Humainary

The rapid convergence of AI SRE products in the observability sector reveals a long-standing flaw: platforms never developed operational intelligence or models of the systems they tracked. Instead, they served as data stores that required people to assemble meaning. lnkd.in/p/eEgXXquX

Google created error budgets as a negotiation tool to balance release speed between developers and operators, not as a comprehensive model of distributed system health. #SRE #DevOps #SLM #SREBook lnkd.in/p/eDfFT_uN

Multi-agent coding systems excel at fulfilling localized tasks and passing test suites, but they risk losing “the program”—the shared, evolving architectural conception that keeps disparate parts belonging to one cohesive system. Agentic coding: death by a thousand plans lnkd.in/p/egj8FNQw

Almost every day, another AI SRE product appears. The demonstrations all follow one script. A chat box takes a question about an incident, a few tool calls run, and seconds later a nicely formatted response lists probable causes and suggested actions. #observability #AISRE #SRE lnkd.in/p/ejDg5fxV

AI lowers software costs and turns features into commodities, ending differentiation through engineering alone. To compete, firms must shift from replicable tactics to strategic “big bets” guided by feedback loops and hard to imitate. lnkd.in/p/eBX-H3yu

Simply centralizing raw telemetry and applying AI does not yield a true “AI nervous system” for production infrastructure. Rather than providing internal regulation and control, it functions as an external diagnostic observer. #Observability #SRE #AISRE #AIOps #DevOps lnkd.in/p/ebTUiwgv

Observability is defined not by telemetry types but by their interpretive role. True observability shifts from raw records requiring human analysis to system-generated determinations that contextualize data and communicate operational state and trajectory at the source. lnkd.in/p/ecYtS_kq

The modern observability industry has fundamentally misunderstood the concept of observability by disconnecting it from its counterpart in control theory: controllability. Observability has become an industry organized around recording events rather than guiding system behavior. lnkd.in/p/eG459pkc

Sampling in observability creates gaps that erase meaning and prevent accurate state reconciliation...This synthetic silence makes intentional deletion indistinguishable from failure, structurally undermining continuity and subject-scale evidence. lnkd.in/p/eYv3jnUk

Traditional observability is reaching a crisis point because collecting, shipping, and storing raw telemetry scales exponentially with system complexity, making standard observability models economically, computationally, and operationally unsustainable. lnkd.in/p/ev4Wfe_f

We have experienced decades of instrumentation. Decades of metrics, logs, traces, profilers, collectors, storage systems, dashboards, correlation engines, anomaly detection, and now AI. The machinery has changed enormously. Yet we still seem stuck below observability www.linkedin.com/posts/hum-ai...

"I’ve just picked up Antonio Damasio’s Natural Intelligence - the Logic of Consciousness, and it lands remarkably close to a question I’ve been pursuing in my own work: What turns sensing into intelligence?" #Intelligence #SituationalIntelligence #Observability lnkd.in/p/eJF36njh

AI agents are semantically inefficient because they constantly serialize internal states into natural language prose only to pay probabilistic models to re-interpret them. To achieve scalable, intelligent behavior, AI must transition from linguistic systems to semiotic agents.. lnkd.in/p/eWRvVeqh

Modern observability has become good at answering questions about data. This creates a strange kind of progress: more searchable evidence and smoother explanations, yet no lasting record of what the system is, what it should be doing, or what its current state implies. lnkd.in/p/eTNw4wXM

They raised the creature. They fed it logs, traces, metrics, alerts, runbooks, dashboards, catalogs, and queries. Then they lowered in the AI brain and announced a new age of operational intelligence. #Observability #DataDog #Dash0 #AISRE lnkd.in/p/epgpUXGj

First came the promise: preserve every event, attribute, dimension, trace, and clue for future use. Then the infrastructure: ingest, retain, index, query, correlate, and search—repeatedly—with AI SRE agents. #Observability #AISRE #AIOps #OpenTelemetry lnkd.in/p/eNHSecQE

Observability is a property of a system: it means a system’s internal state can be inferred from its outputs. OTel transmits data records, but stops there. The vital work of inference, situational intelligence, is left unaddressed. humainary.io/notes/what-b...

This new post, titled “The House Behind the Search Box,” delves into the implications of telemetry hoarding, driven by the pursuit of observability, on an organization’s ability to make informed decisions, prioritize tasks, and continuously learn. www.linkedin.com/posts/hum-ai...

Wide Events: How Observability Turned Uncertainty Into a Business Model humainary.io/notes/wide-e... The observability industry has exploited the concept of “unknown unknowns” to turn boundless data collection into a lucrative business model, shifting away from sound engineering principles

Observability: Searchable Ignorance humainary.io/notes/observ... Observability vendors have reduced observability to the ability to ask questions of data. The system itself stays outside the frame. The

Telemetry is an unsuitable base for projection—projection in the engineering sense, which allows us to ask where a system might go, what follows if a trend persists, or what occurs if I act now rather than later. Those questions sit at the heart of every control system. humainary.io/notes/why-ob...

Observability: The Interpretive Ladder humainary.io/notes/observ...

Presented as a didactic dialogue between a master and an apprentice, this piece explores the core philosophy of software system observability. humainary.io/notes/master...

The #observability #sre industry must acknowledge its unsustainable approach. Complex, unpredictable, and unreliable systems require a deeper understanding of their operations. However, the industry’s solution is to collect more data, which is fundamentally flawed. humainary.io/notes/observ...

It exposes the absurdity of the "log everything and figure it out later" mentality that is promoted by the likes of @datadoghq.com , @dynatrace.com , and @dash0.com. #observability #opentelemetry #tracing #sre #devops What Blind Systems Leave Behind humainary.io/notes/what-b...

Observability is a property of a system: it means a system’s internal state can be inferred from its outputs. @opentelemetry.io transmits data records, but stops there. The vital work of inference, situational intelligence, is left unaddressed. humainary.io/notes/opente...

This booklet offers a cybernetic and biosemiotic perspective on system sensing. It introduces a framework that represents a fascinating shift away from traditional, heavy event-sourcing toward a more holistic, biologically-inspired model of complex sociotechnical systems humainary.io/booklets/the...

Observability vendors have reduced observability to the ability to ask questions of data. They collect telemetry, store it, index it, visualize it, correlate it, and now verbalize it through AI agents. #observability #sre #devops #opentelemetry humainary.io/notes/observ...

The latest edition of the O'Reilly Book, Observability Engineering, is an astonishing +600 pages long! This single number encapsulates the fundamental issues with Observability and @opentelemetry.io. #observability #books #o11y #monitoring #sre #devops www.linkedin.com/feed/update/...

Datadog’s Bits AI and Dash0’s Agent0, despite their polished appearance, primarily automate evidence retrieval, telemetry navigation, threshold-based insights, and natural-language summarization. They lack a genuine “system model” or real-time situational intelligence. humainary.io/notes/puttin...

This note emphasizes that real operational resilience hinges on the “thinking arrow” within us. This internal process transforms raw actions (like incidents, memory traces, and data) into useful knowledge like models and understanding. #SRE #AIOps #Dash0 #Observability humainary.io/notes/the-th...

This booklet is for anyone who has come out of an incident wondering what was missed, or realized too late that they and a teammate were responding to different situations. It trains situational awareness as a disciplined six-station loop. #sre #devops #observability humainary.io/booklets/the...

#OpenTelemetry is architecturally omnivorous, being a centrifugal system. Humainary’s #Substrates is architecturally selective, being a centripetal system. Centrifugal architectures 
 produce reports. Centripetal architectures 
produce awareness. #Observability humainary.io/booklets/tra...

A new booklet delves into the underlying assumptions of OpenTelemetry and Substrates regarding the nature of computational systems. It explores their ontological commitments, epistemological stances, causal models, and theories of attention. #observability @dash0.com humainary.io/booklets/tra...

Language has become the cheapest material. The cost of producing a sentence is now zero (ignoring token costs), so sentences carry no information per unit. The entire medium is debased. Observability is no different.

Systems do not fail
because they lack data. They fail because they cannot
constitute context
nor recognize situation. #observability #sre #devops humainary.io/booklets/sit...

"A book that changes how you look at an on-call dashboard, a post-mortem, a product spec, or an organizational decision. Once you internalize the distinction, you start seeing the same failure mode everywhere: teams that have perfect state but no shared situation..." humainary.io/booklets/sit...

Data is merely recorded difference. Context is the structure that turns difference into meaning. Situation is the living, present moment when that meaning is activated. humainary.io/booklets/sit...

This distillation of systems theory and observability reveals the deeper architectural truths of meaning generation, eliminating the noise of conventional technical discourse that focuses solely on data collection and organization. #observability #systemsthinking #sre humainary.io/booklets/sit...

The observability firefighting market grows in direct proportion to the volume of code entering production. Coding agents are an accelerant for observability vendors. They will generate the very chaos their products are built to manage. humainary.io/why-ive-stop...

Dash0 announced a $110M Series B at a $1B valuation. Its mission: evolve from OpenTelemetry-native observability into “the autonomous nervous system for production.” The language borrows from cybernetics and cognitive science but has very little of the substance. humainary.io/why-ive-stop...

This is what Palantir thinks is situational intelligence. Is this any different than what the typical #observability #o11y vendor sells as "operational intelligence"? Looks all the same?

"The organism that compresses outperforms the organism that deliberates." #observability #situationalintelligence humainary.io/booklets/the...

"Significance is what emerges when an agent under constraint must decide what matters for action right now." #observability #situationalintelligence humainary.io/booklets/the...

The ladder to situational awareness doesn’t exist, because the ground beneath it is always moving. No matter how much you educate teams about awareness or operational intelligence, they cannot reach a level where it is meaningful. #observability #o11y /cc @dash0.com humainary.io/why-ive-stop...

This specification defines structural primitives, behavioral contracts, and lifecycle semantics for computational networks in which typed values flow through governed topologies with precise ordering guarantees. Independent of any programming language, runtime, .... #o11y github.com/humainary-io...

How Palantir, not Anthropic, led US bombs to kill 175 civilians, mostly 7-12yo girls, in a school in Minab, Iran on Feb 28. How embedding automation - not LLMs in this case, just image processing and sensor fusion - in military workflows mutes human judgment, causes errors, & diverts accountability.

Good piece. Also echoes the terrific book "The AI Mirror" by @shannonvallor.bsky.social