the cheap house thing has nothing to do with if you’re foreign and everything to do with a bank not wanting to lend money for something there’s no market for.
you can easily buy these homes with cash. i know because i have a ton of friends that have done exactly this.
if you want a cheap house in a good market like tokyo, osaka or fukoaka, then you can do so through a bank via the normal routes
I just read the book last week. What you said is not true in any useful sense. “Germans were acutely aware…” tries to reduce an entire population and years into one statement. Reality has much more color.
For the germans interviewed in the book, it seems to be true that many had read or heard about the camps or other atrocities, but (1) not the “final solution” which was not in the press and (2) there seems to be heavy desensitization from 1933-1955 when the book was written.
Aside from the tailor that had started the fire at the synagogue, the other 9 interviewees had not directly witnessed atrocities being committed, and instead focused on their personal hardships during the war.
Even though they may have been literate, the people in Mayer’s book were ignorant of the specific realities. Perhaps willfully ignorant, yes, but the nazi regime really did not give any opportunities otherwise.
—
not an expert, just reporting my notes from the book.
i highly recommend all americans read it, its not a long book. it feels eerily familiar, even though many circumstances are drastically different.
Mein Kampf was published 1924 and distributed broadly.
There was not much hidden, the goal of making a big war in the east to conquer new land for the Aryans was there in big letters in the open.
His views towards jews likewise.
So they knew. Maybe largely did not wanted to know. And they did celebrate the victories of the german army as their own. They only stopped celebrating after the victories stopped happening and it was more and more clear that the war will be lost.
In their defense, there is an inexhaustable supply of "take over w my ideology material."
This is a confluence of many conditions. Some long-focused efforts, some architecting and annealing of interests, some individual greed, some long-lasting effects of trauma, and some massive ignorance.
One of the only good points is that the American people are stubbornly allergic to authoritarianism. Yes there are exceptions, but mainly carved out by people trading it for self-interest. Many good surprises like Tucker Carlson's opposition to squashing free speech and the Republican's long-lasting distaste for pedophilia are still out there.
The post above pointing out how we're diff to Nazism is on point. There have been many more authoritarian plays since then. Americans remain conveniently ignorant of them.
Also we're being economically crushed and everyone feels it. Although racism is a powerful tool by this movement, it's actually centered around impoverishing everyone and the dizzying egos of its leaders.
There is no anti-authoritarian party. Are lockdowns not authoritarian? Do mandates to take an experimental vaccine not violate bodily autonomy? How quickly everyone forgets the widescale censorship and lawfare. Snowden had to flee the country and Chelsea Manning was imprisoned during the Obama presidency.
On a more pragmatic level, take the one-party state of California, and the absurd burden of its regulations. These largely prevent the construction of anything new, as seen in the infamous high speed rail project, and the restricted supply of new housing, pricing many young people out of ever owning a home. Perhaps you don't think regulations are authoritarian, yet they're enforced with the power of the state, which wields the monopoly on violence.
One side wants to impose restrictions to avoid loss of life and breakdown of the hospitals. The other wants some people to not exist anymore and are building camps to accomplish that.
" Each
animal mates only with one of its own species. The titmouse cohabits only with the
titmouse, the finch with the finch, the stork with the stork, the field-mouse with the
field-mouse, the house-mouse with the house-mouse, the wolf with the she-wolf, etc."
(from Mein Kampf, Chapter 11)
But if no one would have taken him serious, there would not have been a problem.
But people did take him serious, they seriously believed he was some kind of messias send from god to save his troubled great country.
What I took away from the book was that all these people were very eager to say variants of 'das haben wir nicht gewusst' when at the same time they also describe how the jews were systematically removed from their society and every part of civil society was taken over by the nazi's.
I would add to your statement that almost everyone should read it. It's unnerving to read how 'normal' all these people were in some way and how 'easily' it all happened because the population generally disliked jews.
Based on history books I read (mostly from Richard Evans), they knew. Nazi violence and concentration camps were public knowledge, because the regime needed to generate the fear. Germans prior war were in fact scared a lot.
This particular book is a out what nazi sympatizants and nazi themselves were saying after the war. It is what it is, but there was real motivation to not have own culpability in destruction of Germany in the open. (Which is what they have seen as tradegy, not the holocaust itself all that much)
It's a contentious issue and many historians disagree. However, even many Jews at the time didn't really know what was going on, as evidenced from letters and diaries of the time. Many Jews genuinely thoughts they were Ghettos or concentration camps, which were surely horrible and surely people needless died there, but are far removed from outright extermination camps. So based on that I'm somewhat inclined to believe many didn't really "know" about the extent of the Holocaust.
Of course it's easy to say in hindsight they "knew" or "could have known", but in hindsight everything is easy, right? There were rumours about Jimmy Saville going back to the 70s, but did the British public really "know" what he was up to? Evne Mark Lawson, one of the few people who actually did stop and report a sexual assault (in 2006, see [1]) didn't really know the full extent of things, not really. He may have suspected, but that's not the same.
Another thing is that during the first world war there was a lot of (mostly British) propaganda about atrocities Germans were supposed to have committed, from raped and crucified nuns to Germans killing children for sport to the infamous "German Corpse Factory". This was widely reported and believed during the war, but after the war this all turned out to be a huge load of bollocks. It severely undermined the trust in the media.
There was 21 years between the wars – that's less time than the start of the Iraq war and today. Imagine what your response would be if the US government would say "we found weapons of mass destruction in $country, here as some vague satellite photos as evidence, we have no choice but to invade".
when i’ve done toy demos where GPT5, sonnet 4 and gemini 2.5 pro critique/vote on various docs (eg PRDs) they did not choose their own material more often than not.
my setup wasn’t intended to benchmark though so could be wrong over enough iterations.
A World Model is a theoretical type of model that has knowledge about the "real world" (or whatever world or bounds you define). It can infer causalities from concepts within this world.
Yes, you can use Genie 3 to train other models. Its far from perfect. You still need to train Genie 3. And its training and outputs must be useful in the context of what you want to train other models with. That's a paradox. The feedback loop needs to produce useful results. And Genie 3 can still hallucinate or produce implausible responses. Symbolism is a wide term. But a "World Model" needs it to make sense between concepts (e.g. Ontologies or the relation of movement and gravity).
>The feedback loop needs to produce useful results. And Genie 3 can still hallucinate or produce implausible responses
The solution to this is giving the model a physical body and actually letting it interact with the real world and learn from it. But no lab dares to try this because allowing a model to learn from experience would mean allowing it to potentially change its views/alignment.
Labs have been doing that since Brooks' Subsumption Architecture decades ago. The problem with AI now is that the architectural design, unlike the brain, doesn't have grounded memory and hallucination mitigation. Letting those architectures walk around in the real world would show similar flaws.
Multiple teams already baked memory into designs, some like typical ML and some biologically inspired. Hallucination mitigation needs a ton more research. My proposal was studying the part of the brain that causes hallucinations when damaged in case it's designed to mitigate them. Then, imitate it until we have something better.
What an insensitive, assumptive, stupid remark. You can't possibly know that the person you replied to behaves as you claim. It's 2025, the firebombing of Tokyo is widely recognized now, maybe not by most normies but certainly by any historical adjacent nerd.
I have a team that’s somewhat junior at a big company. We pretty much have everyone “vibe plan” significantly more than vibe code.
- you need to think through the product more, really be sure it’s as clarified as it can be. Everyone has their own process, but it looks like rubber ducking, critiquing, breaking work into phases, those into tasks, etc. (jobs to be done, business requirement docs, domain driven design planning, UX writing product lexicon docs, literally any and all artifacts)
- Prioritize setting up tooling and feedback loops (code quality tools of any and every kind, are required). this includes custom rules to help enforce anything you decided during planning. Spent time on this and life will be a lot better for everyone.
- We typically making very very detailed plans, and then the agents will “IVI” it (eg automatic linting, single test, test suite, manual evaluation).
You basically set up as many and as diverse of automatic feedback signals as you can.
—-
I will plan and document for 2-4 hours, then print a bunch of small “PRDs” that are like “1 story point” small. There’s clear definitions of done.
Doing this, I can pretty much go the gym or have meetings or whatever for 1-2 hours hands off.
I think this is a good use of AI. Change your thinking - the code is, and has always been, a medium between the computer and the human. Where is the human? Where do we define our intent? AI gives us a chance to redefine that relationship or at least make it more fluid.
A well-architected system is easier to develop and easier to maintain. It makes sense to put all the human effort into producing that because, lo and behold, both humans and LLMs can produce much better results within a well-defined structure.
Everyone is responsible for what they deliver. No one is shipping gluttonous CLs, because no one would review them. You still have to know and defend your work.
Not sure what to tell you otherwise. The code is much more thought through, with more tests, and better docs. There’s even entire workflows for the CI portion and review.
I would look at workflows like this as augmentation than automation.
What this actually means is that your manager gets a raise when the AI written code works, and you get fired when it inevitably breaks horribly. You also get fired if you do not use AI written code
1. Mostly written by LLMs, and only superficially reviewed by humans.
2. Written 50-50% by devs and LLMs. Reviewed to the same degree as now.
Software of type 2 will be more expensive and probably of higher quality. Type 1 software will be much much more common, as it will be cheaper. Quality will be lower, but the open question is whether it will be good enough for the use cases of cheap mass produced software. This is the question that is still unanswered by practical experience, and it's the question that all the venture capitalists a salivating about.
I 100% guarantee you there will be plenty of software still written fully by humans—and even more that's written 95% by humans, with minor LLM-based code autocomplete or boilerplate generation.
“We typically making very very detailed plans” - this is writing code in English without tests. Admittedly, since generating code is faster, you get faster feedback. Still, I do not think it as efficient as an incremental, test driven approach. Here you can optimize early on for the feedback loop.
You get faster feedback in code, but you won't know if it actually does what it's supposed to do until it's in production. I don't believe (but have no numbers) LLMs speed up the feedback loop.
yes, but those aren’t released and even then you’ll always need glue code.
you just need to knowingly resource what glue code is needed, and build it in a way it can scale with whatever new limits that upgraded models give you.
i can’t imagine a world where people aren’t building products that try to overcome the limitations of SOTA models
My point is that newer models will have those baked in, so instead of supporting ~30 tools before falling apart they will reliably support 10,000 tools defined in their context. That alone would dramatically change the need for more than one agent in most cases as the architectural split into multiple agents is often driven by the inability to reliably run many tools within a single agent. Now you can hack around it today by turning tools on/off depending on the agent's state but at some point in the future you might afford not to bother and just dump all your tools to a long stable context, maybe cache it for performance, and that will be it.
There will likely be custom, large, and expensive models at an enterprise level in the near future (some large entities and governments already have them (niprgpt)).
With that in mind, what would be the business sense in siloing a single "Agent" instead of using something like a service discovery service that all benefit from?
My guess is the main issue is latency and accuracy; a single agent without all the routing/evaluation sub-agents around it that introduce cumulative errors, lead to infinite loops and slow it down would likely be much faster, accurate and could be cached at the token level on a GPU, reducing token preprocessing time further. Now different companies would run different "monorepo" agents and those would need something like MCP to talk to each other at the business boundary, but internally all this won't be necessary.
Also the current LLMs have still too many issues because they are autoregressive and heavily biased towards the first few generated tokens. They also still don't have full bidirectional awareness of certain relationships due to how they are masked during the training. Discrete diffusion looks interesting but I am not sure how does that one deal with tools as I've never seen a model from that class using any tools.
you want more options for voices that reflects all the types of people in the world. good feedback.
The next part i’m only saying because it reminds me so much of my younger self: The rest of what you said, and how you said it, has a lot of projection and insecurity.
