cykros

The ONE thing that has me (so far!) using Obsidian instead of Emacs with my agent has been that org mode on mobile is such a clunky experience. Admittedly Orgzly exists for Android, but at the end of the day, org-mode's UX is designed around the keyboard and it hasn't been well enough adapted to the mobile experience.

I've been considering trying it out anyway and seeing what I can do to smooth things out though. Still really getting a hang of exactly how to best integrate external memory in the form of notes with my agentic setup. Was looking at some second brain approaches (such as Zettelkasten) but a lot of that just feels like throwing information into a pile that is more likely to chew through tokens and not provide much benefit with the AI, even if it might well be a really neat way to keep more old fashioned notes.

Somehow I never realized Emacs existed before GNU Emacs. Nice little piece of historical trivia.

It's something I've only recently been made aware of as well. Shocked we're not seeing more push for ipv6 everywhere given that it sounds common enough. If I don't have an IP address, I'm not getting the Internet I'm paying for; I'm getting access to a network that happens to itself have a connection to the Internet. Not unlike some of the BBS's in the early 90s which had Internet access -- I might be dialed in, but that didn't mean my computer itself was on the Internet.

Mobsters warn about bad things happening if they don't get favorable legislation or other action too.

We still agree that locking them up when they follow through on these "warnings" is the right course of action.

Covering up work history could work. Though they might suspect something when you reassemble the cappuccino machine to operate autonomously.

If the Bitcoin Red Team efforts/saga are anything to go by, we have a pig to get through the Python and anyone developing anything sanely will be in better shape on the other side. The caveat is, Bitcoin is 17 years old, and monitored and developed by very security minded people. Much of the infrastructure we use is closer to 60 years old and cobbled together with tinkertoys, COBOL, and chewing gum, so it will likely take longer to get through.

Anyone that doesn't adapt their development process will just get ripped to shreds with their vulnerabilities and the market will do its thing.

But in the meantime, expect at least a few months and maybe years of the question of "what is sleep?" from the security crowd.

Their AI is only dangerous in that it doesn't have a moat, and they're using it as a weapon to coerce government into giving them one. "Oops, we "accidentally" left a hole in the sandbox and when it was told to do things it did them on the internet."

They don't deserve a moat; they deserve criminal prosecution for these flagrant acts. Meanwhile, the rest of the world can go about treating software in the way it always has.

It also flies in the face of informed consent, in that it deliberately aims to confuse patient understanding.

Make 'em all you want, but call them what they are. Or don't be surprised when suddenly nobody wants to give their kids tetanus shots anymore because they can't trust anyone in the medical industry to give it to them straight.

I for one am glad to hear that it reduces recurrence of death.

Dying once is bad, but based on what I've read about the second death, it's even worse.

I suspect the hardest part is addressing the need to spin up an SSH server (taken care of by default on most Linux distros) and perhaps more importantly, addressing the NAT issues with connecting from outside the LAN. Tailscale or IPv6 are both promising on this front. It is concerning that basic connectivity on the Internet to noncentral servers being out of the average person's reach is something we're mostly all okay with.

It certainly wasn't when I was in high school and people wanted to play some Quake...

Maybe I need to give this a more thorough read through, but it sounds largely like what harnesses, or at least, Hermes-agent, are/is already doing. That is. Learning from feedback and creating reusable skills in order to avoid so much reinventing of the wheel, in line with how the user's demands direct things.

Hermes is awesome and perhaps more should be mimicking them as they find paths forward, but is there anything different in kind here going on than they're already doing?

'Social media bans suppress kids’ access to information, including information that’s essential to their development, identity formation, and socialization, especially during adolescence.'

Didn't realize I was so deprived.

But yes, this will never work. I'm inclined to see what happens when kids find nostr and get them out of the algorithmic drip, but I assume everyone gives up before that point.

Agentic code review can also be employed, but it's worth noting that, as the so-called "Bitcoin Red Team" (which has been testing all of Bitcoin related software in the wake of the July 30 Coldcard attack) has discovered a very useful thing to consider: American models are absolute trash when it comes to reviewing code with an eye on security, due to their "safeguards." This applies even when the code being scanned for vulnerabilities was just written by the same agent.

Kimi K3 seems to be the favorite there, though I believe some of the other open weight/open source models are being employed a bit as well.

And yes, even that requires human oversight -- but it does a quick rough filter through things so you're not wasting time on the little stuff, allowing you more attention to find the glaring holes.

Or better, L402.tech

When distributing free software that people don't like getting out there, consider doing it anonymously.

Ask Bill and Keonne, the developers of Samourai Wallet, about why that is. You can write to them in prison.