kgcgfva

Pentad.ai -- an operating system for agents... Very real tech challenges.

Well it's BEAM/OTP on Linux presently, but I intend to look at what running a pickled, single-binary on a microkernel might buy me, etc. That said, not using. SeL4 directly.

And, yes, in re: semantic layer, I have't talked bout that part much, yet, but given my experience at Stardog(.com), I've implemented a lakehouse+documents layer that extends convention agentic memory tiers (working, semantic, episodic, procedural) with what I call "institutional" memory, i.e., what an invoice reconciliation agent, say, needs to know from C360 data sources, etc.

One thing that falls out of Model Minimalism is adding native WunderOS model hosting, which I've been working on this week, natively in Zig, NIF'd into BEAM/OTP.

IMO vertical integration in AI infra is underrated; by adding model serving I can exploit a range of optimizations that 'best of breed'/glue code architecture makes harder. This week's example: native semantic entropy implementation -- following Spanda -- in Zig, such that hallucination detection at K=5 (batch size) is 650us per turn, i.e., in the noise.

I've spec'd how to do QLoRA, too, but it's unclear when or if I'll implement it, not least because it's not clear that I should bother given the training data integration issues.

I'm glad someone else is thinking about this stuff!

Good stuff here, much of which resonates with me. FWIW, my notion of an OS for agents isn't as close to the metal yet, ie, my view of agent lifecycle is that they look a lot like WhatsApp or Discord traffic. So, based on that DEEP analysis, I decided to build on BEAM/OTP for the control plane: basically, Elixir for the UX and Gleam for everything else. And the data plane is pure Zig: data sidecar into BEAM/OTP via NIF.

So that gives me a certain freedom for deploy: bare metal, containers, VMs, even K8S. And since there are some tools for pickling all that into a single binary, and running BEAM/OTP on u-kernel sorts of things, I can get all the way to the metal in the way that you are. Whether or when that happens remains to be seen, etc.

Thanks for sharing! See https://pentad.ai/PLRN for more about what I'm up to.

No, if you look carefully these are very different things targeting very different use cases (enterprise-wide agents vs personal agent).

Thanks, softwarewright; would love to hear more about your work. I've been implementing small-model inference in Zug inside WunderOS, but it's not done yet, so nothing public just yet.

A recent paper, Context Compaction Theory (arXiv:2608.01326) puts a floor on how much you can compact an agent's memory. They measured Opus 4.8 against the floor. 4.8 lands on the random-guess line for membership queries. Not very good; but at least it's latent and $-spendy?

I ran the same probes against Wunderblock, which is based on vector-symbolic architecture, and got 84% of the floor with no LLM at all. Bloom filters sit at 55%.

LLMs are overused!

Pentad.ai — operating system for agents which will outnumber human computer users soon. See pentad.ai/platform for tech details.

The problem is cross-organization data sharing in the agentic era: one organization must let another’s agent work over its most sensitive data...memory is a service of the agentic operating system, not a possession of the agent. When the memory of the visit belongs to the host rather than to the visitor, the data room can be rebuilt for agents. The rebuilt room is an agentic data enclave.