Anthropic just dropped a 100% free course on Loop Engineering with Fable 5.
This is the clearest breakdown of Claude Code and agentic loops you'll find anywhere.
People are paying for tutorials that teach less than this one hour does.
Watch it today, then read the step by step guide on building loops below.
I talk to engineers at other companies every day and hear the same thing: one person is 10x'ing their output with Claude but the rest of the org hasn't caught up.
Watching teams adopt AI, I keep seeing the same 4 steps.
I mapped them out here: Steps of AI Adoption https://t.co/kQnRAUMKpP
🔸 Olet, the Beloved Bird of Saint Paisios 🐦 🔸
St Paisios wrote this letter to the nuns of the Monastery of St John the Theologian in Souroti in 1975:
“In your last letter you sent me an icon of Adam and the animals in Paradise. So I thought I would send in my turn the drawing of a bird, my closest friend… I have named him Olet, which in Arabic means ‘child’… Every afternoon I bring him goodies and treats. As soon as I give him something to eat, he takes a little and leaves… When I go to leave he walks behind me at a distance of about 100 meters and I, so that he will not continue coming behind me and get tired, leave him a crumb so that he may occupy himself, and I leave quickly so he will lose me.
Lately he has abandoned his asceticism and seeks good times! He neither eats broken rice or soaked dry bread, but only worms, which he wants me to put on a “plate” - the palm of my hands - and he climbs up there and eats. Progress! One can say: ‘Why do you make exceptions for Olet? Why don’t you do the same for other birds?’ I respond: ‘When I call for Olet to come, he brings with him other birds, friends of his, who run right to the food, but Olet comes out of obedience and love’…
And now that the weather has turned beautiful and he finds bugs to eat, when I call he still comes out of obedience, even though he is full and not bothered by hunger. Well, how can you not rejoice more for this philotimo bird than the other birds? Many times I am moved by such great love that I want to squeeze him in my clutch, but I fear that I will be like the monkey which out of love squeezes its child and in the end suffocates them. So I clench my heart, and I rejoice for him from afar, so I will not harm him.
One day I was late to go to the hill and Olet, because he was chirping a lot, had relaxed early. I left his food and departed without seeing him. The next day I left to go very early, because I was worried that a hawk had eaten him… When he saw me he was like a crazy person because of his joy. I gave him to eat, but he wanted company more than food. I marvel at his asceticism and the love he has, as well as his gratitude. Pray that I may imitate his virtues.”
up . . .
“Laziness throws us down even from heaven, while discouragement hurls us down even to the very abyss of wickedness. We can quickly return from there if we do not become discouraged.” - St John Chrysostom
@derrickcchoi I've been out of the OpenAI eco system for quite sometime. Can you tell me how I'd be able to to do this? Is the codex voice feature still rolling out?
I put GPT-5.6 Sol to the test by building a range of websites and UIs.
For each one, I brainstormed the idea with Codex using my voice, then asked it to build the whole thing in a single shot.
🧵Here are 15 examples:
New in Claude Code: /checkup
Run /checkup to:
1. Clean up unused skills/MCPs/plugins and save context
2. Dedup your local CLAUDE.md against the checked in CLAUDE.md
3. Break up root CLAUDE.md into nested CLAUDE.md's + skills
4. Turn off slow hooks
5. Update your Claude Code to the latest version
6. Enable auto mode by default
7. Pre-approve frequently denied read-only commands
.. And a few other goodies.
/checkup confirms with you before making any changes. Enjoy!
I don't get the codex hype posts. I was using openai with co-pilot for all my workflows before I tried and went all in with anthropic. It's going to take some serious convincing to get me to use openai. I've spent so long learning and implementing anthropic in my workflows that it's here to stay.
Anthropic was the first to market, and they locked down their clients by gating anthropic specific workflows (Claude.md vs agents.md - skills vs w/e openai uses). I've built all my skills to delegate through anthropics models. I've built cost analysis all on top of anthropic... It's constantly gotten the job done and impressed me way more than openai. Openai is late and I'm already full deep with Claude.
IDK what openai will have to do to get me to try the new models, but they're going to need to hit some serious benchmarks for me to care.
A few patterns we frequently use with Fable 5:
Use Fable 5 as an "advisor."
An executor (Sonnet 5) calls Fable 5 for guidance.
Most tokens are billed at the lower executor rate.
The "50% of weekly usage" line on Fable 5 tripped up a lot of people so here's what it actually means, it's not a separate Fable bucket you get on top, it's a ceiling on how far Fable alone can draw down your existing shared weekly pool, so half your normal allowance stays reserved for Opus and Sonnet and the other half is the most Fable can eat before you're into paid usage credits, and since it drains that pool faster per message than Opus does you'll hit the ceiling quicker than the number sounds, and the catch is Claude won't warn you inside the chat window so if you want to actually watch it you go to Settings then Usage where the progress bars for your session and weekly limits live, so check it before you go deep on Fable and you won't get surprised mid-task
Anthropic pulled Fable 5 three days into a two week window over export controls, and the makeup they're offering is one week at half usage, which somehow adds up to less than the thing that got interrupted, and I get that the original suspension was out of their hands but the redo being stingier than what it's replacing is a strange look, what am I missing here