Rod Johnson

Building the future of agent frameworks at Embabel. Creator of Spring. Developer, Entrepreneur, Investor, Author.

[Rod Johnson @springrod]

New blog on why a Context Layer is critical AI infrastructure, and announcing Embabel Worlds: an evolving knowledge graph that understands your business.

https://t.co/P3x36TyQ99

[Rod Johnson @springrod]

@paul_hammant London ~2002…quite a place to be

[Rod Johnson @springrod]

@alina_yurenko Not in the US :-)

[Rod Johnson @springrod]

@jtdavies Yes. Of course there's value in training and some onboarding help if customers want it.

But embedding people on an ongoing basis--and needing thousands of them to embed--indicates something wrong.

[Rod Johnson @springrod]

The present industry fascination with FDEs is strange.

The consultant->FDE rebranding was brilliant from a marketing perspective. Less so for customers or the wider industry.

Are we really back in the 1990s?

New blog:

https://t.co/utTemBaczo

[Rod Johnson @springrod]

@sampullara Still consultants. Better to solve problems with product if possible.

[Rod Johnson @springrod]

@sampullara This is a core feature of Embabel Worlds. Users create "worlds"--their custom knowledge graph--and then can use coding agents to build and evolve apps over them. Skills ensure that apps are written the right way; they're also validated before deploy. And they are instrumented etc

[Rod Johnson @springrod]

@sampullara Agree. And there are existing standards to build on.

The tricky bit is getting the right information to the LLM via progressive disclosure, validation, security, governance and sandboxing.

We selectively present TypeScript interfaces and schema definitions (for Cypher queries).

[Rod Johnson @springrod]

Code mode is the future of tool calling, allowing LLMs to write programs vs call individual tools.

The challenge is to make it safe, through curation, approvals and sandboxing.

One role of a context layer is to provide a surface LLMs can code against.

https://t.co/P3x36TyQ99

[Rod Johnson @springrod]

Vibe coding can help users build great things without code.

But it only works over infrastructure that ensures that apps are deployable, secure, validated, safe and efficient.

Embabel Worlds lets users build safely over their knowledge graphs: https://t.co/P3x36TyQ99

[Rod Johnson @springrod]

@yggy_tech Virtual cypher emphasize things that are definite (even if perhaps with multiple entries linked to a canonical node). Eg a CRM, database, responses returned by an internal app.

For possibly conflicting things and subjective claims we use agent memory -- https://t.co/n6qF6YFDwn

[Rod Johnson @springrod]

New blog on why a Context Layer is critical AI infrastructure, and announcing Embabel Worlds: an evolving knowledge graph that understands your business.

https://t.co/P3x36TyQ99

[Rod Johnson @springrod]

@doclove24 Virtual cypher can be extended by Producers in realms. Producers often preserve metadata. Conflicting sources need to be handled by a pre-processing step or modelling, where you define the graph with a canonical entity and link conflicting sources that queries can reason over.

[Rod Johnson @springrod]

@AlBaker_Dev The Realm spec is public at https://t.co/0VKFgCxRwu. Would love to see how an AgentSpace realm would work. Embabel Worlds can also take events.

[Rod Johnson @springrod]

@sampullara It's interesting how so much of the standard doesn't seem to be taken seriously.

But I think MCP makes a fundamental mistake in assuming that there is one API for agents. There already were API standards: agents need to be able to riff on them, case by case.

[Rod Johnson @springrod]

Anthropic would have us believe that AI might soon be able to end humanity. Meanwhile, we wait for Claude Code to provide a sane way to copy to the clipboard.

[Rod Johnson @springrod]

@myfear Why should this be uncomfortable? When good tools can get more out of a lesser model, we should rejoice!

[Rod Johnson @springrod]

@JamesWard It’s so much easier for them than trying to argue. I’m sure they feel appropriately proud of their brilliance.

[Rod Johnson @springrod]

@JamesWard On Twitter, most of the time it just shows the poster doesn’t really understand X

[Rod Johnson @springrod]

@sampullara I like some non-alcoholic beers. Guinness and some hazy. I find the bubblies acceptable (and very useful for an inclusive choice at social events), but sadly the reds undrinkable. Which is a great pity as I love red

[Rod Johnson @springrod]

Jev is yet more proof that no organization should outsource its AI strategy to a frontier LLM vendor or company that wants to sell tokens

LLMs are great, BUT

- Open weights models are capable of more & more tasks
- AI > Gen AI. You don't always need an LLM

@JamesWard @jtdavies