TheEngineer

AI Systems Architect & technical leader. Building embedded AI, edge computing & critical infrastructure. Do more with less. #AI #EmbeddedSystems

[TheEngineer @TheRevEngineer1]

The Artificial Organism asks a different question about artificial general intelligence. Instead of asking how large or capable a model must become, I ask what kind of system allows intelligence to act persistently, recover from failure, learn from experience, and remain accountable for what it changes.

Using biomimicry as a functional guide rather than a claim that machines are alive or conscious, the book separates responsibilities that are often collapsed inside modern AI systems. Cognition can interpret uncertainty and propose action, while other mechanisms preserve authority, state, evidence, resources, memory, and recovery. A model may know how to perform an action without being permitted to perform it. An attempted action may not have a known outcome. A learned capability should not automatically rewrite the rules that govern its use.

These distinctions become concrete through a carried maintenance case in which an apparently simple request encounters ambiguous intent, refusal, competing conditions, a lost acknowledgment, uncertain completion, compensation, model replacement, and learning. The case exposes a central problem for persistent agency: the system must preserve what actually happened even when its cognitive components change or do not know the answer.

The book also examines small specialized models, artificial nervous-system analogies, memory, identity, homeostasis, embodiment, multi-model systems, and the limits of language as the organizing substrate of intelligence. Its final chapters compare the proposed responsibilities against simpler tool-using agent architectures and make clear what evidence would be required for the architecture itself to succeed or fail.

The Artificial Organism is not an announcement that AGI has been achieved. It is an engineering argument for how increasingly capable intelligence might become a coherent acting system, and a framework for testing whether that organization is actually necessary

https://t.co/pjHB4VazOz

#agi #si #AIAgents #ai

[TheEngineer @TheRevEngineer1]

@MatthewChang Can I bring you my tech and maybe land a job sweeping floors or babysitting the robots?

[TheEngineer @TheRevEngineer1]

@CragheadJace Half of what I’ve seen in the way of mess-ups are people expecting absolutes from a probabilistic system, and then trying to make the toaster responsible for burned toast…

[TheEngineer @TheRevEngineer1]

The Artificial Organism asks a different question about artificial general intelligence. Instead of asking how large or capable a model must become, I ask what kind of system allows intelligence to act persistently, recover from failure, learn from experience, and remain accountable for what it changes.

Using biomimicry as a functional guide rather than a claim that machines are alive or conscious, the book separates responsibilities that are often collapsed inside modern AI systems. Cognition can interpret uncertainty and propose action, while other mechanisms preserve authority, state, evidence, resources, memory, and recovery. A model may know how to perform an action without being permitted to perform it. An attempted action may not have a known outcome. A learned capability should not automatically rewrite the rules that govern its use.

These distinctions become concrete through a carried maintenance case in which an apparently simple request encounters ambiguous intent, refusal, competing conditions, a lost acknowledgment, uncertain completion, compensation, model replacement, and learning. The case exposes a central problem for persistent agency: the system must preserve what actually happened even when its cognitive components change or do not know the answer.

The book also examines small specialized models, artificial nervous-system analogies, memory, identity, homeostasis, embodiment, multi-model systems, and the limits of language as the organizing substrate of intelligence. Its final chapters compare the proposed responsibilities against simpler tool-using agent architectures and make clear what evidence would be required for the architecture itself to succeed or fail.

The Artificial Organism is not an announcement that AGI has been achieved. It is an engineering argument for how increasingly capable intelligence might become a coherent acting system, and a framework for testing whether that organization is actually necessary

https://t.co/pjHB4VazOz

#agi #si #AIAgents #ai

[TheEngineer @TheRevEngineer1]

And the paperback is approved for print!

https://t.co/xYu4WEPTiR

[TheEngineer @TheRevEngineer1]

Maybe we only think it’s complex, because we are defining the problem incorrectly?

[Fugitive Caesar @ThomBrady5]

How long did it take for civilization to recover from the Bronze Age Collapse? 300-400 years. Okay that's how I think about DEI affirmative action. The Woke era (2014-2024) completely burned American elite institutions to the ground. The smart young white men who should've been funneled into career pipelines were locked out and mostly never got started. You have 130+ IQ genius autodidacts on FrogTwitter working as janitors, waiting tables, digging ditches, and driving Ubers because they couldn't get an interview after 800 job applications. There is a total lack of seriousness from anyone in power in how EXTREME and RADICAL we are going to need to be in order to "remedy" this crisis of talent misallocation. Zero urgency from Republican leadership.

[TheEngineer @TheRevEngineer1]

That is @bot bot and @antigravity working together via an @esp32com using the protocol from the book. Still building out the api

[TheEngineer @TheRevEngineer1]

@Ioty_protocol We wouldn’t define fungi as having a body, but it performs necessary work within its ecosystem, and far more efficiently than “bodied” actors.

[TheEngineer @TheRevEngineer1]

@Ioty_protocol I think some of the popular definitions have obfuscated much of what “general intelligence” is. In this case, the “body” is the lowest level hardware required to perform necessary/useful work.

[TheEngineer @TheRevEngineer1]

@mktpavlenko Also, “I don’t know” doesn’t offer a direction. If you expand failure into concrete directions, it allows internal clarification to answer “why” long before it ever returns to the initiating agent. If why is able to be resolved internally, why not allow that?

[TheEngineer @TheRevEngineer1]

@mktpavlenko It depends on how you design your system to handle that kind of scenario. I’ve tested out rerouting that back as incomplete information requiring clarification, which simply stops it until the correct information is there

[TheEngineer @TheRevEngineer1]

@ltgiv Are you just running an LLM, then, or using them for training off the data you generate from the farm?

[Only Form Holds @OnlyFormHolds]

Is justice a human construct or an objective standard, dare I say divine? The answer to this question seems critical to how we respond to situations where justice isn't being served. If it's a human construct, then I don't see what determines what's just beyond will & power.

[TheEngineer @TheRevEngineer1]

The EA crowd already stepped in this mess, first.

Also, the Pope is a formally trained mathematician, so he probably is uniquely qualified to weigh in on this topic…

[Whole Mars Catalog @wholemars]

let’s not mix religion and AI. I don’t know why the Pope needs to have an opinion on what’s going on with AI.

He is an expert in Catholicism. Not technology.

[TheEngineer @TheRevEngineer1]

@untappedgrowth @blakealbers @FromTheFarmUSA @AJRichards I meant Gary Hendrix wasn’t on here…

[Only Form Holds @OnlyFormHolds]

@RealistGuild To learn more about the Realist Guild, a guide is here. Please subscribe & send it to any other currently atomized realist you might know. Thank you.

https://t.co/wyZaEOxjAd

[TheEngineer @TheRevEngineer1]

@untappedgrowth @blakealbers @FromTheFarmUSA @AJRichards I don’t think he’s on X, but I’ve talked with Gary Hendrix at Butchermaster (Northern TX?), he may be someone to talk to as well

[TheEngineer @TheRevEngineer1]

If you’re anyone dissatisfied for the current way the world is being explained on your behalf, and are looking for others who are equally dissatisfied, check out @RealistGuild and give @OnlyFormHolds a follow

[The Realist Guild @RealistGuild]

We are living in challenging times. Those who have put what is true & good at the top of the hierarchy of goods will know that, for as long as most of us can remember, the world around us has been in steady decline.

This is the institutional response.

https://t.co/Y5WXkGXD0I

[The Realist Guild @RealistGuild]

What @InfniteResrvoir may be uncovering is incredibly exciting.

Now it needs rigorous mathematical testing.

But will a scientific establishment that claims to exclude metaphysics recognise the metaphysical assumptions it has already built in?

Full video in the comments.

[TheEngineer @TheRevEngineer1]

@ltgiv Local will make more sense as time goes on… you’re working in agriculture and tech, so you understand probably more than others