Similarly, the whole idea of automated driving becoming the default and humans not being able to drive is something I'm 100% against, since it'd be both a privacy nightmare and take away a lot of people's freedom.
The problem is really more that some places are designed to only be usable in a car, and that's what's poorly designed.
Speaking of which, also that suburbs aren't a bad thing. Again, it feels like people take the car focused design common in the US, and assume that's how these have to be. Thousands of identical houses in the middle of nowhere with everything interesting a 30 minute drive away.
But that's not necessarily true, and (as seen in most of the world), suburbs can be quite nice places to live in. A walkable one (like many in Europe) can be just as valid a place to live as a dense city neighborhood or rural area.
(You may call me 'Cassandra'.)
We all think we’re talking to each other constantly online, but what we’re doing should barely register as communication. And I very much include HN in this.
--
I believe that computer security is a solved problem, but very few people understand the solution, capability based security.[2]
I further believe that because of this insecurity of our computers, people avoid straying from "strange websites" and thus aggregate in the walled gardens.
I further believe that this insecurity will be the cause of losing the war for general purpose computation, and thus losing democracy, what little we have left of it.
It's quite possible this is all part of a plan. Multics was left to die on the vine, and isn't mentioned much any more, because secure computing would make the NSA's job harder.
--
I believe that Von Neuman computing is not the path to cheap Petaflops, but BitGrid[3] is. (It's my personal hobby horse, I may be biased)
--
I believe that AGI has likely already occurred, and we just don't know it yet. Either through MemGPT, or some variant thereof. [4]
It's my belief that during training, an LLM learns superhuman cognitive skills, but they aren't impedance matched to deal with the bag of words output model. Furthermore, it's handy capped by the small context window it has to work with after it's trained.
This is why we keep discovering new things they can do, after the fact.
--
[1] https://theportal.wiki/wiki/Embedded_Growth_Obligations
[2] https://en.wikipedia.org/wiki/Capability-based_security
A person who prefers to work in-office with other people who work in-office (no remote team) is not a terrible person who has no social life outside of work. They simply have a preference.
A company deciding to require in-office work is not evil. They’re making a decision that they think will pay off for them. The same as a company deciding to be remote and give up their office. One isn’t good and one isn’t bad.
I've been watching recent Nature episodes on PBS with my family (as I did in childhood, too), and the other primates, let's just say the lowland gorillas, lead enviable lives without deforesting or otherwise spoiling their home (we're doing that, sadly). We could lead meaningful, engaging lives following the old ways again, and everything we used and made would either decompose safely back into the environment or stay awhile longer as landmarks (earthworks and shaped stones, for example). I don't know how most of the HN participants think about this, but judging by the new and popular topics we as a subgroup aren't putting much effort into activism towards a seasonally-nomadic hunter-gatherer culture.
but now it's owned hard by other bots, and for a site that discussed GPT before it was cool it has little protection from the bots.
and, in general, agit-prop bot, esp. here, reddit, slashdot, etc. is through the roof
I also think that having a lot of libraries and frameworks to choose from is a good thing, not a bad thing. This is perhaps influenced by working in languages that have much smaller ecosystems.
It's very obvious that the reason for the rise in anti-Asian hate crimes in the US is because the US government constantly demonizes China and promotes anti-Chinese rhetoric and tacitly endorses anti-Chinese racism. There's also not much proof that China is committing genocide in Xinjiang.
Rust has extremely repulsive syntax, like C++ and Java had a baby.
General-purpose AI is unnecessary, but possibly cheaper than making many special-purpose AIs. Nobody will ever need AGI, and it's also not desirable.
- LLMs are incredibly useful
- Lisps and functional programming aren't profound at all