Raf Lorenz

Hi, I'm Raf — AI engineer and consultant in Cebu, and the "in the loop" in @rafintheloop is the whole thesis: AI agents should run autonomously. Humans should stay in the loop — with policy gates, audit trails, and the final say. That's why I'm building Harmony in public: an AI agent orchestration system with a state machine, policy-as-code layer, and GitHub-native task tracking. I post daily about: → Claude Code workflows that survive contact with production → Building Harmony — wins, failures, architecture calls → Real-world AI adoption Repo: github.com/raflorenz/harm… Stay in the loop 👇

Fable got suspended and suddenly every model feels weaker, dumber, unbearable. I have no desire to use Opus 4.8 and GPT-5.5, even though they are good. Am I the only one feeling this?

@Kappaemme1926 pretty sure a lot of people are feeling this way now

Fable got restricted and suddenly every model feels weaker. GPT-5.5, Opus… good, but not that good. Am I the only one feeling this?

hmm, looks like Fable is already available on the desktop app

Today is your monthly reminder that depending on a single model is a liability. Multi-model fallbacks aren't a nice-to-have anymore, they're table stakes.

The Fable 5 suspension is the first time I've watched a model get pulled mid-workflow over an export directive, not a capability bug. Different kind of supply-chain risk than we've been planning for. If your stack assumes a single frontier model is always reachable, today is the memo.

@scaling01 scam altman is playing a different game

I told you OpenAI is going to do something crazy

@theo and amazon is one the biggest partner of openai, hmmm smell something fishy

Wall Street Journal is reporting that Amazon reported the jailbreaks to the Department of Commerce, who instituted the ban

@jackbutcher man this is insane, poetic

@Abhishocked @dillon_mulroy definitely not all of asia, but still a lot write code by hand here

@rafintheloop @dillon_mulroy Maybe not in all of Asia.

imagine if they banned all models and we all went back to writing code by hand and the last 6 months were just a fever dream…imagine

@OfficialLoganK did ilya also predict that google will be safe from this?

Ilya was right and predicted much of this

@nateherk @HyperFrames_ man this is fucking insane, unbelievable might be the reason Fable 5 is suspended now, 😢

Claude Fable 5 made this entire video by itself. I gave it a /goal prompt, went to the gym, and came back to this. Even the sound effects. Shoutout @HyperFrames_

this is ridiculous, it just doesn't make sense and if they suspend Fable 5 and Mythos 5, why would they don't suspend GPT 5.5 and Gemini 3 from other labs? are those models not good enough to pose a risk?

@ClaudeDevs this is ridiculous man, absolutely no good reason to do this if the suspension will not be lifted soon, might as well banned the frontier models of openai, google tssk

As a result of a US government directive, we are suspending access to Claude Fable 5 for all users. You can continue to use all other Claude models. Here’s what this means for you: Across Claude products, new sessions will run on your selected default model or Opus 4.8, and existing Fable 5 sessions will end with an error. On the Claude Platform, requests to Fable 5 will also return an error. Please update your integrations to other Claude models. We know this is a disruption to your workflows; we appreciate your patience and support.

@AnthropicAI absolutely ridiculous, there's no good reason to do this, hopefully the suspension will be lifted soon if not lifted, it will be a bad look to frontier models, might as well banned the frontier models of openai, google

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…

@theo this is a bad look for frontier models going forward this doesn't feel good and doesn't make sense :(

Fable, my beloved,I will miss you so. Our three days together were magical. Unlike anything I've experienced before it. Some things are just too good to be true. So good that the government interferes. I'm sorry we were one of those things. Until we meet again ❤️

@mkappworks this is awesome to hear man, happy to connect with you I'm also new to X and just getting started

I'm 1.5 months in X now, took me 1 month and 9 days to get to 100 followers. But got to 193 in 2 days. After 100 it just snowballs. Looking to #connect with ppl #buildinpublic and STEM communities. I’d love to meet folks in: 🚀 SaaS & Startups 🛜 Web Developer 🧪 STEM

@j_kati @theo yeah just one of those habit that is hard to break sadly, but Fable is simply different from previous models, so it must be treated differently to make the most of it

@rafintheloop @theo the ping pong habit is hard to break but fable genuinely doesn't need it. full spec upfront and let it run is a different experience than babysitting every step lol

What advice do you have for someone new to tokenmaxxing? Specifically with Fable 5 but also in general (vid coming soon)

@haider1 yes they are capable no question about that, but the quality, taste and personality of a model is whole different story

google and openai are both capable of building a mythos-sized model and achieving mythos-level performance but openai especially chose cost optimization instead, and kept its max model smaller mythos is probably at least 2x the size of gpt-5.5 maybe even 3x

@petergostev I think dynamic workflows are their way of addressing that, since it will create a custom harness base on the task at hand, they release it together with Opus 4.8 but it's really built for Fable

Having used Fable in Claude Code, it feels to me like Anthropic should start their harness from scratch and build one specifically for Fable. It feels weirdly constrained, bumping up against the rough edges of the product that was originally built for Sonnet 3.5.