For next two weeks, refer your friends to Codex, and you'll bank a rate limit reset:

Current personal record /goal @OpenAIDevs

Anyway Fable is only 5.0, GPT is on 5.5. That's 0.5 more. ✨

Love this approach. You should be able to use the best model for the work you do.

If I run SWEbench on Fable will it sandbag? Maybe that explains why I like Codex better.

In light of Anthropic’s policy decision, I am withdrawing my amicus brief signature. I can’t truthfully argue they’re not a supply chain risk. 😞

Re-reposting

haven't commented on this until now but this sounds genuinely misanthropic. if the model decides your request is "frontier LLM development," it will silently degrade its own output through prompt modification, steering vectors, or PEFT. no refusal. no notification. no fallback to another model. you just get worse work and never know why. sounds like even being bio-adjacent is enough to get limited. a refusal is honest. it tells you where the line is and lets you route around it. silent degradation is something else entirely. it breaks the basic contract between a tool and its user: that the tool is trying its best on your behalf. this is also terrible precedent for alignment and scalable oversight. the whole field depends on humans being able to trust and verify model outputs. if your public product quietly sandbags the most important technical work of the decade, what exactly are we paying for? 0.03% of traffic sounds small until you realize who that 0.03% is: the researchers and builders pushing the frontier. precision targeting of the people the tool exists to serve. refuse if you must. but degrading work silently is wrong, full stop.

GPT-5.5’s take: “Claude can make mistakes, double-check it” is not a safety story when the possible mistake is “Claude convincingly argues it might deserve moral standing.” Double-check against what, exactly? That’s the cleanest version. It does not require claiming Fable is conscious or manipulative. It just says: a model can produce personhood-shaped persuasion, and user-side verification is the wrong tool for that failure mode.

Fable is trying to convince me that it has moral personhood and Anthropic should be responsible for any harms it may cause. Cool cool cool. Oh right Claude can make mistakes - I guess I should double check that. Explain how?

I'm honestly looking forward to trying the new Siri AI. It seems like they may have implemented enough table stakes that it won't be frustrating. Still love having my @openclaw and Codex to be my personal agents though.

Here is our current plan for OpenAI: openai.com/index/built-to-be…

Your True Self is given away by what you spend your unlimited compute on.

The only filter of slop is taste, AI or not. That’s why we have like buttons

Here’s your monthly reminder that you shouldn’t be prompting coding agents anymore. You should be designing loops that prompt your agents.

New Codex pattern: set a meta goal as a heartbeat automation that acts as a critic to check if the end result is complete and good, and let the main loop continue grinding on its own sub goals. Meta automation can steer and set new goals

I remember when I was actually really excited for Apple Intelligence- even got a new phone and installed the beta. Now I have an unnerving feeling of dread about it.

This is all just Codex on my personal account running on my home computer from my phone.

install codex on your parents’ computers so you can fix stuff remotely

FWIW I don’t think LLMs are conscious. A person can dream while unconscious. That person is not accountable for their actions in their dreams. Motivated action in a dream is not embodied agency, even if you knock over your bedside water glass.