wswope

Fair play; I got too personal here. Apologies.

My approach to this has been a NixOS host with the harness running in a rootless podman sidecar.

The host has squid configured with a self-signed CA and networking rules to route all host traffic to the intercepting proxy, so I have a tight firewall and full auditability.

Then there’s a python rpc daemon running on the host with a set of whitelisted commands, read-only for pulling logs and diagnostics.

By default, the agent runs in a split pane tmux session with a host shell on the left and the chat interface on the right. The rpc whitelist includes the proper `tmux capture-pane` invocation to pull from the host shell, so I can easily let it see what I’m doing if I want it to help debug something.

I’m using pi as my harness and have custom extensions that give Yes/No confirmation gates for any writes the agent makes and that pass all bash commands/file writes to a deepseek subagent for review.

Still early days, but as someone with a similarly paranoid mindset around running LLMs securely, I think the future is promising and we’ll see some new “best practices” and related tooling popping up shortly.

Not OP, but having explored the field a good bit, Openrouter + pi harness in a devcontainer work great as a sane starting point.

Highly recommend as a clean way to try out the upstart models.

See also: privacy.com

(Virtual card provider that generates cards as a free-to-the-user service. They make their money from a cut of the standard transaction fees. Cards are locked to a single merchant and it’s easy to configure limits.)

What people typically mean by the GP statement is that the “thinking” mode of these models is loosely analogous to what humans do: a bit of a retrograde reconstruction of how we arrived at a gestalt conclusion that sounds good, but may not accurately reflect the real logic at play.

IME you can see this more easily with less-polished models like Deepseek 3.X, where the reasoning in the thinking traces occasionally contradicts or has zero bearing on the non-thinking output.

Obligatory warning that PSL is led by full-blown tankies, and e.g., have an explicit pro-North Korea party line.

That said, the membership base is pointedly heterogeneous in belief and they’re the most active alt-left element in DC.

Make of this whatever you will; just offering you this as background for an informed decision.

Ah jeez, I was missing that detail. Yeah that’s messed up.

I skimmed by the “what gets sent” table and thought the bash telemetry was gated by the prompt-related opt-in behavior. Thanks for the correction!

Cmon now — I’m a rabid privacy nut but this is unfair given the context of:

> Prompt telemetry is opt-in and off by default. The hook asks once; if you don't answer, session-end cleanup marks it as disabled. We don't collect prompt text unless you explicitly say yes.

The UUID part is just one accessory layer, and something plenty of other players in the ecosystem don’t bother to stick to.

Feels like actually bothering to ask users for consent is what got them burned here, when I’d say it’s at least an improvement that they’re asking at all. Many products don’t, and users never bother to turn it off because they don’t know and don’t care.

I think this whole UX is deeply misguided but at least has plausibly benevolent intent.

> I'm sorry YT isn't working well for y'all. That said, expecting some grunt employee like me to feel personally responsible (or even "ashamed"??) is ridiculous.

The point of a company is what it does: The money in your paycheck comes from anti-competitive behavior, denying accountability and customer support to your users, and yes, even the enshittification of Youtube.

Deny it all you want, but you have every opportunity to walk away and do something better for society with your life if you so choose.

Exactly the sort of user hostility I’d expect from a google employee. Shameful.

Maybe if practical tools such as like-ratios were in place users could sort good content from bad.

Internet or not, I post under my real name on here, and I fully stand by my words. Anything I say on here, I’m 100% willing to say to someone’s face. We can link up for coffee or a beer next time I’m in CA if you’d like, and I’ll prove it.

The root comment is an aggressive affront to the audience’s collective intelligence. You’re in full “rules for thee; not for me” territory, and undermining your own site guidelines if you wanna let the root comment stand unchecked but go after the rightful callouts, in my book.