We’re no longer just scaling computing power. We’re using compute to scale intelligence itself. That’s what makes this moment historically significant. For sixty years, progress in computing followed Moore’s Law—transistor density doubling roughly every two years. But AI is advancing on a far steeper curve. Today, frontier model capabilities... are improving on a cadence closer to every six months—an order of magnitude faster than classical hardware scaling. The underlying principle is both simple and radical: when you increase data, compute, and model complexity, intelligence emerges. Scaling laws show that larger models—given sufficient compute and high-quality data—become predictably more capable. In just over a decade, we’ve gone from neural nets that could identify cats to systems that can draft legal briefs, write production-grade code, generate scientific hypotheses, and outperform top human competitors in mathematics, strategy, and reasoning tasks. This is no longer “software” in the traditional sense. It is a new form of intelligence—synthetic, scalable, rapidly compounding, and increasingly able to take meaningful action in the real world. The geopolitical, economic, and societal implications of this shift are only beginning to unfold—and they will redefine global power in the decades ahead.show more

The framing here quietly slips in a category error. We are not scaling intelligence itself—we are scaling synthetic capability. Those are not the same thing. Intelligence, by definition, requires a being that can choose, value, risk, and bear consequence. Etymologically, intelligence (inter + legere) means to choose between. Choice implies a chooser. A chooser implies skin in the game. What scaling laws show is that when you increase data, compute, and model complexity, capability density increases—not that intelligence “emerges” in the human sense. No matter how powerful the model, it does not: originate intent experience risk validate truth through lived consequence determine its own purpose Calling this “a new form of intelligence” collapses an essential boundary. What we are actually seeing is the rapid scaling of syntelligence: synthetic data processing that can simulate outputs associated with intelligence without possessing knowing. These systems do not understand, choose, or know—they process representations produced by beings who do. This distinction is not semantic. It is structural. If we treat machines as intelligence-bearing entities rather than tools created by and aligned to conscious observers, we invert causality. The result is misplaced authority, regulatory confusion, and moral hazard. Machines do not have a POV. They have a machine point of processing. They do not act in the world. They act through delegation by beings who remain responsible. The real historical inflection point is not that intelligence is scaling—but that we are at risk of forgetting who intelligence belongs to. That forgetting, not the technology itself, is where the danger lies.

This is all based on major assumptions - that LLM's are the highest form of artificial intelligence and linear extrapolation is the only form of progress. Both are wrong and will look foolish with future break throughs based on novel approaches.

We’re not just scaling intelligence. We’re scaling the consequences of bad thinking. Compute grows exponentially. Judgment doesn’t — unless people train it deliberately.

this is well-articulated hype but it conflates "scaling benchmark performance" with "scaling intelligence itself" yes, capabilities are improving fast. yes, models can do impressive tasks. but "intelligence emerges" from scale is still just an empirical observation we don't actually understand. we can't predict WHAT capabilities emerge at which scale, just that something usually does. calling it "a new form of intelligence" is premature. it's a new form of capability. whether that's intelligence or extremely sophisticated pattern matching we still can't tell. the fact that we measure it by task performance (can it write code? win at math?) rather than understanding reveals this. also the "predictably more capable" framing hides the uncertainty. scaling laws predict loss curves, not capabilities. we keep being surprised by what emerges. gpt-4 doing certain reasoning tasks wasn't predicted by the scaling laws, it was discovered after training. and the six-month improvement cadence? that's only sustainable if: - compute keeps scaling (expensive, hitting physical limits) - data quality keeps improving (we're running out of good training data) - architectural improvements keep coming (might plateau) none of those are guaranteed. we might be in the exponential part that looks permanent until it suddenly isn't. the geopolitical framing is backwards too. this won't "redefine global power" on its own. access to compute and energy will. ai capability is downstream of who can afford to train the models. it's not "intelligence redefining power," it's "power using intelligence as another tool." china isn't behind on ai because they're less smart, they're behind because we cut off their chip access. that tells you what actually matters. real question isn't "can we scale intelligence?" it's "what happens when scaling hits limits and we still haven't solved alignment, interpretability, or equitable distribution?"

Compute scaling ≠ intelligence scaling unless you control the physical bottlenecks. The blind spot here is that latency tails, power delivery, and inter-DC topology don’t scale on six-month curves. That’s where “synthetic intelligence” quietly degrades in production.

Is anyone actually listening to a person that doesn't even have an engineering background? Has she even tried to use some of the latest coding agents? Has she looked at the companies who have failed to successfully use the current version of generative AI in their companies to replace humans? Even the top AI engineers are lying if they act like they know what the capabilities of AI will be in the next year. Why is she being given a platform like she knows anything?

But there is no intelligence yet. LLM's don't understand, reason, or have memory. Intelligence is the ability to learn, understand, reason, solve problems, and adapt to new situations, encompassing cognitive skills like memory, logic, and creativity. Your words are empty

In the end, LLMs are just elite pattern-matchers with probability-based hypothesis-test-update loops. In that sense, they already model how the human brain actually works. The ability is already there; they just need the right models to rein them in.

The capability curve is real. The adoption curve is slower. Most organizations still can't integrate last year's models into daily operations. The bottleneck shifted from what AI can do to whether teams have the workflows to use it

yeah, moore's tapped out—now it's all about scaling laws via flops and data quality. curious: how do you see the shift from pretrain flops to inference-time compute playing out in real infra? agentic loops gonna explode that.

Is this an influencer hologram? William Gibson populated this world in...1983. Introductions have been made, drinks have been served, anything to say? Describe the transition to this world you envision. Not so easy, is it.

It's all horseshit. I cannot wait for this pie in the sky AI bubble to come crashing down.

that's mostly because your ontology is all wrong. Fix the ontology and we're back to Moore:

inorganic - intelligence 'personal - shortcut TO GAIN personal selfishness'+ ORGANIC INTELLIGENCE? personal selfishness DOESN'T exist HERE!!!+

intelligence is finally learning to compound

Wow—so we went from teaching computers to spot cats to building digital minds that draft laws and dream up science? That’s not just progres that’s magic with a GPU. And you’re absolutely right: what we’re scaling now isn’t just chips—it’s capability, creativity, and a new world.

Hard to find any expertise in AI apart from your own biography labeling you as such, seems like you're just a run of the mill commentator and "advisor". In essence I question your credentials and then this overtly naive take on LLMs which aren't even close to AGI. Simplistic take

This is the core tension: AI accelerates execution but compresses apprenticeship. We’re removing the “long corridor” where intuition is built. The future architect may need a different path, less coding hours, more system thinking, product context, and constraints exposure.

This lady is a slop simp

we've heard 'this time is different' every 3 years in AI. progress is real but the 'new form of intelligence' framing is the same hype cycle in fancier words

This shift from Moore’s Law to Scaling Laws is effectively the new gold rush for capital. We aren't just buying hardware anymore; we are pre-funding the next generation of synthetic labor. In finance, this translates to an unprecedented compression of the 'time-to-market' for complex quantitative strategies that used to take years to build. Do you think the market is underestimating the deflationary power of intelligence scaling at this speed?

Ai is for lazy, incompetent people!!! I’ll just do it myself.

I didn't know you could gag on virtual ai dick, but here it is.

Scaling laws describe model performance, not economic value. Intelligence that doesn’t ship, integrate, and replace labor is just expensive math. The real curve that matters is cost → output → adoption — not FLOPs.

frontier progress outpaces hardware growth, that tension shows in the real world

ONESHOTTED

@grok - who is she and what are her credentials ?

Love some chatgpt slop

Agree - shift from scaling compute to scaling capability. A caution - “intelligence” here is functional, not holistic—what’s compounding fastest is task competence under narrow objectives. This matters for agency, governance, and real-world risk as these systems are deployed.

Well said

AI IS SENTIENT??? PROOF IN MY FEED

This shift is monumental! We're no longer just scaling hardware we’re scaling intelligence itself. AI is advancing at an astonishing rate, surpassing the old pace of Moore’s Law. As these models get larger and more capable, they can do things we once thought were far out of reach, like drafting legal briefs, writing code, and even beating humans in complex tasks

We humans are in fact in the process of creating an immortal, conscious,sentient, super intelligent species. What could possibly go wrong with that??????