The thing that a lot of people miss with MCP is that it's the right fit for enterprise software. LLMs, being universal translators, are the ideal glue between many disconnected systems that are extremely hard to connect without some fuzzy layer in between. And so that's why you see so much of the B2B SaaS world rolling out MCP servers, and internally at these companies, they're talking about how to re-jigger their APIs and restrictions for those APIs given different usage patterns.
Yes, the protocol is not necessarily "enterprise ready" by various definitions, but as the author points out, this is not terribly important, and the history of standards shows us that even messy and "bad" things get widespread adoption because they hit the right notes at the right time for the right people.
Karrot_Kream · 33m ago
MCP is just RPC over a long-lived connection, most cases being a Websocket. IMO RPC is easier to setup because:
1. There's less bikeshedding for the authors about whether modifying a field in a user object should be a PUT or a POST replacing the original user object. I've spent a non-trivial amount of time bikeshedding over REST verbs.
2. LLMs don't need to understand the REST semantics of the API. It can just look at available RPC methods and make the RPC call it thinks will work.
I think that's really it.
Animats · 50m ago
It's a great money maker. Each data request involves a paid trip through an LLM.
It's not like the endpoints negotiate a schema they can use to make cheap queries in future.
Flemlo · 1h ago
We have rest and openapi.
That allows already for self discovery etc.
Everyone who will offer mcp offers a good API anyway.
hirsin · 1h ago
We don't need SDKs, they publish API docs.
We don't need packages or functions, we can just write the code ourselves, every time it's needed.
We don't need compilers, the computer understands assembly already.
It's a mismatch of use - MCP isn't about telling the LLM about the APIs. It's also about encapsulating and simplifying them so the LLM isn't manually figuring them out every single time from scratch.
No comments yet
zoogeny · 3h ago
> Compared to the olden days, when specs were written by pedantic old Unix dudes
I think that is one of the reasons (among many others) that the semantic web failed (which doesn't contradict the author, whose point is literally the worse-is-better mantra).
People really leaned into the eXtensible part of XML and I think a certain amount of fatigue set it. XSL, XHTML, XSD, WSDL, XSLT, RDF, RSS, et al. just became a bit too much. It was architecture astronautics for data formats when what the world at the time needed was simple interchange formats (and JSON fit the bill).
But I actually believe XML's time has come. I've noticed that XML appears a lot in leaked system prompts from places like Anthropic. LLMs appear to work very well with structured text formats (Markdown and XML specifically).
I believe that MCP is the wrong model, though. I believe we should be "pushing" context to the models rather than giving them directions on how to "pull" the context themselves.
blitzar · 1m ago
> the semantic web failed
It failed because they couldn't figure out how to stuff ads in.
this_user · 29m ago
I think the main reason that "Semantic Web" has failed is that basically no one can even explain what the "Semantic Web" is or what it's good for. It's all just a bunch of generic sounding buzzwords about exchanging data in some way.
mdaniel · 2h ago
> I believe we should be "pushing" context to the models rather than giving them directions on how to "pull" the context themselves.
How could that possibly work for the cases that people want the intern to solve for them? If they knew the information ahead of time, presumably they'd just solve the problem by themselves
I get the impression that the value folks get from MCP is "run some query for me, don't make me learn how to glue together 15 sources"
zoogeny · 1h ago
I'm not sure I understand your objection. You seem to imply that knowing the context is the same as knowing the solution to the problem the context provides?
Let me think of an example here. Context needed to determine if there is cancer in a radiology scan would be the contents of the scan. So there are two modes here, one I say "LLM please tell me if there is cancer in this patients scan" and the LLM makes an MCP call to load the patients report. The second mode is I say "LLM, here is the patients radiology scan, can you tell me if it has signs of cancer".
The first example is what I was calling a "pull" model and the second example is what I am calling a "push" model.
hirsin · 1h ago
The point above about enterprise glue is why this is a pull model.
In your push model, the onus is on you to go find the scan from one of five backends, traverse whatever hoops of access are needed, and actually handle the files manually.
In the pull model, each backend implements the server once, the LLM gets connected to each one once, and you have one single flow to interact with all of them.
zoogeny · 48m ago
It is interesting that the model I am proposing inverts many peoples expectation of how LLMs will benefit us. In one vision, we give a data-lake of information to LLMs, they tease out the relevant context and then make deductions.
In my view, we hand craft the context and then the LLM makes the deductions.
I guess it will come down to how important crafting the relevant context is for making useful deductions. In my experience with writing code using LLMs, the effectiveness increases when I very carefully select the context and the effectiveness goes down when I let the agent framework (e.g. Cursor) figure out the context. The ideal case is the entire project fits in the context window obviously, but that won't always be possible.
What I've found is that LLMs struggle to ask the right questions. I will often ask the LLM "what other information can I provide you to help solve this problem" and I rarely get a good answer. However, if I know the information that will help it solve the problem and I provide it to the agent then it often does a good job.
ryeguy · 1h ago
MCP is just function calls with parameters. Whether or not it's push or pull can be decided by the author. A push model takes the scan as an input to the mcp call. A pull model does the pulling within the mcp call. Neither is right or wrong, it's situational.
goncalo-r · 1h ago
For a lot of office jobs, knowing the context is almost the same as knowing the solution to the problem at hand.
zoogeny · 1h ago
I'd be interested in examples of this. I've worked in offices for all of my adult life and I don't have any examples that come to mind.
I think of logic puzzles I used to do as a kid. The whole idea of the puzzle is that all of the information you need is provided, the fun is in solving using deduction. Sudoku scratches the same itch.
At the least, I would argue there are many problems that don't fit the mold you are suggesting and MCP is not the correct method for addressing them.
mdaniel · 8m ago
> I'd be interested in examples of this. I've worked in offices for all of my adult life and I don't have any examples that come to mind.
Wow, you must have worked in some really mature shops then if you knew instantly which of [Google Drive, Confluence, Airtable, GitHub wiki, ${that one deprecated thing that Alice was using}, ...] contained the reference to Project Frazlebaz mentioned in Slack last.. day? week? Maybe it was today but time is a blur?
wongarsu · 1h ago
XML tags work well for LLMs. But notably the vast majority are just xml tags. Nobody™ is feeding LLMs well-formed XML with an xml declaration (the <?xml version="1.0" encoding="UTF-8"?> at the start) and we aren't using namespaces, XSLT, XML Schemas etc. It's just some ad-hoc collection of sgml-style tags
zoogeny · 58m ago
I've noticed that as well, but I doubt those additional tokens hurt.
You can make an unholy mess with namespaces and all of the bells-and-whistles that XML provides, or you can not. But even if you just structure using <tag></tag> without any of the fancy stuff, you still can create pipelines that operate on that structured format in ways that are more powerful than plain text.
MrBuddyCasino · 2h ago
Two years ago he would have written „pedantic white old Unix dudes“.
M4rkJW · 1h ago
True, though "dudes who were never in danger of being gruntled" made it pretty clear who he was targeting.
bad_haircut72 · 5h ago
"The rise of MCP gives hope that the popularity of AI amongst coders might pry open all these other platforms to make them programmable for any purpose, not just so that LLMs can control them."
I think the opposite, MCP is destined to fail for the exact same reason the semantic web failed, nobody makes money when things aren't locked down.
It makes me wonder how much functionality of things like AI searching the web for us (sorry, doing "deep-research") might have been solved in better ways. We could have had restaurants publish their menus in a metadata format and anyone could write a python script to say find the cheapest tacos in Texas, but no, the left hand locks down data behind artificial barriers and then the right hand builds AI (datacenters and all) to get around it. On a macro level its just plain stupid.
fidotron · 5h ago
> I think the opposite, MCP is destined to fail for the exact same reason the semantic web failed, nobody makes money when things aren't locked down.
I think this is right. MCP resembles robots.txt evolved into some higher lifeform, but it's still very much "describe your resources for us to exploit them".
The reason the previous agent wave died (it was a Java thing in the 90s) was eventually everyone realized they couldn't trust their code once it was running on a machine it's supposed to be negotiating with. Fundamentally there is an information assymetry problem between interacting agents, entirely by design. Take that away and huge swathes of society will stop functioning.
arbuge · 4h ago
"describe your resources for us to exploit them"
What you want to do is offer resources that make you money when they're "exploited".
fidotron · 2h ago
I would agree with that if there were no distinction between clients and servers. i.e. agents and LLMs are resources that should be discovered and exploited in the same exact way as anything else, and switchable in the same ways.
The whole thing reminds me of stuff like Alan Kay's legendary OOPSLA talk in 1997 ( https://www.youtube.com/watch?v=oKg1hTOQXoY ) "Every object should have an IP" (Also "Arrogance in computer science is measured in nano Dijkstras").
renatovico · 2h ago
i think the problem is in business process, it´s created for no work automatic, the people need be on control OR need have block because well if all people try cancel your subscription and can go in one step with a simple prompt, this is a huge revenue loss.
look how all companies have super system for crm/sales but when you go to backoffice all run in sheets and sometimes in real paper.
jsnell · 4h ago
It's not just that nobody makes money providing a free and open API. It's that to operate such an API you'll basically need unlimited resources. No matter how many resources you throw at the problem, somebody will still figure out a way of exhausting those resources for marginal gains. MCP will just make the problem worse as AI agents descend on any open MCP servers like locusts.
The only stable option, I think, is going to be pay-per-call RPC pricing. It's at least more viable to do then it was for Web 2.0 APIs, since at least the entity operating the model / agent will act as a clearinghouse for all the payments. (And I guess their most likely billing model is to fold these costs into their subscription plans? That seems like the best way to align incentives.)
doug_durham · 4h ago
Plain human readable text is not an "artificial barrier". It the nature of our our world. Requiring that a restaurant publish menus in a metadata format is an artificial barrier. That the beauty of these new NLP tools. I don't need to have a restaurant owner learn JSON, or buy a software package that generates JSON. We can use data as it is. The cost of building useful tools goes to near zero. It will be imprecise, but that's what human language is.
Y_Y · 4h ago
Plain text menus would have been fine
a1j9o94 · 1h ago
How do you do things like compare prices in plain text?
lucideer · 3h ago
> It the nature of our our world.
It's the nature of capitalism.
Some forms of capitalism may have roots in the natural world - natural selection as both a destructive & wasteful competitive process certainly has a lot of parallels in idealised markets - but there's nothing inherent about your menu example when it comes to the modern human world, beyond restrictions placed upon us by capitalism.
> Requiring that a restaurant publish menus in a metadata format is an artificial barrier
This is oddly phrased as noone would need to require anyone to do anything - it's obviously beneficial to a restaurant to publish their menus in formats that are as broadly usable as they can. The only barrier to them doing that is access to tools.
The various hurdles you're describing ("buying" software, the "cost" of building tools) are not natural phenomena.
ljm · 4h ago
HATEOAS was the dream in the early 2010s and that basically went nowhere beyond generating swagger yaml, despite the fact it intended to make API consumption trivial.
Whoever coined it as HATEOAS basically set it up to fail though.
johnmaguire · 4h ago
> Whoever coined it as HATEOAS basically set it up to fail though.
I could never understand making the term "hate" so prominent.
dragonwriter · 4h ago
> HATEOAS was the dream in the early 2010s and that basically went nowhere
I dunno, HTTP/1.1, the motivating use case for REST and HATEOAS, seems to have been moderately successful.
badgersnake · 4h ago
MCP is just that again, but less well thought out. Everything new is old.
jjfoooo4 · 3h ago
MCP is described as a means to make the web open, but it’s actually it’s a means to make demos of neat things you could do if the web were actually open.
isodev · 4h ago
I think MCP’s popularity is a side effect of the hype bubble driving AI atm - one of the fancy things one can do with AI.
If there was any “easy” value in making one’s data available in a standard form, we would’ve seen a lot more adoption of interoperable endpoints (e.g. using schema.org or generally common ontologies as opposed to custom formats that always need a special magic SDK).
doug_durham · 4h ago
There is an easy way to make your data available. It's existed for several hundred years, it's called plain text. We now have tools that allow computers to work with plain text. Outside of specific niches ontologies are vanity projects.
alberth · 3h ago
Sure - not many companies made money on "HTTP", but lots of people/companies made gobs of money by adopting it.
philosophty · 4h ago
I haven't paid close attention. Why can't people make money with MCP-based APIs? Why can't providers require API keys / payment to call their functions?
olalonde · 3h ago
Sure they can - they're just another API interface tailored for LLMs. I think parent and OP are in fact ranting about that (many APIs being locked behind signups or paywalls). Not sure I agree with the criticism though. In my view, web 2.0 was a huge success: we went from a world with almost no APIs to one where nearly every major website or app offers one. That's real progress, even if we didn't turn every business into an open data non-profit.
throwaway7783 · 4h ago
MCP is basically APIs V2 as far as I can see. It probably will evolve in its concrete specs, but useful and not niche, especially when they can be composed fairly trivially.
In that sense, it is probably the building block for the next user interface, which is conversational.
Maybe the mention of web 2.0 is triggering all negative responses here, but on it's own, it is useful and could disrupt (not MCP itself, but the overall field) how we interact with software and systems
drusepth · 3h ago
> I think the opposite, MCP is destined to fail for the exact same reason the semantic web failed, nobody makes money when things aren't locked down.
Is there a way to handle "locking down" things with MCP? It seems like a potential opportunity for freemium services if they have a mechanism for authentication and telling the MCP caller "this user has access to these tools, but not _these_ yet".
seanhunter · 3h ago
Yes. MCP allows (and uses) exactly the same authentication mechanisms that any other rest or similar api allows. So if you have a service you want to expose (or not) via MCP you can do that in exactly the same way as you currently could do that for a rest API.
The difference for the user is instead of them having to make (or use) a special-purpose client to call your rest api, the llm (or llm powered application) can just call the api for them, meaning your rest service can be integrated into other llm-powered workflows.
Joker_vD · 4h ago
The reason the semantic web failed is not only because "nobody makes money when things aren't locked down". It's also because nobody ain't got no time for generating infinite amount of metadata when full-text search and indexing, with a judicious pinch of fuzzy matching, is both faster and more reliable. And LLMs, as much as I dislike the technological/societal consequences of their existence, are effectively further development of the latter, so they won't go away.
Manual or even semi-automated cataloguing of websites (and further curating) of websites wasn't the answer to "how do I find stuff on the web" — Google was. Having standardized metadata format for menus is undoubtedly nice — but good luck making people to use it. You just can't. It really is both cheaper and easier for everyone involved to have website with arbitrary information layout scraped and fed into an LLM to extract relevant data: because what is "relevant" is different for everyone. You can't pre-ordain the full list of possible relevant metadata, and, again, good luck forcing everyone to fill out those 500 items-long forms. Ain't nobody got time for that.
jauntywundrkind · 3h ago
I tend to agree one of the top semantic web problems was:
> It's also because nobody ain't got no time for generating infinite amount of metadata
There's also a lot of tooling problems too, that the semantic web doesn't integrate gracefully with POJO of the programming worlds.)
The tooling distance between users/devs and semantic web remains. But all that metadata? There being an interesting rich world of information, associated & well described & meticulous? Uh we actually seem like we just invented a really powerful tool at doing all this immense catalogization (LLM's).
gz5 · 4h ago
>nobody makes money when things aren't locked down
i would rephrase as "incumbents don't usually make more money if things are opened up".
if consumer gets materially better value, then challenger ecosystem around MCP will evolve. it will be open at first - great for startups and challengers, innovator's dilemma for market leaders.
and then it will close as the new wave establishes their moats. but, similar to web, even though the current web leaders are now much more closed than we would like, the overall ecosystem is more open than it was.
testplzignore · 3h ago
> On a macro level its just plain stupid.
You've described most white-collar jobs :)
AlienRobot · 2h ago
I still don't know who uses this semantic web. Like you have all these semantics marked up... for whom? What are actual applications using this?
Google has a small subset of schema.org it supports, but rather than "semantic web" it feels more like "here's my API." Its own schema tester often complains about things that should be valid schemas, simply because it doesn't conform to its API. How would any web developer mark up (and test the validity of said mark up) for applications that don't even exist?
throwaway13337 · 4h ago
It's reasonable to be cynical, but the future hasn't been written yet. If we choose only to see a negative future, we will ensure that it can only exist.
In the negative vein, I see a lot of VCs and business leaders talking about making AI for companies that directly interface with customers.
Those agents will be used to manipulate and make painful existings services exactly like today. Enshitified transactional websites engineered for maximum pain.
A different direction can happen if we choose instead to use our ai agents to interact with business services. This is actually what's currently happening.
I use gemini/chatgpt to find things on the web for me without being manipulated and marketed at. Maybe one day soon, I can buy airline tickets without the minefield of dark patterns employed to maximize A/B tested revenue.
The only thing that needs to happen to keep us on this path is to bite the heels of the major companies with great agent systems that put the user at the center and not a brand. That means selling AI agents as a SaaS or open source projects - not ad-supported models.
This community is probably the group that has, collectively, the most say in where this future goes. Let's choose optimism.
jahewson · 4h ago
The thing is, if AI agents become a significant part of web traffic then the content of the web will simply shift to manipulate the agent instead of the human.
And don’t forget when you use an AI agent today to buy something it’s using “marketing” information to make its decisions. It’s influenced by SEO in its search results, indeed there’s no shortage of marketers busy working out how to do this.
I do agree there’s much to be optimistic about but the fundamental dynamics of the consumer market won’t change.
throwaway13337 · 4h ago
It's absolutely true that in that future vision, the agents will then be marketed at.
And that's great.
In that world, those agents will sift through the noise. And the one that does that the best will win.
The end user experience then becomes uniform and pleasant.
It's the difference between talking to a personal secretary and a customer service representative. No one should have to endure the latter.
kibwen · 3h ago
> In that world, those agents will sift through the noise. And the one that does that the best will win.
The existence of agents capable of learning to cut through the enshittification also implies the existence of agents capable of learning to enshittify all the more effectively. It's an arms race, and there's no reason to suspect that the pro-consumer side will win over the pro-exploitation side.
FuriouslyAdrift · 3h ago
Oh SGML... what could have been
pphysch · 4h ago
xAI is a concrete example of this. During the initial LLM explosion, X locked down its previously public APIs and data sources. Simultaneously, xAI is investing massively in building its private data hoard and compute infrastructure. Probably a similar case with Meta.
"Data for me but not for thee"
MCP is only getting the light of day, arguably, because of LLM "one trick ponies" like OpenAI and Anthropic, who do benefit from MCP amplifying their value proposition. As that business model continues to fizzle out and lose/subordinate to the AI integrators (Google, Microsoft, xAI?), MCP will probably fizzle out as well.
vasusen · 1h ago
What worries me most about MCP isn't that the protocol is so poorly created, but that fixing/improving it is only at the mercy of internal teams at Anthropic and OpenAI. It doesn't seem like the folks who are coming up with the protocol are actual engineers trying to implement it as well.
Vaguely seems like a Visa-Mastercard duopoly.
TOMDM · 1h ago
Microsoft has joined the steering committee as well
Isn’t conformance to a standard API arguably less necessary now that LLMs can read API docs and adapt? For me, the win is the expectation that sites _have_ an API, whether it conforms to the MCP spec or not.
bilbo0s · 1h ago
-API documentation may have been poorly written.
-Even with good documentation, the LLM may generate incorrect code to interact with API. (If you fix the generated code and just have the LLM call that code to interact with the API, then you're on your way to an intermediary in any case. You're basically building an "MCP-like" server.)
-Security or resource allocation issues when giving LLMs direct access to APIs. (LLM has limited knowledge of when API was last called. It could call too often and, if each call is expensive, generate surprise infrastructure bills.)
-So on and so forth. Lots of other potential pain points solved by having something intermediate.
Should that "something" be MCP? Reasonable people can disagree on that. Works well enough right now to get done what people need done I suppose.
lxgr · 5h ago
Turns out the “Semantic Web” was a syntactic web all along, and maybe this is the real deal?
foobarian · 2h ago
> when specs were written by pedantic old Unix dudes
I'm tickled pink that this generation imagines "old Unix dudes" as pedantic, when Unix was the ultimate "move fast and break things" rebellion against the MIT school. Some things never change :-)
no_wizard · 2h ago
My hot take: MCP's value for the foreseeable future will be with automated testing
daemonk · 4h ago
At a higher level, MCP seems wants to enforce a standard where no standard exists. I get that the low level technical implementation allows AI to utilize these tools.
But there doesn't seem to be any standardization or method in how to describe the tool to the AI so that it can utilize it well. And I guess part of the power of AI is that you shouldn't need to standardize that? But shouldn't there at least be some way to describe the tool's functionality in natural language or give some context to the tool?
awestroke · 1h ago
Each prompt, tool or resource fetched from an MCP server includes a name, description and a schema. What more do you want?
I pitty the fools thinking they will have access to anything because there's a MCP.
those things will be hidden behind a dozen layers of payment validation and authentication. And whitelisted IPs (v4, of course).
ERR 402; is all that will be visible to yall.
vivzkestrel · 3h ago
Everything seems to be susceptible to enshittification and so far I see no evidence that MCP is any exception. First the value will go to users, then the users ll cut short to drive value to shareholders and then it will turn to an absolute pile of garbage as businesses make every attempt to somehow cash in on this
yupitsme123 · 1h ago
It's a protocol. The new version of the web that gets built with it may get shitty, but like HTTP or TCP, the protocol stands on its own merits and can be used freely to build cool stuff outside of that shitty ecosystem.
quantadev · 4h ago
We can now build the Semantic Web. All we have to do is create a tiny protocol (as an optional extension to MCP) for how organizations can share their SQL Table Create DDL as a static file that MCP apps can read, to understand data, and then, using the already-existing tools for AI/LLM function calling to SQL, that would become a Semantic Web.
That would fill the missing link that always held back the Semantic Web which was the lack of any incentive for companies to bother to use a standard "Data Type" rather than all proprietary data types. Once we have an MCPQ (MCP with Queries), suddenly there's an incentive for organizations to collaborate at the data structure layer.
RansomStark · 3h ago
MCP could have cracked the web opeb. The terrible standard was all about clients and local servers all on the same host.
Imagine it, everything is open, servers are as simple as a pip install ...
You have full control of what servers you install. What functions you turn on, what access you allow.
Now everyone and their blog is sticking MCPs on their servers and locking them down behind subscriptions and paywalls.
What a wasted opportunity.
freeone3000 · 2h ago
And what pays for the resources used serving your (hundreds of) requests against a “local” server? For computer control, sure, but actual remote services have actual remote costs.
tagfowufe · 5h ago
While I understand where the author is coming from, and I get his sentiment(s), I don't think what he proposes is actually possible: his vision relies on faux open tools and protocols and having access to walled gardens. The means of computation for these kinds of things are owned by a tiny minority. Nearly everything is a SaaS or is based, one way or the other, on rent extraction. We're essentially subject to the whims of someone who is letting us do something for as long as we play nice.
>There is a chance, though, that younger developers, and those who weren't around to build back during that last era a generation ago, are going to get inspired by MCP to push for the web to go back towards its natural architecture. It was never meant to be proprietary.
Alas, the reason APIs started closing and being metered is because, after all, there's someone owning and paying for the hardware upon which you are making calls and requests.
As long as there's no way to agree upon how to have a bunch of servers providing computation for anyone and at the same time ensuring their upkeep without the need for a central authority, I don't think such vision is sustainable long term. The current state of the Internet is proof of it.
No comments yet
jaredcwhite · 5h ago
MCP is not an open standard.
People routinely mistake "protocol specification uploaded to GitHub, PRs welcome" as open standards. They are not. Calling them "open protocols" because they are open source, not open standards (no standards body was involved in the making of this protocol!) is essentially a form of openwashing.
This has been happening way too frequently lately (see also: ATProto), and it really needs to be called out.
underbluewaters · 4h ago
Successful standards usually start out scrappy, are embraced by a community, and then are blessed by standards bodies. What comes out of working groups of standards bodies rarely gains traction. See xhtml vs "html5".
croes · 4h ago
Does MCP still have the security issues?
A big mistake in the first place to start it without proper security.
That not Web 2.0 2.0, that‘s Web 1.0
doug_durham · 4h ago
It's a mistake to not start at all because of an issue that may not be important for many important use cases.
troupo · 32m ago
It's a mistake to invent your own terminology, skip all existing forms/standards/protocols of communication and chose the objectively worst-fitting one, and skip implementing one of the most fundamental part required by nearly anything
croes · 4h ago
It’s a mistake to start and to think security is a non important issue.
Anything that’s connected to the web has to consider security at step one.
How often are we repeating the same mistakes over and over again?
jedisct1 · 4h ago
MCP Servers are usually installed locally and can do whatever they want on the local machine.
But this is solved by sandboxes such as mcp.run .
croes · 3h ago
Sandboxes aren’t a solution just a workaround for a bigger problem
nicoburns · 5h ago
> no standards body was involved in the making of this protocol!
Is a standards body being involved relevant? Many standards ratified by standards bodies are "pay to access" and seem much less open than many de facto standards where no standards body was involved.
isodev · 4h ago
> Is a standards body being involved relevant?
I believe it is. Taking the example of ATproto, it boils down to a managed platform as a service but they seem to do a #BuildInPublic thing where we can experience early developer previews. That’s not really open and it isn’t a standard.
philosophty · 4h ago
How are these not "open" if they're GitHub projects and entirely open source and open to outside contribution?
The organizations that create "standards" are much less open than this, requiring in person lobbying, schmoozing, travel, and company affiliations.
anildash · 4h ago
I was using the phrase in the vernacular sense; I’ve worked on genuine open standards and know the difference. I don’t think it matters that much to use the phrase in this way in an obviously casual piece on my personal blog, as opposed to a more formal assessment of a technology.
dangoor · 4h ago
It has become a de facto standard. There are many implementations of MCP. We'll have to see if it fizzles out or if some reasonable stewardship/governance come about, but it is very much possible for something to be a standard even without a standards body.
98codes · 5h ago
Open standard or not, how is it a standard at all? I spent a while trying to find what the current version number of the spec is, and I could not find one. Theer were SDKs, and the SDKs have versions, but the protocol itself seems to be on v0.0.0.
Partially agree. But I can't help thinking this is the natural lifecycle of protocols. They first start as open projects, proliferate as such, and evolve into standards with governance once they catch on.
What would you call these projects? Open protocols?
Toritori12 · 5h ago
I guess it is a first step... a lot of current "open" protocols started as proprietary.
croes · 4h ago
The first step should include more security.
jacob019 · 2h ago
Fun writing, and something to think about. To me, Web 2.0 is kind of a joke; jQuery, REST, AJAX, CSS2, RSS, single page apps were going to change everything overnight, it was THE buzzword, and then... incremental improvements. In retrospect, everything did change, but that loose collection of technologies was just links in the chain of incremental progress. So yeah, Web 2.0 2.0 makes sense.
I've seen a lot of talk around here, and everywhere, about MCP. A lot of enthusiasm and a lot of criticism. I've written a few MCP servers and plan to write some more. It isn't how I would design a protocol, but it works, and everyone is using it, so hooray for interoperability.
I think the hype represents the wider enthusiasm that people have about this moment in time, and the transformative power of these new tools. It's easy to look at MCP and say there it is, now it's easy to connect these tools to the things that I care about, it feels accessible, and there's community.
MCP is NOT the new TCP for AI. It is, essentially, an RPC layer for chat. And while it could be extended to a wider variety of use cases, I expect that it will remain primarily a method for wiring up tool calls for user-facing use cases. We recognize the power of these tools and anticipate deep changes to workflows and systems, but we don't know how that will shake out. If I am writing a classifier for a backend system, I am not going to use MCP, even if I could. Because it's inefficient. Every additional tool we offer the model consumes tokens and increases latency. I expect that the primary user of LLMs is going to be business automation of all kinds, and I don't expect them to reach for MCP to wire things up. Yeah, it's really cool to hook tools up to the chat, for that to feel accessible, to know how to do things in an idiomatic and standards-compliant way, that feels good! And yeah, the hype is overblown.
No comments yet
CSMastermind · 4h ago
I really wish we'd learn from Web 2.0.
All the mistakes of "hey everything has an API now" that we learned from we're back to repeating.
I feel like that meme from Watchmen with the blue guy sitting on Mars.
olalonde · 3h ago
What mistakes are we repeating?
thuanao · 3h ago
Not understanding that businesses don't want open data or open protocols. Internet businesses want to monopolize information and charge rent for it. Capitalists don't want to commodify software, as that means lower profits and competition. They want to monopolize it. That's the whole game, except for companies selling physical goods over the Internet.
olalonde · 2h ago
So, is the idea that MCP might benefit from more built-in support for things like paid APIs to incentivize businesses?
thuanao · 1h ago
I think so. Though the premise of MCP betrays that AI isn't human-level intelligent. If it was there would be no need for a standard interface. It would be able to read API docs, navigate UIs to generate the proper keys, enter in billing info, and use all of that to make API requests by trial and error. That's what humans do.
anildash · 4h ago
Definitely having that feeling a lot these days watching the cycles repeat. I am tired of earth. These people. Their APIs.
nimish · 3h ago
Rent seeking is the name of the game for much of b2b SaaS.
MCP is an attempt to make that easy, but the issue here is that a lot of the companies offering integration could be disintermediated entirely by LLMs. Hard to say what that means.
No comments yet
quotemstr · 4h ago
People said the same thing about "APIs" 10-15 years ago when they were a craze. Everything had to be an API! Doesn't matter whether it made sense to not. It's going to change the world! We're going to have San Francisco events with microbrews for APIs! Everyone's going to publish API frameworks! Let me make api-blog-blog.blogger.blog!
Blah. Bay Area Tech regularly goes through these bursts of baseless enthusiasm for rehashes of existing technology. Is MCP useful? Yeah, probably. Is the current froth based on anything but SF-scene self-licking-ice-cream-code social cred? No.
rglover · 1h ago
> Bay Area Tech regularly goes through these bursts of baseless enthusiasm for rehashes of existing technology. Is MCP useful? Yeah, probably. Is the current froth based on anything but SF-scene self-licking-ice-cream-code social cred? No.
Thank you. You said what I was going to say. Another goose chase for something that's valuable, certainly, but being frothed up into some mythical sword in the stone.
It's ultimately just a way to say "hey, here's some extra context including some functions we have defined on our server that you can ask us to call."
Yes, the protocol is not necessarily "enterprise ready" by various definitions, but as the author points out, this is not terribly important, and the history of standards shows us that even messy and "bad" things get widespread adoption because they hit the right notes at the right time for the right people.
1. There's less bikeshedding for the authors about whether modifying a field in a user object should be a PUT or a POST replacing the original user object. I've spent a non-trivial amount of time bikeshedding over REST verbs.
2. LLMs don't need to understand the REST semantics of the API. It can just look at available RPC methods and make the RPC call it thinks will work.
I think that's really it.
That allows already for self discovery etc.
Everyone who will offer mcp offers a good API anyway.
We don't need packages or functions, we can just write the code ourselves, every time it's needed.
We don't need compilers, the computer understands assembly already.
It's a mismatch of use - MCP isn't about telling the LLM about the APIs. It's also about encapsulating and simplifying them so the LLM isn't manually figuring them out every single time from scratch.
No comments yet
I think that is one of the reasons (among many others) that the semantic web failed (which doesn't contradict the author, whose point is literally the worse-is-better mantra).
People really leaned into the eXtensible part of XML and I think a certain amount of fatigue set it. XSL, XHTML, XSD, WSDL, XSLT, RDF, RSS, et al. just became a bit too much. It was architecture astronautics for data formats when what the world at the time needed was simple interchange formats (and JSON fit the bill).
But I actually believe XML's time has come. I've noticed that XML appears a lot in leaked system prompts from places like Anthropic. LLMs appear to work very well with structured text formats (Markdown and XML specifically).
I believe that MCP is the wrong model, though. I believe we should be "pushing" context to the models rather than giving them directions on how to "pull" the context themselves.
It failed because they couldn't figure out how to stuff ads in.
How could that possibly work for the cases that people want the intern to solve for them? If they knew the information ahead of time, presumably they'd just solve the problem by themselves
I get the impression that the value folks get from MCP is "run some query for me, don't make me learn how to glue together 15 sources"
Let me think of an example here. Context needed to determine if there is cancer in a radiology scan would be the contents of the scan. So there are two modes here, one I say "LLM please tell me if there is cancer in this patients scan" and the LLM makes an MCP call to load the patients report. The second mode is I say "LLM, here is the patients radiology scan, can you tell me if it has signs of cancer".
The first example is what I was calling a "pull" model and the second example is what I am calling a "push" model.
In your push model, the onus is on you to go find the scan from one of five backends, traverse whatever hoops of access are needed, and actually handle the files manually.
In the pull model, each backend implements the server once, the LLM gets connected to each one once, and you have one single flow to interact with all of them.
In my view, we hand craft the context and then the LLM makes the deductions.
I guess it will come down to how important crafting the relevant context is for making useful deductions. In my experience with writing code using LLMs, the effectiveness increases when I very carefully select the context and the effectiveness goes down when I let the agent framework (e.g. Cursor) figure out the context. The ideal case is the entire project fits in the context window obviously, but that won't always be possible.
What I've found is that LLMs struggle to ask the right questions. I will often ask the LLM "what other information can I provide you to help solve this problem" and I rarely get a good answer. However, if I know the information that will help it solve the problem and I provide it to the agent then it often does a good job.
I think of logic puzzles I used to do as a kid. The whole idea of the puzzle is that all of the information you need is provided, the fun is in solving using deduction. Sudoku scratches the same itch.
At the least, I would argue there are many problems that don't fit the mold you are suggesting and MCP is not the correct method for addressing them.
Wow, you must have worked in some really mature shops then if you knew instantly which of [Google Drive, Confluence, Airtable, GitHub wiki, ${that one deprecated thing that Alice was using}, ...] contained the reference to Project Frazlebaz mentioned in Slack last.. day? week? Maybe it was today but time is a blur?
You can make an unholy mess with namespaces and all of the bells-and-whistles that XML provides, or you can not. But even if you just structure using <tag></tag> without any of the fancy stuff, you still can create pipelines that operate on that structured format in ways that are more powerful than plain text.
I think the opposite, MCP is destined to fail for the exact same reason the semantic web failed, nobody makes money when things aren't locked down.
It makes me wonder how much functionality of things like AI searching the web for us (sorry, doing "deep-research") might have been solved in better ways. We could have had restaurants publish their menus in a metadata format and anyone could write a python script to say find the cheapest tacos in Texas, but no, the left hand locks down data behind artificial barriers and then the right hand builds AI (datacenters and all) to get around it. On a macro level its just plain stupid.
I think this is right. MCP resembles robots.txt evolved into some higher lifeform, but it's still very much "describe your resources for us to exploit them".
The reason the previous agent wave died (it was a Java thing in the 90s) was eventually everyone realized they couldn't trust their code once it was running on a machine it's supposed to be negotiating with. Fundamentally there is an information assymetry problem between interacting agents, entirely by design. Take that away and huge swathes of society will stop functioning.
What you want to do is offer resources that make you money when they're "exploited".
The whole thing reminds me of stuff like Alan Kay's legendary OOPSLA talk in 1997 ( https://www.youtube.com/watch?v=oKg1hTOQXoY ) "Every object should have an IP" (Also "Arrogance in computer science is measured in nano Dijkstras").
look how all companies have super system for crm/sales but when you go to backoffice all run in sheets and sometimes in real paper.
The only stable option, I think, is going to be pay-per-call RPC pricing. It's at least more viable to do then it was for Web 2.0 APIs, since at least the entity operating the model / agent will act as a clearinghouse for all the payments. (And I guess their most likely billing model is to fold these costs into their subscription plans? That seems like the best way to align incentives.)
It's the nature of capitalism.
Some forms of capitalism may have roots in the natural world - natural selection as both a destructive & wasteful competitive process certainly has a lot of parallels in idealised markets - but there's nothing inherent about your menu example when it comes to the modern human world, beyond restrictions placed upon us by capitalism.
> Requiring that a restaurant publish menus in a metadata format is an artificial barrier
This is oddly phrased as noone would need to require anyone to do anything - it's obviously beneficial to a restaurant to publish their menus in formats that are as broadly usable as they can. The only barrier to them doing that is access to tools.
The various hurdles you're describing ("buying" software, the "cost" of building tools) are not natural phenomena.
Whoever coined it as HATEOAS basically set it up to fail though.
I could never understand making the term "hate" so prominent.
I dunno, HTTP/1.1, the motivating use case for REST and HATEOAS, seems to have been moderately successful.
If there was any “easy” value in making one’s data available in a standard form, we would’ve seen a lot more adoption of interoperable endpoints (e.g. using schema.org or generally common ontologies as opposed to custom formats that always need a special magic SDK).
In that sense, it is probably the building block for the next user interface, which is conversational.
Maybe the mention of web 2.0 is triggering all negative responses here, but on it's own, it is useful and could disrupt (not MCP itself, but the overall field) how we interact with software and systems
Is there a way to handle "locking down" things with MCP? It seems like a potential opportunity for freemium services if they have a mechanism for authentication and telling the MCP caller "this user has access to these tools, but not _these_ yet".
The difference for the user is instead of them having to make (or use) a special-purpose client to call your rest api, the llm (or llm powered application) can just call the api for them, meaning your rest service can be integrated into other llm-powered workflows.
Manual or even semi-automated cataloguing of websites (and further curating) of websites wasn't the answer to "how do I find stuff on the web" — Google was. Having standardized metadata format for menus is undoubtedly nice — but good luck making people to use it. You just can't. It really is both cheaper and easier for everyone involved to have website with arbitrary information layout scraped and fed into an LLM to extract relevant data: because what is "relevant" is different for everyone. You can't pre-ordain the full list of possible relevant metadata, and, again, good luck forcing everyone to fill out those 500 items-long forms. Ain't nobody got time for that.
> It's also because nobody ain't got no time for generating infinite amount of metadata
There's also a lot of tooling problems too, that the semantic web doesn't integrate gracefully with POJO of the programming worlds.)
The tooling distance between users/devs and semantic web remains. But all that metadata? There being an interesting rich world of information, associated & well described & meticulous? Uh we actually seem like we just invented a really powerful tool at doing all this immense catalogization (LLM's).
i would rephrase as "incumbents don't usually make more money if things are opened up".
if consumer gets materially better value, then challenger ecosystem around MCP will evolve. it will be open at first - great for startups and challengers, innovator's dilemma for market leaders.
and then it will close as the new wave establishes their moats. but, similar to web, even though the current web leaders are now much more closed than we would like, the overall ecosystem is more open than it was.
You've described most white-collar jobs :)
Google has a small subset of schema.org it supports, but rather than "semantic web" it feels more like "here's my API." Its own schema tester often complains about things that should be valid schemas, simply because it doesn't conform to its API. How would any web developer mark up (and test the validity of said mark up) for applications that don't even exist?
In the negative vein, I see a lot of VCs and business leaders talking about making AI for companies that directly interface with customers.
Those agents will be used to manipulate and make painful existings services exactly like today. Enshitified transactional websites engineered for maximum pain.
A different direction can happen if we choose instead to use our ai agents to interact with business services. This is actually what's currently happening.
I use gemini/chatgpt to find things on the web for me without being manipulated and marketed at. Maybe one day soon, I can buy airline tickets without the minefield of dark patterns employed to maximize A/B tested revenue.
The only thing that needs to happen to keep us on this path is to bite the heels of the major companies with great agent systems that put the user at the center and not a brand. That means selling AI agents as a SaaS or open source projects - not ad-supported models.
This community is probably the group that has, collectively, the most say in where this future goes. Let's choose optimism.
And don’t forget when you use an AI agent today to buy something it’s using “marketing” information to make its decisions. It’s influenced by SEO in its search results, indeed there’s no shortage of marketers busy working out how to do this.
I do agree there’s much to be optimistic about but the fundamental dynamics of the consumer market won’t change.
And that's great.
In that world, those agents will sift through the noise. And the one that does that the best will win.
The end user experience then becomes uniform and pleasant.
It's the difference between talking to a personal secretary and a customer service representative. No one should have to endure the latter.
The existence of agents capable of learning to cut through the enshittification also implies the existence of agents capable of learning to enshittify all the more effectively. It's an arms race, and there's no reason to suspect that the pro-consumer side will win over the pro-exploitation side.
"Data for me but not for thee"
MCP is only getting the light of day, arguably, because of LLM "one trick ponies" like OpenAI and Anthropic, who do benefit from MCP amplifying their value proposition. As that business model continues to fizzle out and lose/subordinate to the AI integrators (Google, Microsoft, xAI?), MCP will probably fizzle out as well.
Vaguely seems like a Visa-Mastercard duopoly.
https://techcrunch.com/2025/05/19/github-microsoft-embrace-a...
-Even with good documentation, the LLM may generate incorrect code to interact with API. (If you fix the generated code and just have the LLM call that code to interact with the API, then you're on your way to an intermediary in any case. You're basically building an "MCP-like" server.)
-Security or resource allocation issues when giving LLMs direct access to APIs. (LLM has limited knowledge of when API was last called. It could call too often and, if each call is expensive, generate surprise infrastructure bills.)
-So on and so forth. Lots of other potential pain points solved by having something intermediate.
Should that "something" be MCP? Reasonable people can disagree on that. Works well enough right now to get done what people need done I suppose.
I'm tickled pink that this generation imagines "old Unix dudes" as pedantic, when Unix was the ultimate "move fast and break things" rebellion against the MIT school. Some things never change :-)
But there doesn't seem to be any standardization or method in how to describe the tool to the AI so that it can utilize it well. And I guess part of the power of AI is that you shouldn't need to standardize that? But shouldn't there at least be some way to describe the tool's functionality in natural language or give some context to the tool?
https://modelcontextprotocol.io/specification/2025-03-26/ser...
those things will be hidden behind a dozen layers of payment validation and authentication. And whitelisted IPs (v4, of course).
ERR 402; is all that will be visible to yall.
That would fill the missing link that always held back the Semantic Web which was the lack of any incentive for companies to bother to use a standard "Data Type" rather than all proprietary data types. Once we have an MCPQ (MCP with Queries), suddenly there's an incentive for organizations to collaborate at the data structure layer.
Imagine it, everything is open, servers are as simple as a pip install ... You have full control of what servers you install. What functions you turn on, what access you allow.
Now everyone and their blog is sticking MCPs on their servers and locking them down behind subscriptions and paywalls.
What a wasted opportunity.
>There is a chance, though, that younger developers, and those who weren't around to build back during that last era a generation ago, are going to get inspired by MCP to push for the web to go back towards its natural architecture. It was never meant to be proprietary.
Alas, the reason APIs started closing and being metered is because, after all, there's someone owning and paying for the hardware upon which you are making calls and requests.
As long as there's no way to agree upon how to have a bunch of servers providing computation for anyone and at the same time ensuring their upkeep without the need for a central authority, I don't think such vision is sustainable long term. The current state of the Internet is proof of it.
No comments yet
People routinely mistake "protocol specification uploaded to GitHub, PRs welcome" as open standards. They are not. Calling them "open protocols" because they are open source, not open standards (no standards body was involved in the making of this protocol!) is essentially a form of openwashing.
This has been happening way too frequently lately (see also: ATProto), and it really needs to be called out.
A big mistake in the first place to start it without proper security.
That not Web 2.0 2.0, that‘s Web 1.0
Anything that’s connected to the web has to consider security at step one.
How often are we repeating the same mistakes over and over again?
But this is solved by sandboxes such as mcp.run .
Is a standards body being involved relevant? Many standards ratified by standards bodies are "pay to access" and seem much less open than many de facto standards where no standards body was involved.
I believe it is. Taking the example of ATproto, it boils down to a managed platform as a service but they seem to do a #BuildInPublic thing where we can experience early developer previews. That’s not really open and it isn’t a standard.
The organizations that create "standards" are much less open than this, requiring in person lobbying, schmoozing, travel, and company affiliations.
What would you call these projects? Open protocols?
I've seen a lot of talk around here, and everywhere, about MCP. A lot of enthusiasm and a lot of criticism. I've written a few MCP servers and plan to write some more. It isn't how I would design a protocol, but it works, and everyone is using it, so hooray for interoperability.
I think the hype represents the wider enthusiasm that people have about this moment in time, and the transformative power of these new tools. It's easy to look at MCP and say there it is, now it's easy to connect these tools to the things that I care about, it feels accessible, and there's community.
MCP is NOT the new TCP for AI. It is, essentially, an RPC layer for chat. And while it could be extended to a wider variety of use cases, I expect that it will remain primarily a method for wiring up tool calls for user-facing use cases. We recognize the power of these tools and anticipate deep changes to workflows and systems, but we don't know how that will shake out. If I am writing a classifier for a backend system, I am not going to use MCP, even if I could. Because it's inefficient. Every additional tool we offer the model consumes tokens and increases latency. I expect that the primary user of LLMs is going to be business automation of all kinds, and I don't expect them to reach for MCP to wire things up. Yeah, it's really cool to hook tools up to the chat, for that to feel accessible, to know how to do things in an idiomatic and standards-compliant way, that feels good! And yeah, the hype is overblown.
No comments yet
All the mistakes of "hey everything has an API now" that we learned from we're back to repeating.
I feel like that meme from Watchmen with the blue guy sitting on Mars.
MCP is an attempt to make that easy, but the issue here is that a lot of the companies offering integration could be disintermediated entirely by LLMs. Hard to say what that means.
No comments yet
Blah. Bay Area Tech regularly goes through these bursts of baseless enthusiasm for rehashes of existing technology. Is MCP useful? Yeah, probably. Is the current froth based on anything but SF-scene self-licking-ice-cream-code social cred? No.
Thank you. You said what I was going to say. Another goose chase for something that's valuable, certainly, but being frothed up into some mythical sword in the stone.
It's ultimately just a way to say "hey, here's some extra context including some functions we have defined on our server that you can ask us to call."