Just to add on, armchair quarterbacking is a thing, it’s easy in hindsight to label decisions as the result of bad intentions. This is completely different than whatever might have been at play in the moment and retrospective judgement is often unrealistic.
I switched from Android to iPhone last year, and this just isn’t true. There’s so many tiny issues with android apps that just don’t exist on iPhone, because the android apps have to work on all these different devices. You don’t even have to look for the kinds of apps you’re talking about because things like Safari and Apple Podcasts work really well. I know people have a lot of complaints, but things on the iPhone really do “just work”.
Reading about this, it sounds like it has a dangerous amount of permissions. While AI can be useful, it often goes off on unexpected tangents and giving it access to control your computer remotely sounds like a terrible idea.
-4 makes sense if you understand that the input -2 is a unary minus operation. So typing -2 then hitting square only squares 2, not (-2). This is the same in eg Python so I'm not sure it's very controversial. I agree it's unexpected, though.
I think Europe hasn't developed this kind of political manipulation ability. Europe seems to operate in the mode where as long as the political institutions are still standing, everything is felt to be alright. US democrats also operate in this mode.
Hi HN — I built this because most AI app builders frustrated me.
They’re usually cloud-only, expensive, and generate code that looks good in demos but breaks or hallucinates when you try to extend it. I wanted something I could actually use for real projects.
So I built a local-first full-stack generator that runs inside VS Code and writes files directly to your machine.
It can scaffold:
frontend (React/Next/etc)
backend APIs
database models
basic auth/CRUD flows
The focus is reliability over “magic”. I try to keep generation constrained and file-aware to reduce hallucinations and produce clean, editable code.
Cost is ~$0.5–$1 per project using small models, so it’s cheap enough to use daily.
Still early, but I’m already using it to ship internal tools and small client apps.
Would genuinely love feedback from other devs — especially where it breaks or feels unreliable.
Do you also wish you could only get telephone calls from people using American made handsets, and that your email client asked you before receiving emails from other email clients, and that you couldn’t get SMS’s from other smartphone manufacturers without opting in one at a time?
Being able to reject spam , regions, specific people, specific topics, all makes sense. Wanting to approve/reject the program used to make the connection is a pretty useless way to segment communications - how will you determine “questionable” clients, and what when there’s a person you want to chat with and a person you don’t both using the same client?
I built a local-first AI full-stack code generator that creates complete web and mobile apps from prompts.
Most AI app builders run everything in the cloud, are expensive, and often hallucinate or generate messy code that’s hard to edit. I wanted something cheaper, private, and more predictable.
So this tool:
runs locally in VS Code
saves all files directly on your machine (no remote servers)
generates frontend + backend + database together
produces editable, production-style code (not locked templates)
costs ~free for now per project using small/efficient models
focuses on reducing hallucinations with constrained generation + file-aware context
You can scaffold things like:
CRUD dashboards
SaaS apps
landing pages
REST APIs
full-stack MVPs
The goal isn’t “magic one-click apps”, but practical, reliable code you can actually maintain.
Still early, but already using it to ship internal tools and small client projects.
Would love feedback from other devs — especially on reliability and workflows.
If we're talking about big science&technology categories, I'd say:
Controlled fire (if you can consider it a "man-made creation") -> essential for food and a lot of manufacturing
Wheel -> essential for transportation, but also to make flour (millstones), and a lot of other stuff (e.g. turbines are, basically, specialized wheels)
Controlled electricity and electromagnetism -> artificial light, modern communications, not to mention medical advancements like X-rays
Insulin and pecillin -> millions of lives saved
the printing press -> knowledge becomes easier to spread
If we extend this to all kinds of human "inventions", including law, philosophy, religion, and so on, the list is even longer.
Orchestration buys parallelism, not coherence. More agents means more drift between assumptions. Past a point you're just generating merge conflicts with extra steps.
Is that happened? I'm not an expert on the history of Bretton Woods, but my understanding is countries that sent gold to be stored in the US retained ownership over it, and repatriation was never refused.
If they had converted gold to dollars, this only suspended their ability to "rebalance" between gold and dollars, and no value was lost.
Yegge's been around a long, long time and this is about within a standard deviation of his normal writings, at least in style. I haven't read much of his LLM/AI related stuff, but none of Gas Town left me with any sort of "huh" reaction, knowing the author.
The section on multi-column indexes mirrors how I was taught and how I’ve generally handled such indexes in the past. But is it still true for more recent PG versions? I had an index and query similar to the third example, and IIRC PG was able to use an index, though I believe it was a bitmap index scan.
I am also unsure of the specific perf tradeoffs between index scan types in that case, but when I saw that happen in the EXPLAIN plan it was enough for me to call into question what had been hardcoded wisdom in my mind for quite some time.
Further essential reading is the classic Use The Index, Luke [0] site, and the book is a great buy for the whole team.
This is the flight where one pilot tried to pull up to recover from the stall, and the warning for dual input (which Airbus just averages together) was snoozed by the system yelling about the other errors and was reduced to a light they didn't notice. The captain commented towards the end"no don't climb". The stall alarm was the one the system chose to display over all others and was mishandled (by the pilot who didn't know how to recover from a stall).
Boeing there's physical feed back, when one control moves so does the other.
This was not the first time pilots were having conflicting input without noticing.
The small boutique mail hosts are also so much more tedious to deal with than any of the big players. So it depends on your recipients how much effort it is.
An ad hominem argument is an argument constructed around characteristics of a person outside the bounds of what is being discussed. Inferring someone's opinion[1] about the subject under discussion from their text, and explicitly marking so in my text when doing so, is just "debate". Am I wrong? Say I'm wrong and cite why.
Don't call me "disgraceful". Why? Because THAT is an ad hominem attack. In fact the clear offense being taken makes it pretty clear to me that my point landed closer than maybe you're prepared to admit.
[1] You cleverly skipped the point where I even admitted I might be wrong!
