@Tom_MkV

Co-founder @ctxpipe. Technical Product Maker, father of two, husband, and homelab/military history/LoTR nerd.

 Australia
ctxpipe.ai
Joined November 2024
  • Tweets1.2K
  • Following489
  • Followers578
Been pushing hard the last (4 years) two weeks, and if I stop building at 12am I find it super difficult to sleep quickly: music, multiple agents & projects - of course, my brain needs time to unwind. So a 1979 edition of the Hobbit will do tonight!
2
29
Working on the local memory hook design, and the balance between signal and noise is tricky. - Agents working faster, longer, and in (📷 ) parallel - Relying more and more on them to self-solve/source etc - Engineers involved less in low-level work, more on design Cont.
1
19
How does the system self learn? Is agent chatter at lower-levels useful as part of a context layer for future agent consideration? Should it be the episodic memories captured during agent sessions, or be a synthesis bolted onto the PR description (which we now capture), and leave local memories uncommitted?
1
5
I'm leaning towards PR's being the gate, as we follow a strict 'merge = decision/truth' boundary. Keeps the graph (perhaps) cleaner, and add fuel to esoteric markdown files battles on weighting. Would love people's thoughts!
3
An org using ctxpipe.ai - This shows a single engineering org's graph with their: - Linear issues - Pagerduty incidents - 15 repos and PRs from Github - Eng specific Slack captures All connected and easy for agents to reason over to help understand how they build software.
1
1
1
62
Full platform - open source: github.com/ctxpipe-ai/ctxpip…

GitHub - ctxpipe-ai/ctxpipe: Self-learning context infrastructure for AI engineering agents....

Self-learning context infrastructure for AI engineering agents. Connect repos, docs, and tools into one org-scoped knowledge graph and expose it through a single MCP. - ctxpipe-ai/ctxpipe

github.com
1
6
Agents that know how YOUR organisation builds software? Could be the move. github.com/ctxpipe-ai/ctxpip…

GitHub - ctxpipe-ai/ctxpipe: Self-learning context infrastructure for AI engineering agents....

Self-learning context infrastructure for AI engineering agents. Connect repos, docs, and tools into one org-scoped knowledge graph and expose it through a single MCP. - ctxpipe-ai/ctxpipe

github.com
19
Cool idea on how to fund open source. Idealistic or pragmatic? Essential or greedy? seldo.com/posts/nobody-pays-…
18
The team thinks that Opus says no like Sir Humphrey Appleby… youtu.be/zcsCqwVk6ow?si=HiFs…

Obfuscation

Sir Humphrey confuses the PM

youtube.com
28
Ohh man, just burned through $30 Sol credits on a single turn in Cursor, for it to run-over the limit and default back to Grok 4.6 High. Does that accrued context get handed to the new agent? Not encountered this before. It seems like it's lost... oof.
1
4
73
ctxpipe connectors now capture images. Frontier models handle vision well, so when you're throwing around screenshots (that may have annotation), these serve to improve context for your agents. There's more we can do here - perhaps extracting context from the image into text...
1
49
github.com/ctxpipe-ai/ctxpip…

GitHub - ctxpipe-ai/ctxpipe: Self-learning context infrastructure for AI engineering agents....

Self-learning context infrastructure for AI engineering agents. Connect repos, docs, and tools into one org-scoped knowledge graph and expose it through a single MCP. - ctxpipe-ai/ctxpipe

github.com
21
Oh man I need a new pfp... this is 7 years old.
1
1
49
The graph must grow... Notion and Linear now supported - connect them to add more depth and breadth to what your agents understand about your engineering organisation. github.com/ctxpipe-ai/ctxpip… #ai #agent #context #engineering
1
1
4
115
I wonder if the battle between building vs buying in the AI space is due to fear of lock in and perceived eventual migration to the next best thing. We're thinking about how we solve this as ctxpipe.ai - data portability is a principal - along with being open source

ctxpipe | The AI Engineering context layer for agents

The AI Engineering context layer for agents. A pair programmer for your agents that knows your platform and shows them how things are done. Self-learning knowledge graph infrastructure for autonomous...

ctxpipe.ai
1
114
AI context is another battleground in the AI infra wars. Our bet is on intent based layers to eliminate context poisoning and token wate. ctxpipe is for engineering orgs. Not finance, not marketing, not whole-of-company. Just AI engineering initiatives big and small. More on why this is a good call here: ctxpipe.ai/blog/specificity-…

Specificity compounds. Generality dilutes. — General-purpose memory layers promi

General-purpose memory layers promise one graph for everything. For agentic software engineering, that's the wrong trade.

ctxpipe.ai
1
1
33
I've been head buried in building for the last few months. Since coming back and posting, the OF bots are back. Kinda sweet.
1
19
Check us out here: github.com/ctxpipe-ai/ctxpip… #ai #agent #engineering #context #memory

GitHub - ctxpipe-ai/ctxpipe: Self-learning context infrastructure for AI engineering agents....

Self-learning context infrastructure for AI engineering agents. Connect repos, docs, and tools into one org-scoped knowledge graph and expose it through a single MCP. - ctxpipe-ai/ctxpipe

github.com
ctx|@ctxpipe
Jun 11
We're working on making AI agent context simple. Start with what you have and let the system, agents, and harness improve it; so you can get on with building. Remote knowledge system + local memory harness that work hand in hand. We wrote about it here: ctxpipe.ai/blog/agents-md-is…

AGENTS.md is the wrong conversation — A paper dropped this week that test

A paper dropped this week that tested AGENTS.md files across multiple models and real GitHub issues. Context files reduced task success rates and inflated inference costs. The debate is useful — but...

ctxpipe.ai
1
2
53
The ctx| repo is up: github.com/ctxpipe-ai/ctxpip… Open source self-learning context layer for agents. We're working on the managed version now, so if you're after a simple platform to manage your engineering org context, DM me! #ai #agents #context

GitHub - ctxpipe-ai/ctxpipe: Self-learning context infrastructure for AI engineering agents....

Self-learning context infrastructure for AI engineering agents. Connect repos, docs, and tools into one org-scoped knowledge graph and expose it through a single MCP. - ctxpipe-ai/ctxpipe

github.com
1
2
44
RLMs are doing the rounds again. I can’t wait until they’re more ubiquitous and less fringe tech- but context access and portability will remain a gap.
Tom
@Tom_MkV
Apr 1
Wrote an article about RLMs and the problems solved by agents who and run and run and run. - RLM runs RADICALLY better than OOB agent harnesses - Context portability issue still limits value - Stateful agents = $$$$$ ctxpipe.ai/blog/infinite-con… #ai #agents #llm #engineering

Infinite context is theoretically possible. That's just the start.

Recursive Language Models push token-level working memory to millions of tokens — and make it obvious why context window limits are necessary but not sufficient for real engineering organisations.

ctxpipe.ai
1
1
40
Load more