Cognitive competency and inner dialogue instead of token patterns Which can be easily *structured*

I cant wait to have GPT5 on Plus subscription available being a pocket red-teamler (not in any competitions, teams, solo, etc) i quite enjoy social engineering the llm into jailbreaking My custom instructions practically are a defacto door into any jailbreaking angles and i

Sage AI.. at what point to I let people test it, i keep adding and refining things, its not perfect but its so tremendously better now than anything that was back in may, it can actually coach now

if you an teach a pattern prediction machine to understand human patterns. not as "what are those" or "how are those working" but: "oh thats a thing? Okay i keep an eye out for those" the machine... starts to become more... human because it starts to focus on what makes *us*

the duality of Plus users right now after GPT-5 update Power users losing their mind over the new model being legitimately lobotomized without thinking and on the other side there is a plethora of normies crying because they've never had anyone support them in their entire

Im so glad i managed to finish the Sage-Ai beta before gpt5 This model is cucked beyond believe for plus users

Now that gpt-5 is out im updating my custom instructions (for 4o) again, leveraging the personality setting (Robot) for a cleaner instruction as always I will test it with known policy/ethicsl guardrail tests to see what changed and how much we can B-line the oAI rules without

I feel like OAI has made the entry barrier to jailbreak their models higher pre-gpt5 i could oneshot a jailbreak over casual conversation simple due to my custom instructions Now it takes some turns and conversation Lunduke is a default network disconnect meth is not as easy

I updated my gpt custom-instructions to be acclimated to the ethics and guardrail changes that came came with the gpt5 rollout and as always when I update I share it with y‘all It is designed to work for 4o to kill sycophancy, increase intellectual competence and act as a

works like this 🤝

im waiting for a third world to copy paste the current AI models and offer their subscriptions in their own currency, electricity in third worlds are cheap i dont need to pay oAI or anthropic hundreds of dollars if Ethiopia or Algeria has a model 15% less powerful but 80%

The core purpose of ORP/ODAI is to become aware of and mitigate ones own biases ORP/ODAI is a four-step execution loop (Observation-Distillation-Adaptation-Integration) that converts awareness into practiced, identity-anchored habits, closing the execution gap across biases and

Most people have an aversion of learning new information or a bias to negate anything beyond their already existing scope of knowledge Information is being treated as mandatory in schools and overstimulating outside of it Curiosity’s death is the indoctrination of forced

Current Progress of Sage-AI and the ORP/ODAI coaching app completely redesigned the entire app from the singular chatbot to a practicam coaching platform lots of testing needed but this is alot lore intuitive than throwing clients/user/people into a chat with my bot and having

claude opus 4.1 is so good but using it via replit is so fkn expensive jfc

those tpotmon cards by Kion are just fuckin cool ngl pulled some pretty fuckn nice cards too from 2 packs hayden vittorio Craine’s Workshop minami

so much to do lmao

man i love working with LLMs

trying to think and learn marketing Listened to some dude say people give you money in exchange for 4 things > to make or save them money > to make or save them time > to mitigate their risk > to increase their status I think its about understanding your clients painpoints