There are very few stateful MCP Servers out there, and the standard is moving towards stateless by default.

What is really making MCP stand out is:

- oauth integration

- generalistic IA assistants adoption. If you want to be inside ChatGPT or Claude, you can't provide a CLI.

> What is really making MCP stand out is:

> - oauth integration

I don't see a reason a cli can't provide oauth integration flow. Every single language has an oauth client.

> - generalistic IA assistants adoption. If you want to be inside ChatGPT or Claude, you can't provide a CLI.

This is actually a valid point. I solved it by using a sane agent harness that doesn't have artificial restrictions, but I understand that some people have limited choices there and that MCP provides some benefits there.

Same story as SOAP, even a bad standard is better than no standard at all and every vendor rolling out their own half-baked solution.

Oauth with mcp is more than just traditional oauth. It allows dynamic client registration among other things, so any mcp client can connect to any mcp server without the developers on either side having to issue client ids, secrets, etc. Obviously a cli could use DCR as well, but afaik nobody really does that, and again, your cli doesn't run in claude or chatgpt.

Stateful at the application layer, not the transport layer. There are tons of stateful apps that run on UDP. You can build state on top of stateless comms.

Wow this is the perfect prompt injection scheme

Would have been even more absurd if code AND PRs were all AI generated by different coding agents

Nothing is stopping that from happening tbh

I think you don't need to add the @, just prompt: Figma, etc... And of course check that you are connected to the app in your settings

You need to either @-mention, or click + below. ChatGPT then prompts to connect.

Will be interesting how difficult the submission process will be, and which apps will get featured by chatgpt.

Maybe an ad based system coming soon?

Chatgpt apps are MCP servers with a UI resource (can be a react component or vanilla js) that gets shown in an frame one the tool is called by chatgpt. So you can't just port an existing app, but you can reuse the same backend Api wrapped inside an mcp server, and some of the components that you need to adapt to openai ux requirements. I practice this means developing an app from scratch.

You'd be surprised! In B2C at least, almost every company we talked to is building a ChatGPT App out ouf fear of missing out this agentic wave like so many missed mobile 15 years ago

If you want to quickly get started, we developed a react-based, type-safe framework to build chatgpt apps: https://github.com/alpic-ai/skybridge

Have you tried it and compared to other small models?

I have extensively. Here's an example of Mistral Small Creative reading weather and snow data creating a ski report:

The mountain is wrapped in a fresh, high-density powder blanket after a wild overnight storm that dumped 14" at the summit, 5" mid-mountain, and 3" at the base—with an extra 1-3" of light snow today topping it off. The snow is upside-down (heavier on top), creating a challenging, almost "wet concrete" texture in spots, especially on SE and SW aspects, where the wind scoured and redeposited it unevenly. Thin coverage is the name of the game—expect to bottom out in rocky or icy patches across all aspects, particularly on flatter terrain and wind-exposed slopes.

Skiing today is a mixed bag of adventure and caution: the summit’s NW aspects offer the deepest powder (up to 15"), but the upside-down snowpack means you’ll need to adjust your line to avoid post-holing or sudden transitions to bare ground. Mid-mountain feels like a crusty, wind-packed playground—fast and fun on groomers but sketchy in the trees where the snow is thin and uneven. Down at the base, the 3" of fresh stuff is just enough to soften the edges of icy patches, but don’t expect deep turns—it’s more about playful bumps and quick carves before you hit the harder stuff underneath.

Wind effects are still visible: NW slopes are wind-loaded and dense, while SE faces feel scoured and thin. Freezing levels dropped below base, so lower elevations may have a slushy, refrozen crust by midday. Avalanche risk is moderate—watch for wind slabs on lee aspects and thin, unconsolidated snow over harder layers.

Bottom line? If you’re after deep, untracked turns, head early to the summit’s NW aspects—but expect a workout as you navigate the heavy, upside-down snow. For a more mellow (but still fun) day, stick to groomed runs mid-mountain, where the 5" of fresh stuff softens the ride. Helmets and poles are a must—this snow isn’t forgiving if you take a spill. Conditions will firm up fast as temps rise, so get out there early for the best experience.

------------

vs Gemini Flash 3 doing the same at 4x the cost for the same prompt:

Conditions are currently overcast with light snow falling across all elevations, adding to the 14" of new high-density powder recorded at the summit and 5" at mid-mountain over the last 24 hours. While the base areas are seeing a fresh 3" coating, freezing levels remain low, ensuring the ongoing precipitation stays frozen throughout the day.

Skiing today features a heavy, "upside-down" snowpack that requires strong legs to navigate, especially where SE winds have scoured slopes and loaded the NW aspects. Expect a physically demanding experience as you maneuver through this dense, wind-affected snow, and exercise caution as the coverage remains extremely thin; you will likely feel the firm ground or obstacles beneath your skis in localized spots across the mountain.

I hope MCP will prosper inside this new structure! Block donating Goose is a bit more worrisome - it feels like they are throwing it away into the graveyard.