Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social

On the move: @mike-hole.bsky.social

[Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social @mikehole]

That sure is a click bait start to this tweet.

[Avi Chawla @_avichawla]

Anthropic's in trouble, again!

They spent years building what's now fully open-source.

What made Claude feel different from a normal app is that the agent could act inside the interface instead of only talking in a chat box.

For instance, Claude Artifacts let an agent render real UI, charts, dashboards, and interactive components that assemble live inside the response.

Every major AI product tried to replicate it.

But the problem was that unlike reasoning, planning, tool-calling, etc., none of it shipped natively with LangGraph, CrewAI, or Google ADK.

So teams started building an owned version that required engineering the entire interface layer from scratch.

Most teams, however, just settled for shipping the agent as a backend API in a chat box since rendering the UI is only one piece of it.

To actually make it work, the interface layer also needed real-time streaming, state kept in sync between agent and UI, conversations that persist across sessions, and reconnection when a user refreshes mid-run.

@CopilotKit is now the only open-source framework that actually lets you build your own full-stack Claude-like apps.

It decouples the agent from the interface, talking over AG-UI (an open protocol for agent-to-user communication).

Being a standard protocol, the frontend never needs to know whether it is talking to a LangGraph or a CrewAI agent. You can change the backend anytime and the UI will never notice.

In practice, CopilotKit's interface layer gives several pre-implemented React building blocks that wire the agent directly into the app, like:

- generative UI, so the agent renders real components instead of text
- chat windows, sidebars, and popups, or a fully headless setup
- shared state, so the agent and app stay in sync
- human-in-the-loop approvals, where the agent waits before acting
- persistent threads that store the whole session, including the agent-user interactions and generated UI, not just text

And because that full history is captured, those interactions can feed a self-learning layer that also improves the agent from real usage over time.

The interface layer that Anthropic spent years engineering in-house is now literally available to any developer/team.

CopilotKit is open-source with 30k+ GitHub stars, and AG-UI, the protocol underneath, is already supported across every major agent framework: LangGraph, CrewAI, Mastra, Google ADK, and more.

CopilotKit GitHub repo → https://t.co/wkQ1taF0rM

(don't forget to star it ⭐ )

If you want to go deeper, I found a detailed breakdown by Shubham Saboo recently on the three Generative UI patterns, with implementation.

Read it below.

[Avi Chawla @_avichawla]

Anthropic's in trouble, again!

They spent years building what's now fully open-source.

What made Claude feel different from a normal app is that the agent could act inside the interface instead of only talking in a chat box.

For instance, Claude Artifacts let an agent render real UI, charts, dashboards, and interactive components that assemble live inside the response.

Every major AI product tried to replicate it.

But the problem was that unlike reasoning, planning, tool-calling, etc., none of it shipped natively with LangGraph, CrewAI, or Google ADK.

So teams started building an owned version that required engineering the entire interface layer from scratch.

Most teams, however, just settled for shipping the agent as a backend API in a chat box since rendering the UI is only one piece of it.

To actually make it work, the interface layer also needed real-time streaming, state kept in sync between agent and UI, conversations that persist across sessions, and reconnection when a user refreshes mid-run.

@CopilotKit is now the only open-source framework that actually lets you build your own full-stack Claude-like apps.

It decouples the agent from the interface, talking over AG-UI (an open protocol for agent-to-user communication).

Being a standard protocol, the frontend never needs to know whether it is talking to a LangGraph or a CrewAI agent. You can change the backend anytime and the UI will never notice.

In practice, CopilotKit's interface layer gives several pre-implemented React building blocks that wire the agent directly into the app, like:

- generative UI, so the agent renders real components instead of text
- chat windows, sidebars, and popups, or a fully headless setup
- shared state, so the agent and app stay in sync
- human-in-the-loop approvals, where the agent waits before acting
- persistent threads that store the whole session, including the agent-user interactions and generated UI, not just text

And because that full history is captured, those interactions can feed a self-learning layer that also improves the agent from real usage over time.

The interface layer that Anthropic spent years engineering in-house is now literally available to any developer/team.

CopilotKit is open-source with 30k+ GitHub stars, and AG-UI, the protocol underneath, is already supported across every major agent framework: LangGraph, CrewAI, Mastra, Google ADK, and more.

CopilotKit GitHub repo → https://t.co/wkQ1taF0rM

(don't forget to star it ⭐ )

If you want to go deeper, I found a detailed breakdown by Shubham Saboo recently on the three Generative UI patterns, with implementation.

Read it below.

[Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social @mikehole]

@TheCinesthetic @charlesarthur Damn right!

[Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social @mikehole]

@Aaronontheweb @petabridge Agent framework? Will ag-ui be easy to add?

[The Shovel @TheShovel]

“They Looked Iranian”: Security Footage Shows Trump Negotiating With White House Patio Umbrellas for Six Hours

https://t.co/IBdGhvQ2jd

[Count Binface @CountBinface]

I see Question Time has reached its inevitable endgame. It was bound to happen.

[Rocket @masoladosa]

When you trust Donald Trump and try to cross the Strait of Hormuz

[Have I Got News For You @haveigotnews]

Donald Trump says the bombing campaign in Iran has led to "people walking around with no legs". That's yet to be proven, although Trump has shown evidence that people can walk around with no brain

[Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social @mikehole]

OMFG!!!!!

[Grady Booch @Grady_Booch]

I have discovered a truly marvelous way to make my AI conscious, but I have run out of tokens.

[Harry Eccles @Heccles94]

Tweet of the year?

[Nature Chapter @NatureChapter]

Real-life Ice Age! 🤣

[Max Crypto @MaxCrypto]

Day 59 of 2026:

[Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social @mikehole]

So OpenClaw is cooked.... Literally! 😉

[Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social @mikehole]

@Grady_Booch You could get your AI to print your own for you https://t.co/mLDi98Uehd full disclosure I work as part of the group that Peecho belongs to.

[Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social @mikehole]

JFC GitHub when last week!!!!! 😠😠😠😠

[staysaasy @staysaasy]

I have a very strong belief that a lot of the AI-oriented hyper-FOMO, end-of-the-world campaigning on X stems from three main sources, all of which are gross and unfortunate.

The first is that many people were trained on / trained by crypto influence campaigns to affect a "you're falling behind" induced FOMO as a way to get eyeballs and capture dollars, even if those dollars were a total grift. (most of AI is not a grift, but crypto taught a certain nihilistic amorality that encourages active grifting)

Next - many people in tech are off-beat, somewhat misfits, or were marginalized in high school. There's a lot of tech people on X with big chips on their shoulders. Telling people that they don't get AI and will be relegated to a permanent underclass is emotionally satisfying as a form of (essentially) retaliatory bullying. If you notice, a lot of the X personalities who seem generally well-adjusted seem to have more measured takes on AI.

Finally, the AI wave is centered in the SF bay area, like so many technology trends before it. Something about those redwoods just makes the people want to build startups. The SFBA absolutely loves a cult, and it particularly loves a millenarian-style cult that is based upon the world as we know it ending and ushering in a new, incomprehensible age. AI is perfect, and espousing the most extremist, apocalyptic views of the future is the best way to gather the most extreme cult members. This is really just a cultural affectation, but it means that a certain end-of-days way of communicating that sounds ridiculous in Tokyo / London / NY (or even LA!) plays totally fine in the bay.

All of this combines to make AI discourse completely insufferable.

[Ryan Deiss @ryandeiss]

I understand why people think AI will make workers less valuable.

...but allow me to offer an entrepreneur's perspective.

(As opposed to the douchebag, scarcity-minded media/economist perspective.)

When AI increases productivity, the value of the person using it goes UP…not down.

If I have an ad campaign delivering a 300% return on ad spend, I don't think "How can I spend less?" I think "How can I spend MORE?"

The same is true for human capital.

Entrepreneurs get this because we're abundance-minded.

We create something from nothing and call it Tuesday.

Economists don't get this because they're scarcity-minded. They literally believe all resources are finite, so a new technology must replace an older one.

It's never worked that way.

The pessimists look smart in the short term, but it's the optimists who get rich in the long term.

Don't listen to the doomsdayers.

Treat AI like the productivity gift that it is, and double down on the PEOPLE who are willing and able to use it.

[Mike Hole 🏴󠁧󠁢󠁷󠁬󠁳󠁿 @mike-hole.bsky.social @mikehole]

@Grady_Booch It's the replies from people who obviously don't know your background etc. they make me smile the most. I just wonder how they saw your tweet in the 1st place.