'm missing out on something but have abstained so far.

But damn the stuff looks cool.


The CEO/Founder of Rabbit is on the board of TE.


as innovative as rabbit r1


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: