A recent RPi is sufficient to handle a few plants - though, yes, recompilation will take time. The ESP is a beautiful piece of software, hence I highly recommend it. My native language has an expression that describes this situation perfectly: the appetite grows with eating. Next thing you know you have 2k or more entities, and your HA even handles some video feeds.
The important thing is that it's pretty much always easy to make an upgrade thanks to the good design of the backup system. Don't forget to set up backups in either case, it's a sin to not use such a complete system :)
The cool thing is that it's very easy to migrate it to better hardware. HA backup and restore system is highly reliable. For this reason I can definitely recommend an RPi to start with, and who knows perhaps it will be enough forever, but if not then moving is a matter of ~one evening.
I used to run HA on an RPi, but eventually migrated it to a similar NUC. The RPi eventually just wasn't powerful enough (peak compute needs), while the NUCs are still quite cheap. And you can run a surprising amount of Proxmox VMs and LXCs on barely a few cores and gigabytes of RAM.
This is how you can install GrapheneOS on these. Also, if you're wondering how does the security of something like this work: if you change the boot hash then the phone forgets all the hardware-stored secrets, for example the disk encryption keys.
I used to read forums for schizophrenics (self disorders fascinate me, look the term up if you want to understand schizophrenia), and it was the consensus there that out of all the recreational drugs cannabis caused the greatest deterioration in one's mental state. Those are generally fairly sick people, but I don't think one can just ignore this signal. I personally went catatonic once after consumption, and I'm not schizophrenic at all. And that was in Amsterdam, so it wasn't some trash spiced up by a 17-years old dealer with whatever he found in his grandma's medical cabinet.
This may be a good place to exchange some security ideas. I've configured my OpenClaw in a Proxmox VM, firewalled it off of my home network so that it can only talk to the open Internet, and don't store any credentials that aren't necessary. Pretty much only the needed API keys and Signal linked device credentials. The models that can run locally do run locally, for example Whisper for voice messages or embeddings models for semantic search.
Does anyone have an answer to the problem of an OS for a laptop? I'm thinking about strong security here, less so about privacy (which is doable, for example via a Linux distribution).
Because the patient is usually unable to handle such information correctly (the medical system sometimes too). And the whole-body-scan type of tests additionally pre-select for the high anxiety types.
Nah, it's the 50 tabs in my browser that I no longer have to open before I know the correct incantation to type. I bet someone knows how many tabs an average deveoper is keeping opened when working (like Google or someone with a nasty spying browser extension) - would be cool to see that plotted against the fraction of Github commits that are co-authored by Claude. I predict a fairly tight negative correlation.
Do you think some of them are honestly like that? I can never quite figure out how many levels of irony^H^H^H delusion there are. Spoken as a person that would totally have his job, but just because it most certainly pays plenty and is likely fun to do.
I think they don't mind that you don't trust them if you're considered bottom of the barrel anyways - as far as ad revenue goes. I don't mind being there myself.
I don't understand why people think they know why stocks move up or down. There are clear cases from time to time, but in general the market doesn't explain itself. The reasons may not even be explainable in a way that's comprehensive to a human.
Reminds me of this question - why did the USSR collapse? You can describe dozens of influences which acted all at the same time, but there isn't a one paragraph summary answer.
GrapheneOS does this by default - only power delivery when locked. Also it's a hardware block, not software. Seems to be completely immune to these USB exploit tools.
AIs are exceptional at sensing personalities from text. Claude nailed it here, the author felt so good about the "holy cow" comments that he even included them in the blog post. I'm not just poking this, but saying that the bots are fantastic sycophants.
A lot of this stuff doesn't work by changing people's mind on topic X, but rather by saturating the informational environment so that people declare epistemological bankruptcy. For example, one thing that you can quite often hear from a Russian that has been confronted with something unpleasant is "well, who knows what's true". This is usually not a figure of speech, not some kind of washing down of facts, but rather an accurate representation of their mind.
Between being fooled and being uninformed the latter is much more pleasant.