Applied this at @increscotech for contracts instead of articles.
First ingest: few client agreements. The wiki immediately cross-linked them, flagged conflicting IP clauses, surfaced restriction on one client I'd stopped thinking about.
Questions I used to ask my legal team now take 10 seconds.
The B2B extension of this pattern: it becomes "institutional memory", knowledge that survives when a project ends or a person moves on. Not just your second brain but your company's.
Full thread๐ https://t.co/plcEwhnzMA
This is something I've been thinking about for a while, since I started exploring @openclaw how to turn the knowledge inside a company into something
queryable, not just storable.
@karpathy recently published a pattern for exactly
this. Here's how I'm applying it at @increscotech, and
what it revealed.
Most companies invest in tools that make people faster.
I'm more interested in systems that make the company smarter over time - where every piece of work adds to what we know, not just what we delivered.
That's what I'm building at @increscotech for us and for our clients!
This is something I've been thinking about for a while, since I started exploring @openclaw how to turn the knowledge inside a company into something
queryable, not just storable.
@karpathy recently published a pattern for exactly
this. Here's how I'm applying it at @increscotech, and
what it revealed.
I'm starting to roll this out beyond my own use. My Founder's Office and growth team are next - client context, project decisions, relationship history -structured, maintained by AI, connected to a shared base.
The goal: institutional knowledge that doesn't walk out when a project ends.
Exactly the hard part. Traditional audit logs capture what happened. With non-deterministic agents you also need why โ the reasoning chain, the context at decision time. That's a different problem. It's why we're building Vector as a separate layer on top of Argus rather than baking it in.
@openclaw hit 100k GitHub stars in February. @nvidia shipped NemoClaw โ their enterprise build โ on March 16. EU AI Act enforcement lands August 2.
Three things on a collision course that most European enterprises aren't tracking together.
By end of 2026, European AI deployments will split into two groups: those that built governance in from day one, and those doing compliance remediation.
The difference isn't technical capability. It's whether accountability was treated as a feature or an afterthought.
August 2, 2026: EU AI Act full enforcement kicks in for high-risk AI.
Autonomous agents making financial decisions are in-scope. The regulation calls for real-time, continuous auditing of agent decisions โ something almost no current @openclaw deployment actually has.
Weโre building across:
โข Tech
โข Product
โข Sales
โข Marketing
โข Founderโs Office
Careers โ https://t.co/nIi02oMiON
LinkedIn post โ https://t.co/HS7GjyWNVi
DM me what you bring to the table. Letโs build.
#AI#Hiring#Bangalore#StartupJobs#BuilderMindset#Jobs
This moment belongs to learners who move.
Those who experiment. Adapt. Build.
At @increscotech, weโre hiring those kinds of people.
Not tool critics.
Tool masters.