> It is not possible one of their extraordinarily high paid engineers did not know how to deploy an airgapped environment for the models to run in.
It is, in fact, possible.
Do you not remember 2022-2023 when the air was singing with the sounds of young engineers in San Francisco elated they were creating an AI god?
And now a couple years later we're supposed to pretend they're competent?
Sorry. I've been in the tech industry a long time. There's a lot of incompetence. The margins provided by software allow a level of endemic incompetence that would simply not be possible in other industries.
I suspect bailouts will be off the table. Simply because the people who support OpenAI and the tech industry are on the same side of the political aisle as the people protesting data centers.
"Big Tech" as a pejorative started on the Right. That never changed.
Remember the hubbub about companies having "free speech" while they deplatformed and censored critics? That never sat well with a certain group of people. And they're the kind of people with long memories, that might have guests entertained as a visible sign of submission, and not as friends.
The more interesting revelation will be if certain persons can psychologically come to admit the people they are against are politically aligned with them, and their opponents are more intelligent than they are willing to let on.
Agreed. Specs are more important than ever. Constraints: types, contracts, protocols, schemas, APIs and such are what ground the boundaries of systems.
Without them, code drift is a guarantee.
Does anyone trust that a set of human sentences typed into a box is more reliable and precise than objectively exact descriptions of data?
This is not to diminish the utility of these code factories. But surely, if anything provides resistance to drift, it's codebase consistency, standards, and things which must objectively be true for the software to function.
The entirety of "agents" and tool calls is a process of combining LLMs to get better results. Is it the same LLM in many cases? Yes. But it doesn't have to be.
It's the natural move that happened after people realized you couldn't throw away half a century of AI research.
Most of these focus on costs. But it is simply the case that the one-shot output did not scale for harder problems on workflows.
Would it be the ones that tried to backdoor SSL and encryption in the 90s?
Would it be the ones who argued for censoring speech on social platforms?
Do those individuals in both cases share the same political ideology as the Bay Area?
I'll provide a counter thought experiment. What if, a certain politician whose campaign was derailed by people who censored him and kicked him off every tech platform simultaneously... thought it would be very funny to "commoditize his complement"?
"But what about criminals?! What about terrorists?! Won't someone please think of the children?!!!"
Every argument against open source AI is moot. Outside of totalitarianism, you cannot control people building algorithms with math and software.
And even that won't work in the long run. People are not going to stop their AI printing presses because the Church of Venture Capital needs regulatory scarcity to command valuations.
They don't. "Emergence" is some ridiculous hype that a bunch of kids with money peddled to get valuations.
And yes, if "emergence" and the "there's something else going in there!" ridiculousness was true, there would be some other architecture which can do more than its symbol prediction.
Everyone can see the code. Everyone can see the math. Everyone can see the failures of the claims. And there's now about 5 years of everyone watching the lies come undone one-by-one.
But much like Ben's point that commoditization is a relatively novel concept to many in tech, it's not the consumer AI applications at risk of commoditization. They have distribution there.
It's the literally millions of engineers who are updating codebases with tools replacing workers partially or wholly. It's the supply-side where there's compression, and no need for distribution.
I would argue, given the enormity of the existing SaaS stack and how it integrates with the human machinery of personnel, that's where volume is. And that is clearly cheaper and a home run.
Commoditizing a ~$100B AI consumer market is no small feat. Commoditizing 20% of the $500B SaaS market, to say nothing of the underlying systems in the who-knows-how-many trillions "Big Tech" market (you're obligated to say that like the Kool Aid man), is shocking.
" - The highest tier Chinese models are not more economical than US frontier models. Try GLM 5.2 and see how much it costs to do real work. I did, and it was more expensive than GPT 5.6."
This is a flatly false statement for most things powering backend applications. The AI consumer "doing real work" model, either for analysis, chat, or coding could well be more cost effective with closed frontier models.
But most of these internal glue business SaaS applications where engineers are integrating are not those tasks. It is those tasks which 1) drive immense amount of domain-specific data into the platform over time, and 2) are most encouraging of driving open model independence with no vendor lock-in.
Anyone on this site who has actually used ML models (more accurate in many cases) knows there's a lot of kludge that simply does not need a 5 minute agentic feedback loop to solve the problem. And they were solvable a year ago with lower class models. The token economics are exceptional and the anecdotes of a16z saying 80% of startups are productionizing open models is only surprising to people who think running your company on OracleDB in 2026 is a sound engineering decision.
With what accuracy? And with how many intermediate tokens?
We can replace multiplication with any class of problems which should go from 0->100% solution almost immediately if there was actually a concept learned.
There is not. Because they are plain ol' fits. And there are no "emergent" features that pop out without having a sufficient set, where "sufficient" is absolutely gigantic and equivalent to memorizing enough of the space to compress the problem. LLMs are Rain Man.
They interpolate within a known distribution. Search allows places outside of distribution to be explored.
This paper should be required reading [1]. You can explore the curves yourself. You can see exactly what it's doing. And you also have this nagging thing—which you know and I know—that all these models converge and do not diverge upwards. An "emergent" "hyperintelligence"—a characteristic that could be found if something was actually learned and combined with a new concept—would not have this problem.
Exponentials on exponentials added to compute and data and the problem classes still sit at not great places, and require agents, feedback loops, and trial and error to solve. The models are the problem, but more importantly, the people selling things these models could never do are the problem.
Edit: It should be mentioned, if there's some scary neural architecture that's super-de-duper and doing something beyond the very obvious next string prediction that LLMs clearly do, it can't do what absolutely ancient ML models could do; a network to multiply two floating point numbers should pop out somewhere without symbolic computation, no?
It does not. There's no magic other than the run-of-the-mill SV fake it til you make it magic. And that magic has failed.
I am not confused. Because herein these forums, I predicted everything that was going to happen years ago.
And the "insane inefficiency" of deep learning is fully to be expected from how it works. As well, there are provably no—literally no—emergent properties in these models. The choice of metric was a convenient, sloppy, and embarrassing fault of the field. It should be discredited; the field should be embarrassed; expectations on messaging should have changed; and it did not.
Why? Because the industry is full of charlatans, and this is a highly profitable enterprise telling people that this would lead to AGI.
Multiply two floating point numbers without a tool call. Still can't, because it's a curve fit.
So, in summary, nothing you just said is relevant. There are no emergent capabilities, simply 1) search, + 2) the original set of learned feature vectors from throwing tons of data at this.
When I see Chinese (open source) emerging AI dominance...
what I see is a chronically corrupt set of Bay Area companies who stole open source software, pretended there was an AI god around the corner, and wanted to be regulated so the federal government could backstop their failed gambit.
I think this is actually a perfect case of an impressive thing being done exactly by mimicry.
Why do LLMs still have trouble on floating point math without forking out to a tool, but they can perform symbolic manipulation just fine?
Because symbolic manipulation is just rote work and textual stepping through symbols. A side poster commented on the number of prior attempts on this problem which were close, but not quite.
Starting from a known "close" solution (which this did), and using exploration to search around the space is exactly something an LLM would and could be good at (clearly).
The "transformer LLMs are next-token predictors with some in-GPU processing of bounded complexity with respect to token count" remains undefeated. Both because that is mathematically what they are, and also because we don't have counterexamples to that effect that don't require some higher-order tooling wrapping the systems.
What discipline, outside academia, could position "nothing is neutral" as insightful and helpful commentary?
So is doing nothing, consuming resources and goods created by others, and living in the default state of humanity: brutality, poverty, and early death.
If it sounds if I'm being glib in this statement, I am. This article is an unreasonable amount of words which eventually boil down to "we have choices in what we do": an exactly equivalent statement to "X is just a tool, it depends on how we use it." Which, of course, is the thesis he intended to rail against.
Babies sleep better when their head is near their mother's heart. This seems to be the obvious reason for the left-handed cradling bias [1].
If a baby sleeps better, it cries less. If it cries less, it attracts fewer predators and helps both parents sleep better at night and have more energy. It also allows the mother to do things with her dominant hand if she is right-handed.
Given the left-handed cradling bias exists even with left-handers, it means there is something specific with left-handedness and infant rearing. A baby in the left hand and a tool (or weapon) in the right is biologically efficient.
Most studies take this from the perspective of evolutionary advantage of the individual. They should take it from the perspective of evolutionary advantage of the family, without which the baby does not survive.
If the bias confers evolutionary advantages, that is also important for the longer childhood humans have compared to primates, which supports our larger brains. Any differential here would have a feedback effect.
Wouldn't it be interesting if a key reason humans are the way we are is a mother's love ♥?
There is also a bias for how babies are held [1]. It holds even with left-handers. Holding a baby's head near the mother's heart helps the baby get to sleep. Which means the baby doesn't cry (and attract predators) and also gives the parents more time to sleep at night.
It also allows right-handed mothers to do something with their dominant hand while cradling the baby in that position.
It is, in fact, possible.
Do you not remember 2022-2023 when the air was singing with the sounds of young engineers in San Francisco elated they were creating an AI god?
And now a couple years later we're supposed to pretend they're competent?
Sorry. I've been in the tech industry a long time. There's a lot of incompetence. The margins provided by software allow a level of endemic incompetence that would simply not be possible in other industries.