I agree in the short term. But in the long term, the owners of the compute will become disgustingly rich without a wealth tax, nationalization, or local AI becoming competitive and ubiquitous. There are still problems to solve; the alternative is an absolute oligarchy.
Tokens are not fungible for hosted models but there are resellers like Github Copilot and OpenRouter. These allow you to purchase capacity and choose which hosted models to burn your tokens on. Obviously there is some margin being taken in order to provide this flexibility, but there's probably some offset with volume contracts.
Sounds like parent was using async to mean that JS has single-threaded concurrent functions by default.
Disclaimer: I'm just the messenger here so you might hate what follows, the community is choosing the words.
Concurrency as a keyword is increasingly being used to differentiate single-threaded event handling from parallel processing (threads). As long as event handling is quick enough, it can give the illusion of parallel processing. And in some cases it can outperform parallel processing because the cost of context switching between events can be lower than context switching between threads. See https://medium.com/@caophuc799/nginx-architecture-and-why-ca...
The async keyword, in practice, simplifies the callback boilerplate required to handle the event model. When used in communication, async is increasingly being used as a word to describe "Single-threaded Concurrent Functions Built On Hidden Event Handling", because there is no succinct way of saying STCFBOHEH in English.
So: JS has async/STCFBOHEH/context-switching on its functions by default whereas Java's functions are (by default) executed in order without async/STCFBOHEH/context-switching.
And the I stands for Internet. I guess we're considering local BT networks to be "the internet" now.
Perhaps the BT connection bridged the device to the internet, but we have no indication of that and the internet was used as the vector. It could have been a local wireless software update interface so maintenance doesn't have to open up each pump to apply the patch.
Awesome, thanks! I am in fact guilty of using logrotate but thankfully I haven't been burned yet. Although perhaps with the advent of containers piping logs to streams I've unknowingly absolved myself.
Also, you have an amazing website. Looking forward to reading more.
What's your recommended alternative to the newsyslog style of external logfile rotation? I'm not much of a sysadmin but it might be useful to know at some point. Thanks in advance!