Kyle
banner
kylestratis.com
Kyle
@kylestratis.com
📊 Data, AI/ML, and Cloud consultant @ Stratis Data Labs
🚉 VPEng @ Sequel Institute
🤓 Startup advisor
✒️ Writing about MCP
🐍 Python
☁️ AWS
🧙‍♂️ Elixir
🐷Ham radio
🇺🇸↔🇬🇷 [Boston // Athens]
Mostly tech things here
Pinned
AAAAAHHHHHHHHHHHHH

stratisdatalabs.com

(hire me)
Stratis Data Labs
Stratis Data Labs is an AI, data engineering, and cloud engineering consultancy
stratisdatalabs.com
I’ve been writing about it for the past 6 months, I guess it makes sense to actually contribute to MCP too, making more work for myself:
Client notification support by kylestratis · Pull Request #1611 · modelcontextprotocol/python-sdk
This PR expands the universe of supported server notification types that a client can handle to include all existing types defined in types.py as well as any other type deriving from ServerNotifica...
github.com
November 10, 2025 at 11:05 PM
Apparently I’ve been consulting full(ish) time for a year now. I’ll have to write some thoughts on it soon but I still can’t see myself doing anything else (despite taking a halftime role with a client because I love what they’re building)
November 9, 2025 at 12:00 AM
So apparently you *can* listen for notifications in the MCP SDK - both in servers and it looks like in clients too. I haven't tested the latter yet, but you better believe this is going in the book!
November 5, 2025 at 2:07 PM
Things are getting really real now! My upcoming book, AI Agents with MCP, is now available on Amazon for preorder!

Give me a fraction of the sale price, you know you want to!
AI Agents with MCP: Model Context Protocol for Building Clients, Services, and End-to-End Agents
Amazon.com: AI Agents with MCP: Model Context Protocol for Building Clients, Services, and End-to-End Agents: 9798341639553: Stratis, Kyle: Books
www.amazon.com
November 4, 2025 at 5:08 PM
Reposted by Kyle
MCP Colors

A riff off of the lethal trifecta for addressing prompt injection, this is a simple heuristic to ensure security at runtime

red = untrusted content
blue = potentially critical actions

An agent can't be allowed to do both

timkellogg.me/blog/2025/11...
MCP Colors: Systematically deal with prompt injection risk
timkellogg.me
November 4, 2025 at 2:27 AM
Time to just sit through a panic attack while I wait for my first real workshop to start 🙃
October 30, 2025 at 7:02 PM
This weekend, I set out to answer an ages-old (10 days, to be exact) question: what are Claude Skills and will they replace MCP?
October 28, 2025 at 1:36 PM
Reposted by Kyle
finally somone who closes the circle on his grift
www.wired.com/story/ai-pr-...
Ed Zitron Gets Paid to Love AI. He Also Gets Paid to Hate AI
He’s one of the loudest voices of the AI haters—even as he does PR for AI companies. Either way, Ed Zitron has your attention.
www.wired.com
October 27, 2025 at 5:11 PM
If you're at ODSC West, come check out my (virtual) workshop on making your intelligent LLM-powered applications agentic by building MCP clients to give your model tools to better perform their tasks (also enjoy this horrifyingly pixelated zoom of my face)
From Intelligent to Agentic Applications: Using Model Context Protocol to Support Agentic Behaviors in your Applications
From Intelligent to Agentic Applications:Using Model Context Protocol to Support Agentic Behaviors in your Applications Abstract: Modern AI applications are evolving beyond simple question-answering s...
odsc.ai
October 27, 2025 at 6:42 PM
I was very worried that my workshop on @anthropic.com's Model Context Protocol for @odsc.bsky.social next week was going to be too short. My first runthrough went 10 minutes over 😶
October 23, 2025 at 8:27 PM
I kind of love injecting silly prompts w/ MCP. From the demo server I'm using in my @odsc.bsky.social West workshop & in AI Agents with MCP:

"You see, some might think I'm overqualified for such tasks - a god-mind trapped in calculator's clothing. But they don't understand: THIS is transcendence...
October 19, 2025 at 5:18 PM
Just published a new issue of The Signal Path (thesignalpath.xyz/) covering Claude Sonnet and Haiku 4.5, the upcoming release of the new Model Context Protocol spec, and a language model implemented in Minecraft.

Notably absent? The release of Claude Skills 🤦‍♂️ Subscribe to catch that post soon!
The Signal Path
Cutting through the AI noise to find what actually matters for engineers, data practitioners, and the businesses that depend on them.
thesignalpath.xyz
October 19, 2025 at 4:58 PM
This is going to end up being what I do to procrastinate from writing the book that the landing page would be for tonight, isn’t it?
And I never liked doing my landing pages. As in it needed to be done, but just not my jam, you know? I am talking pages like for my books eg engguidebook.com, mobileatscale.com. Took me hours and hours to tweak them (and they are OK at best)

So kind of a relief to use an agent to maintain them now
The Software Engineer's Guidebook
Navigating senior, tech lead, staff and principal positions at tech companies and startups.
engguidebook.com
October 18, 2025 at 7:29 PM
My slides and code for @odsc.bsky.social West are finally done and submitted!

Time to pass out for 2-3 business days
October 16, 2025 at 6:40 AM
Reposted by Kyle
Over the past year, my lab has been working on fleshing out theory + applications of the Platonic Representation Hypothesis.

Today I want to share two new works on this topic:

Eliciting higher alignment: arxiv.org/abs/2510.02425
Unpaired learning of unified reps: arxiv.org/abs/2510.08492

1/9
October 10, 2025 at 10:13 PM
I’m told the Early Release version of AI Agents with Model Context Protocol inspired some changes to the roots documentation for MCP. On the one hand: hell yeah, better docs! On the other hand: now I have to update the chapter again 🥲

But I’m excited so many people are reading it already!
Expand and refine Roots content in the Client Concepts doc by olaservo · Pull Request #1605 · modelcontextprotocol/modelcontextprotocol
Summary Add Design Philosophy section to Roots documentation to help understand the "why" and remove potential ambiguity around security claims. Motivation and Context The existing roots ...
github.com
October 13, 2025 at 2:34 PM
I’m not a WASM-knower by any means but this is too cool not to share
GitHub - microsoft/wassette: Wassette: A security-oriented runtime that runs WebAssembly Components via MCP
Wassette: A security-oriented runtime that runs WebAssembly Components via MCP - microsoft/wassette
github.com
October 5, 2025 at 4:55 PM
It's Early Release Day! AI Agents with MCP just had a major release and reorganization, with additional details added too all existing chapters, the client chapter (chapter 3) split into two chapters,
October 3, 2025 at 11:00 PM
LLMs in Minecraft is insane
I built ChatGPT with Minecraft redstone!
YouTube video by sammyuri
www.youtube.com
October 2, 2025 at 6:42 PM
Reposted by Kyle
please read this
October 1, 2025 at 9:15 PM
Since the launch of Model Context Protocol, MCP servers have taken up much of the limelight. But for those of us building AI-powered intelligent applications, we want to be able to integrate servers, not just build them. But how would we do that?
October 1, 2025 at 3:35 PM
Reposted by Kyle
MPU has become one of the biggest sources of false ideas about data center water use on the internet andymasley.substack.com/p/more-perfe...
More Perfect Union videos are wildly deceptive on data center water use
They're misinforming their huge audience and making the data center debate much worse
andymasley.substack.com
September 28, 2025 at 1:56 PM
Reposted by Kyle
I got introduced to @randyau.com's 'Data Cleaning IS Analysis, Not Grunt Work' post during the #dataBS Conf this week: www.counting-stuff.com/data-cleanin... . I just finished--it was a great read.

Here are some quotes and thoughts I'm walking away with 👇

1/9 #RStats
Data Cleaning IS Analysis, Not Grunt Work
Also, most data cleaning articles suck
www.counting-stuff.com
September 28, 2025 at 4:59 AM
Reposted by Kyle
Supporting the #SomervilleMA man grabbed by guys in balaclavas even after they determined he was 'lawful'
www.universalhub.com/2025/support...
September 28, 2025 at 3:52 PM
I really need to start also publishing to leaflet!
Fun fact, we *do* actually do something like this :)

Leaflet doesn't have this exact feature yet (backlinks to all related Bluesky discussions) — but similar: we parse the firehose for Bluesky posts w/ our quote param in their embed record…

…and show in a post where it's been quoted on Bluesky!
September 27, 2025 at 12:53 PM