New MCP Specification Addresses the Main Barrier To Enterprise Adoption
- Reference: 0184775004
- News link: https://slashdot.org/story/26/07/30/2053239/new-mcp-specification-addresses-the-main-barrier-to-enterprise-adoption
- Source link:
> This week, the [1]Model Context Protocol (MCP), an open source standard for how AI systems interact with external tools and data sources, saw its largest update since [2]its introduction . Most notably, MCP's protocol core is now stateless, so requests are [3]no longer dependent on a session tied to an individual server instance . This change has the potential to address long-standing barriers to scalability.
>
> The blog post [4]announcing the specification , written by lead maintainers David Soria Parra and Den Delimarsky (who both work at Anthropic), says: "The highlight of this release is a stateless protocol core -- MCP is transforming from a bidirectional stateful protocol into a request/response stateless protocol. It was one of the most highly-requested features from developers who were eager to get better reliability and scalability for their MCP servers."
>
> [...] There is also a new deprecation policy that ensures at least 12 months between when a feature's formal deprecation is enacted and when the feature may actually be removed -- with a narrow exception for critical security updates. This is again in keeping with the general "let's make this work better at enterprise scale" theme of the new specification.
This update is "MCP's most important since remote MCP first launched over a year ago," Soria Parra wrote. Other additions include "Multi Round-Trip Requests, header-based routing, cacheable list results, authorization hardening, a formal extensions framework, and updated Tier 1 SDKs."
A full list of changes can be found [5]here .
[1] https://github.com/modelcontextprotocol/modelcontextprotocol
[2] https://slashdot.org/story/25/04/01/208221/mcp-the-new-usb-c-for-ai-thats-bringing-fierce-rivals-together
[3] https://arstechnica.com/ai/2026/07/with-a-stateless-makeover-new-mcp-spec-targets-enterprise-scale/
[4] https://blog.modelcontextprotocol.io/posts/2026-07-28/
[5] https://modelcontextprotocol.io/specification/2026-07-28/changelog
AI doesn't understand you. (Score:2)
[1]AI Doesn't Simulate Reality [x.com].
“It simulates fluency. It calculates text. We are living through the greatest psychological gaslighting in the history of technology.
Every day, Silicon Valley founders stand on stages and promise that if we just feed enough internet data into a massive GPU cluster, a conscious mind will magically emerge from the server rack. They show us models that can write poetry, pass the bar exam, and simulate empathy, and they tell us we are standing on the brink of Artificial G
[1] https://x.com/rodeojuly/status/2081321804384088210
Re: (Score:3)
And?
AI is a tool. My hammer doesn't need to grok the reality of the nail to be useful for hammering it in. By the same token, AI doesn't need to simulate reality to do useful things in reality.
Re: (Score:1)
I don't let my hammer drive a car.
Re: AI doesn't understand you. (Score:2)
And Waymo was working pretty well before Gemini came along.
Re: AI doesn't understand you. (Score:1)
Unfortunately no. Wayno has failed at all but the most happy-path uses. Give it anything even mildly unusual and it fails. It will do fine for a small sliver of urban core driving but it will not be replacing truck drivers ever.
Re: AI doesn't understand you. (Score:2)
Urban core is way harder to do than highways, technically speaking. Inter-state regulations are more of an obstacle there.
Re: AI doesn't understand you. (Score:2)
The parent has the right of it. I'm honestly not that interested in AGI or machine conscious. Whatever Claude Code and Codex are doing is useful to me.
Getting the word out. (Score:2)
Great. Amazon will soon be flooded with self-published books written partially by AI telling about the new changes.
So it fixes the security? (Score:2)
I love the fact that this article claims the specification fixes the "main barrier to Enterprise Adoption (of AI slop)" right after an article saying basically AI will never be secure and you need to stop pretending it ever will be.
So it's unreliable, removes business knowledge from the employees of a company and puts it in the hands of a hostile third party, and now it turns out it's insecure, but adoption of the slobtasystem is only being held back by some specification? Really?
Congrats (Score:2)
Congratulations, you just reimplemented REST with additional steps.
deploy tron! (Score:3)
deploy tron!
Re: (Score:2)
Haha! I came here to ask whether the Model Context Protocol also wanted to kill Tron.
Re: (Score:2)
That's like asking if TCP/IP wants to Kill Bill.
Re: (Score:2)
Really? I don't see the letters K or B in TCP/IP.
Re: deploy tron! (Score:2)
And there is no T R O or N in MCP. Are you some kind of idiot?
Re: (Score:2)
Yeah, I can't make any sense of your posts either.
Re: deploy tron! (Score:2)
MCP and TCP/IP are protocols, and Tron and Kill Bill. Protocols don't want to kill movies. Asking if a protocol wants to kill a movie makes zero sense, which was the point of my initial reply to the GP. The second sentence in my reply to you was rhetorical. It is self evident based on the fact that you couldn't follow the logic.
Re: deploy tron! (Score:2)
You haven't seen Tron, obviously.
Re: deploy tron! (Score:2)
If you are taking about the shitty remake then you are correct. It doesn't matter though, because protocols can't kill. That's not how protocols work.
Re: deploy tron! (Score:2)
I could explain the joke, but it's more entertaining not to. You'll figure it out if you've seen the original.
Re: deploy tron! (Score:2)
I won't figure it out because the premise of Tron is stupid. I'd sooner watch the Matrix all the way through than Tron and I don't have that much lack of taste. I prefer good movies where off the chart level of suspension of disbelief aren't required.
Re: deploy tron! (Score:2)
Master Control Program, as played by the most excellent Ed Dillinger
Re: deploy tron! (Score:2)
Thanks for explaining the "joke" ... Now I see why it definitely wasn't funny.
Re: deploy tron! (Score:2)
The Matrix is from 1999. Tron is from 1982. Totally different eras of the computer age.
Re: deploy tron! (Score:2)
I'm well aware. I was writing software before the IBM PC existed, and before the Apple Lisa hit the market. I didn't mention Tron and the Matrix because I thought they were both released at the same time. I can't imagine why you thought that, even if I didn't already know that information, it would in any way be pertinent.