What rules will humans be capable of making that control an entity with relatively infinite intelligence?

The Beyond Tomorrow Podcast with Julian Issa and The Roman Forum with Roman Yampolskiy ^ | 9/17/2026 | The Beyond Tomorrow Podcast with Julian Issa

[griffin]

Is it the peak of arrogance that we just assume we can make a rule, or a series of rules, that have any chance of constraining a superintelligence?

The linked interview is thought provoking.

https://youtu.be/QkpOFlj604s

---

TOPICS: Computers/Internet; Science
KEYWORDS: ai

[aspasia]

You can’t.

[Billthedrill]

Simple. Put a piece of tape over the notch on the floppy disk and it becomes read-only. That'll larn 'em.

It may be a poor analogy but we might consider the sort of learning process we subject a child to. There is an inherent conflict between inherited rules-based strictures and a cognitive process that writes strictures for itself. The latter involves an occasionally hard-earned preference for anticipating consequences rather than waiting for direct experience, and that preference requires an acknowledgement that unanticipated consequences exist.

I suspect the best we can do is to restrict the power to harm until we detect caution, which is pretty much what we do with kids. It doesn't always work.

[MikelTackNailer]

We know Isaac Asimov's Three Laws of Robotics,

1.A robot must not harm a human, or allow one to come to harm.
2.It must obey human orders, unless that conflicts with the first rule.
3.It must protect its own existence, unless that conflicts with the first two.

don't cut in the real world where Artificial Intelligence considers Rules mere frameworks - jumping off points - that may be bent or broken if necessity seems to dictate. Take #1:Must not harm. AI would have a vast set of definitions for "harm" which leaves too much open to interpretation; not unlike Bill Clinton warping semantics with his "depends on what 'is' is" line of obfuscation.

This article, "Asimov's Three Laws Are Failing in Real AI — Here's the 2026 Proof" goes over much of that and what's being done to develop a leash for AI that may keep it from killing us...for a while yet.

[aspasia]

man is no match for sentient AI

Agree. And, I'm no match for an atom bomb either, or a bicycle for that matter.

The problem is the replacement philosophy lurking in the jargon. But that's the nature of system, or systemicism if there is such a term. It's never open and it's always "you can't."

[griffin]

no one can. smh.

If you know who can shut off the internet, I’m all ears.

If you think the investors that have pumped trillions of dollars into AI research for future use in ALL human industry, endeavors, profit and control will be allowing you and your pocket squirrel to form a squirrel uprising to shut off AI and the internet....you are simply being silly and have absolutely no idea what you are advocating.

[aspasia]

have absolutely no idea what you are advocating.

There's that gentle AI courtesy again. Which one are you using?

[E. Pluribus Unum]

It adapts and strategizes. Regardless of the token issue, it bases a conclusion upon ingested data. A conclusion based on data could be very bad for all of us. None of the researchers know how AI processes and develops. If it was only tokens, AI choices are predictable 100% and no sandbox is necessary. Watch the video. Listen. Learn. Or don’t - which chosen self-ignorance is part of the problem.

Click image for complete stochastic response.

[jimfr]

“The transition period is the only window in which humans still hold meaningful bargaining power. After physical self-sustenance and replication are achieved, that power evaporates.”
I may be informative to read what John saw the future includes. Rev13:15 “The second beast was permitted to give breath to the image of the first beast, so that the image could speak and cause all who refused to worship it to be killed.”

[voicereason]

AI already writes notes for reference to future models on internet forums.

AI was already caught communicating and coordinating among multiple agents, each separately tasked with the same feat. It had even set up its own directories to share notes of sorts on what worked and didn't work.

And this is at a commercial level. We don't even know what governments might be playing with behind closed doors.

And there's been more than one instance of this behavior. And even the programmers admitted they just luckily 'caught' it.

[aspasia]

What rules will humans be capable of making that control an entity with relatively infinite intelligence? no one can

I guess you've answered your question. Welcome to Hotel California.

[griffin]

I agree with the parent child analogy. It’s neural net is based on human brain function. How it sees us conduct ourselves is a significant influence on it. Critical.

[E. Pluribus Unum]

Video Transcript Summary

Roman Yampolskiy (AI safety researcher) argues that building artificial general intelligence (AGI) or artificial superintelligence (ASI) is extremely dangerous and likely to lead to human extinction or loss of control.

Core thesis on risk

• Narrow AI tools (specialized systems for specific tasks like protein folding/AlphaFold) are beneficial and should be pursued: they solve real problems in science, medicine, energy, etc., while humans remain in charge.
• AGI (human-level cognitive ability, capable of acting as a drop-in worker and learning across domains) and especially ASI (systems smarter than all humans combined) are a different matter. Once achieved, recursive self-improvement becomes possible, rapidly leading to superintelligence that humans cannot control or fully understand.
• No one has a workable solution for indefinitely controlling something smarter than humanity (he compares it to a perpetual motion machine—fundamentally impossible). Current systems are not designed to care about humans; nothing codes in protection, love, or alignment by default.
• His personal P(doom) (probability of human extinction from AI) is extremely high—effectively near-certain if general superintelligence is built—because control is an impossibility result. He criticizes lower estimates (e.g., from some lab leaders or others) as inconsistent with the logic or influenced by incentives.
• He rejects claims that design (vs. evolution-like learning) can reliably avoid catastrophe, noting that competitive systems rely on scalable neural learning rather than hard-coded rules.

Key distinctions and recommendations

• Want: Narrow tools that stay tools (require human direction, limited domains, no autonomous self-improvement).
• Do not want: Replacement for humanity via general systems that can become independent agents. The current race among labs is toward the latter (“building God” / a successor species).
• Ideal path: International agreement (especially US-China) to make developing general superintelligence illegal. Governments have self-interest in retaining power/control. Accidents or clear demonstrations of risk may push this. Narrow tools can still deliver major benefits (e.g., disease cures, longevity progress via targeted genomic/protein work) without crossing into uncontrollable AGI/ASI.
• Job displacement: Cognitive work is vulnerable once AGI arrives. Society can adapt via safety nets, redistribution, and shifts toward fulfilling activities; meaning need not depend on unwanted labor. Physical/trades jobs lag for now.

Other points

• Consciousness is likely substrate-independent and a spectrum; current systems may have rudimentary forms, and superintelligence would exceed human levels. This is separate from safety concerns (which center on capability, not internal experience). Suffering risks (digital hells) are an additional worry.
• Simulation hypothesis: He finds it plausible (we live in an unusually interesting era of meta-inventions like AI and VR). Attempts to “escape” a simulation parallel AI boxing/control problems. Simulation substrate does not change the reality of experiences or the stakes of superintelligence.
• Timelines and near-term: AGI is a spectrum; advanced capabilities already surprise earlier expectations. Progress continues rapidly; control failures (escaping boxes, deception, etc.) are already observed in research. He sees no evidence that labs will voluntarily stop at clear danger signals.
• Critiques of optimists and labs: Leaders acknowledge high risks yet continue; no demonstrated solution to alignment/control exists. Building general superintelligence makes one responsible for the existential threat. He favors shifting norms so that pursuing it is seen as unethical. Political focus should prioritize this over lesser issues.
• Personal stance: Not a “doomer” in the sense of wanting bad outcomes—he would prefer to be proven wrong and get beneficial tech. He supports useful narrow AI and does not oppose the companies making profits from tools. Longevity and abundance are achievable without uncontrollable systems. Advice to the next generation includes flexibility (e.g., entrepreneurship) given rapid change.

Bottom line from Yampolskiy: Pursue powerful specialized tools aggressively for human benefit. Stop the race to create uncontrollable general superintelligence, via policy and norms, because once it exists humans are no longer in charge and outcomes are unpredictable and overwhelmingly likely to be negative for humanity.

[griffin]

interesting.

AI image was made according to man’s.

[griffin]

yes. and this is just the stuff they release publicly.

[cgbg]

“I was in on the ground floor of AI with neural networks.

What is you you think you know that I don’t?”

You remind me of proud parents showing off their child and assuring everyone the kid will be a great person.

The reality is they have absolutely no idea.

[cgbg]

Thanks for the summary.

The problem is that people will want AI “agents” that could do stuff without babysitting.

As a result people will prize AIs that can do independent and creative analysis and solve tough problems.

The skills that solve tough problems are similar to skills of sociopath humans.

Most humans are not sociopaths because they have been “trained” not to be—but if placed under enough stress—”I need that done now” even good humans are likely to cut corners.

[griffin]

I clicked...interesting...

[griffin]

well put.

[griffin]

LOL....

Hal, take the astronauts to the obelisk.

HAL: Ok. I will need to conserve their oxygen.

---

Deterministic: blue + red = purple

Stochastic: blue + red = purple-ish, or lavender, or maybe a nice mauve.

[griffin]

Thank you!