These are two different layers in the stack that do two different jobs. Agent to agent communication across remote machines is definitely lubricated with tailscale, I’m not sure what value moving up the stack would bring at this stage?

My read here isn’t about connecting to remote machines in the sense of ssh, but presenting a unified view across herdr instances running on multiple remote machines.

This stinks of the same enshittification moves Docker made 10 years ago.

Tailscale's connectivity solutions should allow private connection between agents and between human-agents, by layering on top of the existing connectivity substrate. (I was stating "general impression", and I haven't looked into details yet.)

What I need is to have a general solution to connect me to my agents running somewhere. The motivation is that any hosted solution deprives me of ownership of my data. And I am planning to share more private data with my agents, that cannot be done to have hosted agents today. (I plan to build a hosted solution myself where users can declarative define their own agents, and completely own their own data without my service save anything, I'll ask users to connect to their own storage solutions like Google doc)

I of course can build everything myself. But I think a tailscale offering could be very convenient.

I am guessing cloud flare should have building blocks of their own, but it's a different layer than tailscale's.

The ad in itself probably has no practical value. The surprising nature of its staying power is curious. And curiosity is a strong signal. Good luck.

100% convinced their raw output is intended as further inputs, and my workflows have been comfortable and efficient treating it as such. If you really need to read slop, you ask your agent to give it to you in a style that works for you. I can imagine a world where the slop from others doesn’t hit us directly but gets personal mediation.

The only time I ever used it (as a turk?) was when Jim Gray went missing at sea and satellite imagery of vast regions of the pacific were fed through Mechanical Turk. Seems a trivial problem now, but was not then, and needed humans to take a look.

Well, if they didn't want a dominant investor, they didn't get one, they got acquired.

I've been running a desktop Mac in addition to MacBook and iPad for the past decade (currently a first gen Studio still holding quite strong). It's primarily a focus thing, secondarily it's a way to keep work happening locally while I'm elsewhere, and never have the frustrations of docking and screen geometries all re-arranging, etc. The iPad is at the opposite extreme, it's for recreation, burn out management. The MacBook exists for when I actually need to do "real work" outside of the house.

At this point we should be sharing prompts not code for solutions like OpenLogi. Reverse engineering many things can typically be one-shotted (ignoring things like.. capturing a pcap or whatever to aid the effort). Let others add their own taste to the solution.

Isn't it wasteful to have everyone go off and remake the same software with slightly different bugs in it instead of one that fixes almost all of the bugs

Terribly wasteful, yes, though perhaps if the idea caught on the major labs could stand some chance of long-term profitability.

Seriously, I wonder how many tokens are generated (wasted) every day on someone recreating another agent memory, context management, code graph, etc.

I can’t wait until Qwen 4.5 27B or Gemma 5.

It then becomes a roll of the dice if someone ends up with working software.

You're forgetting the other half of using the software you created - you still need to maintain it!

Please no, I couldn’t care about your prompts, there is nothing special about them. Anyone can look at the generated artifact and reverse engineer their own prompt in no time

The grey beards in the mid-90s cut their teeth on much more constrained systems! The web crawling these days probably has all the ones who have passed spinning rapidly in their graves.

Paper companies would rejoice at the idea of OpenTelemetry

I’m either out of touch with contemporary terminology but Opus 5 dropped “pre-mortem” on me today. Figured I’d just figure it out with more context.

It's project management jargon: https://en.wikipedia.org/wiki/Pre-mortem

Modifying those games were the first times I ever touched code! There was one that existed somewhere between DOS 2.0 and 3.2 that simply asked your name, then asked how the hell you got that name. It was maybe half a dozen lines, but easy enough to modify output and learn a bit about variables.