Hacker Newsnew | past | comments | ask | show | jobs | submit | weinzierl's commentslogin

Here is vunderba's site, for anyone who got curious too:

https://mordenstar.com/

I just skimmed but looks like exactly the content I love. It's a shame search engines are so narrow minded and such content so hard to discover.


Images could be doctored convincingly at least since the 1920s and yet people used "pics or it didn't happen" unironically nearly 100 years later. Video will not be different. People believe what they want to believe and will never let get truth in their way.

These processors were very very different from what we have today.

They usually only had a single general purpose register (plus some helpers). Registers were 8-bit but addresses (pointers) were 16-bit. Memory was highly non-uniform, with (fast) SRAM, DRAM and (slow) ROM all in one single address space. Instructions often involved RAM directly and there were a plethora of complicated addressing modes.

Partly this was because there was no big gap between processing speed and memory access, but this makes it very unlikely that similar architectures will ever come back.

As interesting as experiments like LLVM-MOS are, they would not be a good fit for upstream LLVM.


> ... there was no big gap between processing speed and memory access, but this makes it very unlikely that similar architectures will ever come back. ...

Don't think "memory access" (i.e. RAM), think "accessing generic (addressable) scratchpad storage" as a viable alternative to both low-level cache and a conventional register file. This is not too different from how GPU low-level architectures might be said to work these days.


Great point. And you can even extend that to think like a 6502 or GPU programmer on an AMD, ARM or Intel CPU as well if you want the very best performance. Caches are big enough on modern CPUs that you can almost run portions of your code in the same manner. I bet TPUs at Google also qualify.

When I grew up in Germany it always made me proud that 100% of taxis were Mercedes Benz. If a car can withstand the rough demands of taxi service, it has to be good. And even in South America back then German cares were ubiquitous, especially Volkswagen.

When I was in Brazil this spring[*] I rode a lot of Uber and they were 100% BYD - 100%, no exception. It's not that my head hadn't known that German auto was dead but seeing it playing out like this hit hard.

[*] northern hemisphere


BYD recently went live with a highly automated, large scale manufacturing facility in Brazil. The BYD Dolphin Mini sells for ~$22,500, and the manufacturer already has 200 showrooms open across the country.

https://www.politico.com/news/magazine/2025/11/22/cop-brazil...

https://www.byd.com/us/news-list/First-BYD-Electric-Vehicle-...


They are everywhere. The only limit to adoption is that many people live in buildings where chargers can't easily be installed.

The limit to adoption is non existent or slow public charging infrastructure.

Both limit adoption. Public charging infrastructure would solve the problem of nit being able to charge at home.

You have to consider that South America is the most dangerous continent.

You can't just leave your car charging unattended in a public space. It has to be done at home or somewhere closed (which would make it expensive) or you would have to watch over your car (which would take a lot of your time).


How do parking lots even work in South America then?

At supermarkets and malls they usually have guards.

Lots where people park overnight are gated.


There you go then.

Not having public charging means you can't travel more than << 50% range away from home.

I'm surprised Mercedes was ever price competitive for taxis even in Germany, I mean the average VW would do the job just as well at half the cost.

These days they're both priced like they're selling Ferraris anyway so yeah. The ID Buzz starting at 70k EUR is such a joke.


Mercedes and BMW serves a wider band of value than they do in the USA, where they have purposefully cultivated themselves as a pure luxury brand. For example, the 1 and 2 series from BMW or Mercedes A class will never go the USA, even though I’m sure there is a market for them.

They used to make them with quality construction. Now it's all engineered with plastic bits that will only last for the first 5 years the rich owner will be using it before tossing it out.

I remember all taxis in Portugal being beige Mercedes in the 80s,when Portugal wasn't well off. I guess their durability is what made them worthwhile.

I can't wait for BYD to enter the European market with a true minibus like VW's ID Buzz. There is a rumor that the M9 is coming in 2026.

Mercedes taxis in Europe are not appointed like the cars they sell in North America. They are just normal cars there.

I remember being in a Mercedes in France in the 80's and noticing it had manual crank windows. My dad in the US (even then!) hated added electronics in cars so he went to a Mercedes dealer and they explained that in the US we could only get fully loaded models.

Tesla has no moat

1. Batteries - BYD has them beat 2. Self Driving tech - other players are better 3. Luxury brands already provide the luxury aspect & even better built cars 4. in the US they're being saved by US protectionism. in Europe etc - we already see the chinese brands making inroads for EV sales


I mostly agree on all points, but what self driving tech is better? I've periodically looked at the options, and nothing really seems to compare in North America. Maybe BYD and others have great tech, but stuff like Blue Cruise works hardly anywhere in Canada, and to me, that makes it virtually useless.

He’s probably thinking of robo taxi self driving. So that would be e.g. Waymo.

I don’t think anyone has better self driving for consumers out atm, but you could argue that’s because other companies are not using their customers as beta testers. I’ve seen demos that may indicate Mobileye has tech that’s just as good if not better. But they don’t release it to end users until it’s fully ready.

I don’t think Tesla has any special sauce, and that when the tech is actually ready for unattended full self driving in a consumer car, other car makers will come out with solutions around the same Tesla. One difference is maaaybe Tesla will be able to update old cars (probably with a hardware update). While I think others will only support it on new cars.


Yeah this.

The argument for years has been something like:

> Tesla will solve self-driving and everyone will be left unable to compete. Also, AI is advancing rapidly and will solve all kinds of problems for society.

But apparently it will not solve self-driving for anyone else but Tesla.

I gave up trying to argue with Tesla fans years ago. They are immune to logic which invalidates their priors.


Certainly waymo is better, but you can't buy it. Yet, anyway.

http://comma.ai isn't self-driving, just really good cruise control (better than Blue cruise, imo), but most importantly, you can get it today. (And for less than $8k.)

TBF all I really want is highway driving. I probably even prefer signalling lane changes manually. Comma looks pretty neat.

That’s just Tesla auto pilot and smart cruise control on most new models. It’s definitely nice, but not full self driving.

"I've tried all the options I have here in the country that banned the major players and I don't think any are better."

“I like to be snarky and mean to people on the internet who aren’t as lucky as me to have diverse consumer choices.”

Protectionism on inputs kills manufacturing. Imagine having to pay 15% more for all inputs and trying to compete with someone who doesn't have pay that.

Well, at least domestically you don’t have to compete with someone who doesn’t have to pay that because their product is probably tariffed directly.

Internationally, yes if you manufacture the international product in the home country, but AFAIK in auto at least there are usually satellite factories and have been for some time, and those wouldn’t be subject to home country tariffs would they?


Precisely, you need to set up plants overseas to dodge the input tariffs instead of onshoring manufacturing for export. That causes reductions in manufacturing investment compared to the alternative.

Compared to the alternative where the domestic plants don’t exist at all because they are competing domestically with low cost foreign products?

In that hypothetical they can't compete because of labour protectionism and immigration restrictions, not because there's some intrinsic reason it's too expensive to manufacture in one country versus another.

Not only that, Teslas nowadays look like they belong in a museum - so fucking old and outdated. I own 2014 Tesla S, my neighbour has 2025 Tesla S, same fucking car - literally. My car was THE shit back in the day, lots of broken necks looking at it … now, 12 years later, someone (fewer and fewer) is paying $90k for exactly the same car. Tesla X was great looking - circa 2016… Tesla 3 is like a Kia and Model Y is just 3 that is blown up a bit.

> My car was THE shit back in the day, lots of broken necks looking at it … now, 12 years later, someone (fewer and fewer) is paying $90k for exactly the same car

That's what Porsche also discovered, the hard way.


I live in the US and I would happily buy a Chinese EV if the price was right. However, there is no price for which I will buy a Tesla, ever.

Tesla also has a big Elon problem in that the blue cities where self-driving Taxis will be most profitable may opt for Waymo or boycott Tesla over politics.

I just value my life enough that Waymo seems to be the better route. Tesla hasn’t shown that they’ve solved the problem while waymo definitely has.

I’m not going to allow “Nazi salutes” to be downgraded to “politics”. It’s still too soon to be a Nazi. Hopefully it will always be.

I think the issue is to create an ICE is a very complicated process requiring lots of specialist knowledge, skills and technologies. An EV is just much simpler, comes down to who has the cheapest batteries. Europe and Japan are great at the former, the latter no chance.

Im sure some of it is personal bias from experience with them but I don't think ICE are as complicated as some people think. 90% of the extra shit on them are unnecessary for it to work but what those things are and what they do and why it broke or failed or how important they are is essentially obfuscated from the general public so they seem like overly complicated magic. The vast majority of cars I see do not fail or get trashed due to engine failure from design flaws or anything, most get trashed because people stop caring about them and treat them like trash and don't replace that $15 sensor, others think they can't afford the maintenance because car manufacturers don't give a fuck about having to take 3 hours disassembling other unnecessary shit to replace a 30 cent sensor that they know will eventually need replaced, but the only number the customer seems to look at is the total cost of the mechanic quote. They think because something is a $1,000+ repair that something seriously is worn or old and that the car is on its last legs, instead of the reality of that one part being a huge pain in the ass to replace but it is otherwise a good reliable motor for another 100,000+ miles. And of the cars that do get trashed because they have actual major mechanical problems, the vast majority of the problems have to do with the body work rusting out and/or suspension components needing replaced after being used for 3x their expected lifetime, which an EV is not going to improve in any way.

Like ive seen people junkyard perfectly working and good cars because it is over 150,000 miles and some service guy who is looking for work/money told them they need to do scheduled maintenance some time soon and they thought the car was too old and was junk. And yet very few cars ive seen would not make it over 300,000 miles if they spent even 1/10th the money of their new car for maintenance on their old.


Thought the comment was somewhat helpful. Sparked considering the various anti-patterns in automobile design and searches came back with several others that have been vaguely thought about, just never really identified very clearly for me.

  - Inaccessible Components (Poor Physical Layout): One of the main ones you're talking about.  Take out the engine to repair a light on the dashboard.

  - Integrated, Non-Modular Systems: Minor damage or failure ruins an entire assembly.  You dinged the bumper, replace the entire front.

  - Lack of Standardization: Even from year-to-year, designs change and mechanics have to learn yet another system.

  - Forced Replacement over Repair: Object is "black box", thou shalt replace, not repair.

  - Dead/Onion/Boat Anchor Components: No longer used, maybe need it, build stuff on top of it, layer after layer, "can we even remove it"?

  - Spaghetti Wiring/Code and System Coupling: Single modules that route all over the car, another "can we even remove it"?

  - Proprietary Diagnostics and Restricted Data Access: Don't have the special tools, you can't repair, or even find out what's wrong.

Ok fine, take that as a baseline for "not very complicated". EVs are less complicated, and take less maintenance.

EVs are very complicated cars anyway. They need maintenance in service as well as ICE cars. Yes, not so often you need to change liquids, but service is required. Also good luck with water/rodent damage to 400v parts.

Optimizing costs while producing a safe, reliable, durable vehicle isn't exactly simple and requires an entire supply chain to be in place, not just a single company. Look at how many auto mfgs there are in the world that turn out terrible cars. EVs dramatically lower the parts count which helps but you still a lot of expertise to make a safe, reliable car.

My grandfather was a mechanic and told me how replacing a dashboard light in some models required removing large portions of the engine to access the socket.

Europe and Japan are great at the former, the latter no chance.

Europe and Japan should be totally capable of producing super inexpensive batteries. They just don't, at the moment.


How? By building entirely automated factories, they way they do for medicine production?

To be fair taxis have unique requirements. Taxis in the UK were like 80% Prius for a long time because they drive very long distances and hybrids are very cost effective for city driving where you're doing a lot of low speed driving and don't have convenient recharging opportunities. But most people aren't in that situation.

Still, I think BYD are kind of killing it.


> But most people aren't in that situation.

Those who commute from the suburbs actually save even more as hybrids achieve their lowest consumption going a steady 50-70km/h.

Of course the same people could just get an EV and charge at home, but in terms of cost-effectiveness hybrids still win in this use case.


Unless you have a really cheap electricity at home (not like in EU) the best price per km has old Prius with LNG fuel. Also reliable, there are tons of them with 500k+ km on the clock.

Over here LPG is the gas of choice and indeed taxi drivers regularly have that installed in their hybrids.

I've been riding a German electric motorbike for a couple of years, and before that, German electric mopeds.

I think there is a lot of innovation in the German electric vehicle industry. I am quite excited for BTM, my bikes manufacturer, to design and release new versions of their platform. This model is distinctly German.


Note the Mercedes as taxis in Germany are not the high end luxury car imports we are accustomed to seeing in the US. Mercedes makes a lot of more affordable cars for their domestic market we never see here!

Sadly the German car industry has lost its way in the EV transition and is now vainly trying to get the EU to rollback the sun setting of ICE car sales in 3035.

Meanwhile the Chinese are eating their lunch.


3035? :)

Even 1010 years isn't enough time.

In the UK, for a very long time they were Skoda Octavias.

I know of two ex-taxis that were scrapped at about five or six years old - one was taken off the road because of a deep paint scratch down to bare metal from about half way along the front wing to the rear door, rendering it beyond economic repair - with over half a million miles on the clock each.

Neither had been outside the Greater Glasgow area since they were dropped off on the transporter.


Who remembers the ICMB meta tag?

https://en.wikipedia.org/wiki/ICBM_address

(It's funny that the Wikipedia article has it in the section Modern use)


This and schema support for JSON will always be an ill fitting afterthought. If you really, really need the strictness and correctness (and you most probably don't) XML has you covered.

To me schemas are absolutely essential for high quality software, I have a hard time imagining a project where I wouldn't want to use them.

I also often make heavy use of schema based editors to auto-generate UI.

Perhaps I'd feel differently if I ever had to write a validator myself, but they seem to exist in all the popular languages as is.


An e.V. in Germany mostly makes sense if it is also charitable. You will have a hard time justifying that for an arbitrary FOSS project.

There are about 100 categories that are predeclared as potentially charitable and you have to fit into one of them. Most of them are weirdly specific like homing-pigeon breeding or model plane flying.

The only two that are broader and remotely realistic candidates for a FOSS project are religion and education.

If you don't want to start a cult you are left with education. That is as how organizations like Chaos Computer Club do it. Education means education for the general public though and it is not enough if you offer occasional courses for a niche topic. It has to be something that potentially interests everyone. The tax office is checking that and it is on you proof it to them regularly.


https://techcultivation.org is a German broad-spectrum FOSS charity.

Chaos Computer Club is not a charity.


Ah yes, you are right, I confused the parent with my local "chapter" (Erfa Kreis).

I believe the "Chaos Computer Club München e.V" is (or was at some point in time) charitable. At least their bylaws have the magic words:

"Durch die genannten Zwecke sollen Kultur, Bildung und Wissenschaft gefördert werden."


Thanks for bringing the "Center for the Cultivation of Technology" to my attention. They're noteworthy because they are not a club but a charitable company (gGmbH). In the end the need to meet the same criteria but it is an interesting choice.

The single most annoying thing with these old displays was the flicker. Whenever I use one of my real old home computer era monitors it is the only thing that makes it unbearable after a while.

But I'm not surprised they don't go overboard with that in the emulators. They'd probably have to add PSE warnings if they did.


My sister tried to go through broadcast school, with great difficulty especially when she got through the video editing classes. Turns out she has photosensitive epilepsy and all the exposure to CRT monitors made her quite ill. You couldn't convince her to go back to the CRT days for all the tea in China.

"Grant of Copyright License. Subject to the terms and conditions of this Agreement, You hereby grant to Core Devices and to recipients of software distributed by Core Devices a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare derivative works of, publicly display, publicly perform, sublicense, and distribute Your Contributions and such derivative works."

A few years from now we will see the usual HN thread were contributors lachrymosely complain about how their precious work was stolen by a good-turned-evil organization.


Please note our CLA explicitly include a clause to require Core Devices to distribute all contributions under an OSI-compatible FOSS license (e.g. GPLv3). So no contributions can be 'stolen'.

https://ericmigi.notion.site/Core-Devices-Software-Licensing...


But OSI-compatible FOSS licenses include pushover ones like MIT, so even though you couldn't steal all of the contributions to make a proprietary fork, any other company then could.

The CLA very well allows them to use all of the contributions to make a proprietary fork. The license is irrelevant for that[1].

Calling it "stealing" doesn't help.

[1] I'd usually call it a distraction, a sleigh of hand, smoke and mirrors, but we have to give Eric credit for not burying the CLA.


well, their new app mentioned in the post is GPL.

but in the rest of the ecosystem a bunch of the inherited code is already Apache or MIT. so, i presume you have already forked the other repos to relicense them. can you drop a link?


I'm the first to agree that contributions can't be stolen in this scenario but read the threads I'm referring to. People feel that way anyway if you stop supporting a component or distribute your focus between a free and a paid tier.

What we need is more awareness that looking at the license alone is not enough to make an informed decision if contributing to a project is aligned with the contributors attitude and personal goals.

With that in mind: Thank you for putting the CLA right in the repo where it belongs and people can easily find it. Many organizations put a license upfront and bury the CLA. For a particularly bad example try MonoDB.


> lachrymosely

Learned a new word, thank you!


The h is peculiar, as the latin word reads `lacrimosus`

Interesting etymology:

https://www.etymonline.com/word/lachrymose

Greek dakryma

"-d- to -l- alteration in Latin"

"The Medieval Latin practice of writing -ch- for -c- before Latin -r- also altered anchor, pulchritude, sepulchre. The -y- is pedantic, from the former belief that the word was pure Greek."


Pebble was bought a long time ago by google yea? So this already happened

not exactly, fitbit acquired pebble in 2016, and google acquired fitbit in 2021. so while google did end up with the IP in the end, they did the Right Thing this year and open-sourced PebbleOS.

i don't normally praise google, but i am glad their open-sourcing of pebbleOS here enabled this new revival of Pebble devices.


Unlike many I was never really unhappy with TeX, I still think it is amazing. I used to be skeptical about Typst but after using it for a while I cannot really imagine any scenario where I'd go back to TeX or LaTeX again.


I feel that LaTeX is still a good option for broad use, as it is the default that "everyone" knows. It may happen that Typst will be forgotten in 10 years, I doubt that will happen with LaTeX.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: