@DaseindustriesLtd comments on "Culture War Roundup for the week of August 17, 2026

Culture War Roundup for the week of August 17, 2026

This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.

Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.

We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:

Shaming.
Attempting to 'build consensus' or enforce ideological conformity.
Making sweeping generalizations to vilify a group you dislike.
Recruiting for a cause.
Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.

In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:

Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.
Be as precise and charitable as you can. Don't paraphrase unflatteringly.
Don't imply that someone said something they did not say, even if you think it follows from what they said.
Write like everyone is reading and you want them to be included in the discussion.

On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.

Jump in the discussion.

No email address required.

DaseindustriesLtd late version of a small language model 5hr ago

I don't know why you assume "particularly not you" would have predicted GPT-3. GPT-3 follows from GPT-2 follows from Vaswani et al. Why do you think plenty of intelligent people, including Dario Amodei and Ilya Sutskever, got so agitated when that paper came out and started scrambling for commercialization? Do you believe they were that into machine translation as a business? Of course folks like Hinton, Sutskever's teacher, were running multi-decade research programs premised on connectionism being enough for intelligence in general, not just for some particular capability demo. I was pretty AGI-pilled all of my conscious life, admittedly mostly for shallow sci-fi reasons, and got convinced of inevitable singularity once I saw (and later launched) DeepDream. That degree of quasi-creative flexibility was an unambiguous proof, for me, that we have really grasped the sufficient basic primitive of universal learning. That's the biggest piece that biological life needed to go from insects to humans, and hardware&software cycles are inherently absurdly faster plus capital can grow exponentially, so what exactly could the counterargument even be?

In my opinion, this is "out there" on purely rhetorical grounds – skepticism is a product of incomplete information available in the popular discourse, and skeptics who profess to have some principled objections are continuously humbled and forced into retreat (like Chollet or LeCun, the French are annoying in this way; or hey, remember when Hlynka tried to burst my/others bubble here with some prose about mathematical Truth versus wordcel nonsense? I wonder how he feels about the recent math results from OAI/Ant). The coarse-grained logic is settled since Samuel Butler, the timelines and mechanisms were getting constantly refined, now we have a good idea of the necessary mechanisms and a thicket of more or less cheap ways to compensate for any particular – and still temporary – technological block.

Vernor Vinge, 1993:

I believe that the creation of greater than human intelligence will occur during the next thirty years. (I'll be surprised if this event occurs before 2005 or after 2030.) From the human point of view this change will be a throwing away of all the previous rules, perhaps in the blink of an eye, an exponential runaway beyond any hope of control. I think it's fair to call this event a singularity. It is a point where our models must be discarded and a new reality rules. As we move closer and closer to this point, it will loom vaster and vaster over human affairs till the notion becomes a commonplace. Yet when it finally happens it may still be a great surprise and a greater unknown.

Shane Legg, 2009:

UPDATE 11 April 2009: Note that these predictions do not take into account my apparent bias towards predicting that things will happen faster than they actually do (see previous post). The required compensation for technology events appears to be about 50% more time. Thus if you want the “Shane meta predictor”, then take 2033 as the expected date, perhaps with a standard deviation of 7 years.

Shane Legg, 2011:

I’ve decided to once again leave my prediction for when human level AGI will arrive unchanged. That is, I give it a log-normal distribution with a mean of 2028 and a mode of 2025, under the assumption that nothing crazy happens like a nuclear war. I’d also like to add to this prediction that I expect to see an impressive proto-AGI within the next 8 years. By this I mean a system with basic vision, basic sound processing, basic movement control, and basic language abilities, with all of these things being essentially learnt rather than preprogrammed. It will also be able to solve a range of simple problems, including novel ones.

Kurzweil, according to Google's AI, “predicts that Artificial General Intelligence (AGI) will arrive by 2029. He first made this prediction in his 1999 book The Age of Spiritual Machines, and has maintained it through subsequent books like The Singularity Is Near and The Singularity Is Nearer.”

and in 1988, “pioneering Carnegie Mellon University roboticist Hans Moravec predicted that hardware matching the computing power of the human brain would arrive by the late 2020s, enabling human-level artificial intelligence. He argued that processing capacity, driven by hardware scaling, would naturally unlock general machine intelligence.’’

While nowhere close to these giants, largely on account of my age, I believe I have a pretty decent prediction track record in this field (I have plenty of receipts here, such as talking about DeepSeek when their V1 Coder and LLM came out in late 2023; since then their architecture and much else became the default paradigm in the industry).

Nobody knew Opus 5 would be doing what it's doing now in 2026,

What about Opus 5? It's been programmed since o1-preview at least. Of course the specific product name, timing, costs, benchmark numbers and downstream capability priorities are all unpredictable but trivial. Anthropic surviving and remaining well-resourced enough to compete was very likely, the substance of the rest follows necessarily.

you have literally no idea what 2030 or 2034 will look like. None.

That's fair, that's the predictable part, as Vinge has predicted. But then again, as Yud said, you don't know how specifically a superhuman AI will beat you in chess; you can just safely bet on it doing so. I have a pretty clear idea that at this rate we will have AI doing wildly superhuman knowledge work by 2030. This is the most salient and important part, unlike the specious water use trolling. And thus the real problem people should have with AI, I believe, is still the same I outlined 3 years ago.
I strongly believe there are no remaining valid objections to this big picture. You're welcome to make some.

Context

dailydogma DaseindustriesLtd 1hr ago

I don't know why you assume "particularly not you" would have predicted GPT-3.

You used to post on an engaging variety of important topics with AI in the background. Now it seems AI, specifically, for some reason, Chinese AI, is to you the only important topic. Not even on frontier labs which still don't have chatbots that can do, e.g., decent philosophy or social criticism, but DeepSeek, which is still quite bad. I see you have moved on to GLM and Kimi which are at least like 2nd gen models. I still haven't done anything useful with DeepSeek though. But frontier models even are reddit bots that have a swarm of hackernews user coding brain lobes hooked into each other with wires, they are not even high human level unitary intelligences.

In my opinion, this is "out there" on purely rhetorical grounds – skepticism is a product of incomplete information available in the popular discourse

When exactly do you think we're getting UBI? Then when am we're getting radical life extension? And complete freedom? Because you write on Chinese AI models like Deepseek 2028 will be granting these things. I am more pessimistic. I think we could be dead before it happens. I think LLM technologies are pseudo-intelligent and could stall for decades just like the early internet. I'm out there too when it comes to popular discourse. When I said you are out there I mean it's not adding up even for me. For every 20th century prolog pseudogenius you mention in your writing there's several who think LLMs aren't the answer. Why are the latter stupid but the former are «giants»?

he coarse-grained logic is settled since Samuel Butler

Yes, but when matters. We'll have robots and AI eventually. And humans will certainly be supplanted or entirely die out eventually. Is it in the next decade or after my natural life time? What about that of my grand children? You have no idea but you act like it's in the next decade or could be. I'm not seeing it.

I believe that the creation of greater than human intelligence will occur during the next thirty years.

So by 2023. Wrong. It appears he sadly died 1 year after his failed prediction, after a career of assuming the 20th century would continue indefinitely under advanced technology:

Fairmont High is an advanced high school in a future where everyone has access to almost unlUSimited computation and bandwidth. A distributed network of tiny nodes (akin to the localizers in Rainbows End), which can be found on clothing and all over streets and houses, allows people to superimpose an augmented reality on their real environment, as well as be virtually present as a hologram anywhere in the world where the nodes are available. Most students learn via distance education but are able to attend physically when asked. One exception is Bertie Todd who lives halfway across the US. Bertie has contacts in Siberia and admits to being on a committee that controls a distance learning graduate "student" at MIT. It is hinted that Bertie is a virtual entity himself, though Juan Orozco has visited his home virtually and cannot see why anyone would fake that.

Fairmont High believes that it is important that their students be able to operate both on-line and off-line; thus, one of their exams must be completed while isolated from the network. This exam is a project of the students' choosing, and one grading criterion is to generate income for the school. It turns out that Bertie has been subtly manipulating some students, and entices Miriam Gu to team up with Juan despite them never having spoken before and their different personalities (another positive grading criterion). Bertie has also been supplying Juan with memory-enhancing drugs which enable him to structure large queries to the network in ways that ordinary people cannot, and Juan has been helping him with some of his outside research.

Miriam has decided to investigate strange goings-on in Torrey Pines State Natural Reserve for their project. Nodes are not permitted because of the toxic components that would remain when they failed. Just before they go offline, Bertie mail-drops (via FedEx personal rocket) some new biodegradable network nodes that his MIT persona has been developing. He asks Juan to test them in the forest as part of his unlimited exam, where projects can make use of any resources available (doing two projects concurrently is a third positive grading criterion). Bertie does not tell Juan how to interact with the nodes' software, in the hope that Juan will be forced to send a high-powered signal from inside the forest to ask for outside help – not only failing the offline restriction of the exam, but also allowing Bertie to receive data from the nodes and steal Miriam and Juan's project.

Under true singularity there will be no US or governments at all, no high schools, and no less-than-super humans. They will look down on these people as laughing stocks, as those people look down on apes today.

As for Kurzweil:

Kurzweil advocates universal basic income (UBI), arguing that progress in science and technology will lead to an abundance of virtually free resources, enabling every citizen to live without the need to work: "We are clearly headed toward a situation where everyone can live very well".[64] According to him, the major hurdle to introducing UBI is not its feasibility, but political will, which is slowly emerging.[65] In a 2018 TED Talk, he predicted that "in the early 2030s, we'll have universal basic income in the developed world, and worldwide by the end of the 2030s. You'll be able to live very well on that. The primary concern will be meaning and purpose."[65]

I hope he's right, but like I said I'm just not seeing it.

Of course the specific product name, timing, costs, benchmark numbers and downstream capability priorities are all unpredictable but trivial.

I am talking about timing and capability, not name or company provenance. What will the frontier model be capable of in 2030? We don't know. We don't know if there will be JS devs left then. Or in 2040. Or even 2050. Much less if we will have blue collar work robots.

I have a pretty clear idea that at this rate we will have AI doing wildly superhuman knowledge work by 2030.

Great, like automating all of labor and relieving me from all labor by 2032?

I strongly believe there are no remaining valid objections to this big picture.

Big picture is too easy. If I don't get radical life extension then none of it mattered enough to consume all of your posting effort. The problem remained other people and technology was not a savior. I mainly post about how evil other people are and how they lie about this and why. I would love for technology to free good people from their grasp in 10 years but 3 years ago GPT 4 was helping me code. They still help me code but I have to fill in less. They aren't very good at anything other than code and shell commands. You can't write with them, for instance. And they still have made no direct contribution to science. I can see that only beginning to change in 5 years. I can see 70% of the current programming job market still existing in 10; again I would love for it to go faster because working is a waste of my time.

jkf DaseindustriesLtd 3hr ago

By this I mean a system with basic vision, basic sound processing, basic movement control, and basic language abilities, with all of these things being essentially learnt rather than preprogrammed.

Seems like we are still missing a few of these bullets, even if we add his 50%?

DaseindustriesLtd late version of a small language model jkf 2hr ago

No, we have such systems, for a certain notion of "basic". It can be built as a simple omnimodal Transformer. Or something like this. It was just underrated how much can be done with language or images alone, without robotic embodiment – these criteria were informed by the assumption that human-level intelligence is more tightly coupled to our mode of existence, and sample efficiency would suffer catastrophically if we just, like, pretrained on a large web corpus.

Liang Wenfeng explicitly says that he won't bother with embodiment because there's a more fundamental problem within the current paradigm:

With the current generation of AI technology, if you can describe a problem very clearly and give it complete context and instructions, it already surpasses humans. But there is a condition, a prerequisite: you give it the complete context and the complete instructions. And this prerequisite is very hard to meet.
For example, today’s meeting: we have a very long and rich context. Perhaps each of us has decades of context. AI doesn’t have that. What AI can have today is the ability, within a limited context, to do better than humans. But it still cannot replace humans. What’s missing is continuous learning. Because humans can continue learning.
… But if AI had continuous learning — if it could, like your employee, come into the company and learn for two months — then it could replace anyone. So the next step is still missing “learning to learn.” We can understand AI development as a staircase. Last year’s step was CoT — the chain of thought. We found that by using chain-of-thought reasoning, we could push intelligence higher. By letting AI think for itself, the ceiling rises and AI can do more. We crossed one step. This year’s step is Agent. We found that with agents, even more things can be done: the range of capabilities expands, and the upper bound of intelligence rises. Why is it a staircase? Because every later step builds on earlier ones. Agent uses CoT, and CoT uses the previous step — the language model. So no step was wasted.
Thus, the development of AI, the direction of intelligence, is traceable. This year’s step is Agent, but Agent will also eventually reach the end of its step. When it solves all solvable problems, it still won’t replace your employees — it will just have reached its ceiling.
Like CoT: after CoT reached its ceiling, it already surpassed the best humans at Olympiad math and programming. But it stopped there; that technology didn’t reach AGI. See, the direction of intelligence is traceable.
… Where we stand now, at the Agent step, we can see the next bottleneck: continuous learning. The next problem to solve is how to make continuous learning work. This is visible, and relatively clear. It’s the obstacle right in front of us. You must cross it, and there must be a way to cross it, but it will take time. After continuous learning, we may arrive at a singularity. That singularity is: when the model can continuously learn, it can do everything humans can do. It will be able to develop its own next version, conduct research itself, and create the next, more advanced AI model. So it will reach a singularity — self-iteration. But this singularity is not truly a singularity; it’s also a gradual process.
The process may be a long, gradual shift, not a sudden mutation. But habitually, people call it a singularity, because earlier prophets thought there would be a singularity. But actually, it’s not a singularity — it’s a continuous process. And after this step, I think the next is embodied intelligence.
This is our speculation — our view of the timeline: first solve learning-to-learn, then reach the self-iterating intelligence singularity, and only then embodied intelligence. After embodied intelligence, AI enters the real world: it can do housework for you, care for the elderly. We think this is an ideal roadmap. Everyone has different views; there is no right or wrong. We just think this roadmap is the easiest. The reason is that each step requires very little that is new. With this roadmap, we don’t have to work overtime. But if the roadmap were reversed — say, embodied intelligence first — then you’d be doing very hard, exhausting labor. We don’t want a roadmap like that. We want to do it the easy way.
If we first solve continuous learning, then the self-iterating singularity, then embodied intelligence, the path is easy. Because later you can use earlier technologies to help develop later ones. After the singularity, embodied intelligence no longer needs to be built by humans — the model itself will produce it. So this answers the question about our long-term goal.

We've made 1 million token context easy and cheap. With LLMs simply greping over a codebase, writing their own memos and using RLM-like tricks it's not hard to have them operate over many millions of tokens continuously. Then there are techniques to compress a given context into a higher-density "cartridge" prefix. Given strong priors from pretraining, this can functionally substitute for most of true continuous learning, in the sense that agents will be able to do long-range tasks well beyond the pretraining distribution. I'm pretty optimistic about the trajectory here. Robots will continue to develop in parallel for a while but that's just because this is "the easy way". We could merge it already.

What is this place?

Why are you called The Motte?

New post guidelines

Rules

Recommended Posts And Communities

Recommended Realtime Chats