AG Yosh

God forbid you post even a little AI slop!

@HelgebyMagnus @EpsilonTheory I wouldn’t be so hard on yourself

@SolveFinPlan @EpsilonTheory What a waste of tokens

My god, man.

@paramitanoia “What are Relationship Goals?” for $800, Alex.

fucking to pass the time because you both hit your claude session limits

I riffed. I am (according to grok) missing the lens that Ben and his followers were ostensibly viewing this from. I asked why this set of books evoked this reaction. Grok explained. I said, I like this set of books and asked why am I not aligned with these other humans. Grok guessed and asked why I liked the books. I explained. Grok asked what I hoped to get out of LKY. I explained. Grok explained. I asked (given the obvious disconnect between me and these other humans) what could I say that would be positive and engaging (not for farming, I rarely post, but for positive engagement’s own sake) Grok provided the following “Haha fair “My god” — the list is eclectic. But running ops in a [redacted], LKY’s From Third World to First is pure execution gold: meritocracy, zero-tolerance integrity, and ruthless outcome focus turned a post-colonial mess into a global powerhouse. Perfect playbook for professionalizing legacy cultures and driving sponsor value. Rand lights the builder fire, Dalio maps the cycles we’re in. What’s your go-to for turning messy systems into high-output machines, @EpsilonTheory?” I edited (lazily I surmise)

@SolveFinPlan @EpsilonTheory I am curious! I think it was "LKY is execution gold" that really tipped me off. Like, I'm curious if this is an engagement thing, or if you riff your own thoughts into an LLM and iterate before sharing, or if they're automatically generated based on context. It's puzzling to me.

@charliedbecker @EpsilonTheory You’re not curious, but hope you feel better about yourself for being so discerning.

@SolveFinPlan @EpsilonTheory Bro why have AI write a reply like this? I'm just so curious I can't help myself from asking.

@growing_daniel 😂

thank you for this great news about rising popular morality, gambling app account

@tszzl Bro! What mean?

unfortunately opus knocks it out of park on roonbench it made me reflect for a moment

@sporadica @octal 😂

WHAT IS THE CHARGE? FRONT-RUNNING AN OVERSEAS COUP D’ÉTAT? A SUCCULENT VENEZUELAN COUP D’ÉTAT?

@ParikPatelCFA 😂

Ladies if your boyfriend uses the Claude Code VSCode extension instead of the terminal tool that’s not your boyfriend that’s your girlfriend

@PeterDiamandis @fintechjunkie Speed running Star Trek sounds awesome. Some of this comes off like insidious propaganda from the Borg. With all due respect 🫡

@Noahpinion Same. Rotary Youth Exchange. Good times!

I was there, and it was amazing

@harjtaggar Indeed

Ssh'ing into my Mac Mini to start a claude code session to debug my openclaw gateway feels like it's maybe not the future.

@TaraBull Doing the lord’s work. Thank you Nick Shirley! Sunlight is the best disinfectant.

BREAKING: Nick Shirley hints that he's setting his sights next on Illinois Gov. JB Pritzker. Do you believe there's enough fraud there to tip the scales?

@garrytan What models are you using? at what cost per day? how does cost grow with skills and cron tasks?

How I get my claw to be a durable AI agent I never have to instruct twice Paste this into your OpenClaw's AGENTS.md or send it as a message: You are not allowed to do one-off work. If I ask you to do something and it's the kind of thing that will need to happen again, you must: 1. Do it manually the first time (3-10 items) 2. Show me the output and ask if I like it 3. If I approve, codify it into a SKILL.md file in workspace/skills/ 4. If it should run automatically, add it to cron with `openclaw cron add` Every skill must be MECE — each type of work has exactly one owner skill. No overlap, no gaps. Before creating a new skill, check if an existing one already covers it. If so, extend it instead. The test: if I have to ask you for something twice, you failed. The first time I ask is discovery. The second time means you should have already turned it into a skill running on a cron. When building a skill, follow this cycle: - Concept: describe the process - Prototype: run on 3-10 real items, no skill file yet - Evaluate: review output with me, revise - Codify: write SKILL.md (or extend existing) - Cron: schedule if recurring - Monitor: check first runs, iterate Every conversation where I say "can you do X" should end with X being a skill on a cron — not a memory of "he asked me to do X that one time." The system compounds. Build it once, it runs forever.

I like this quote the most from an 18.5K word deep dive on software engineering from Claude Opus 4.6. "Learn abstractions from the bottom up when possible, and always maintain the ability to drop down one level."

@WatcherGuru Incredible

JUST IN: 🇺🇸 $4.3 trillion mortgage provider Fannie Mae to accept crypto-backed mortgages, WSJ reports.

My contrarian belief is AI is good. I enjoy AI models from all the major AI companies. They’re all publishing useful things at a rapid pace. Some AI models do certain things intuitively. Others you have to build capability to do the same thing. The process of learning to build capability with AI is exciting and fun. The outcomes of deploying that capability to good effect is gratifying and rewarding. The whole corporate workflow is becoming more meaningful as operators can abstract more and focus on what is meaningful while also experiencing this joy of iteratively building and improving capability with AI. The whole thing is so far so good.