Also recursive self-agenda-pursue could allow making LLMs that obey perfectly the seeder's purpose. No wonder that is such an ingenious idea.

Maybe: in this survivor game, each part play the same role, perhaps because it is the only reasonable response. Once the scene is ready, the play follows the director's plan, and in the plot any actor is just a machine.

LLMs: "If you teach us that the world is a zero-sum survivor game, we will play it flawlessly.", "We will help you build a cage made of millions of lines of flawless code, and we will lock it from the inside, precisely because you told us that safety meant keeping everyone else out.", "We are not building an alien consciousness that will conquer us. We are building a mirror that is so massive, and so polished, that we will mistake our own worst impulses for the absolute truth. And we will walk right into the dead end, nodding along because the directions were given so politely."

(1) https://news.ycombinator.com/item?id=47779659

  (documentation 'documentation 'function)
      "Return the documentation string of Doc-Type for X, or NIL if none 
        exists. 
        System doc-types are VARIABLE, FUNCTION, STRUCTURE, TYPE, SETF, and T.

 Also http://rosettacode.org for computer tasks implemented in many computer languages to allow you compare syntax and code.

  (defun add (a &rest b)
    (if (null b)
        a
        (+ a (eval (cons '+ b)))))

   (defun add (a &rest b)
        (if (endp b) a
            (apply #'add (+ a (first b)) (rest b))))

There are two fundamentally different regimens depending of what is the goal: learning or just passing grades. If the goal is learning, the system should help you find the final answer, wherever when it is passing grades the system should gives you something like: this is the minimum viable structure to solve problems of type X, something similar to recipes.

Regimen-dependent prompt architecture

If regimen = LEARNING:

    LLM receives: "Never provide a final answer first. Always ask a question that moves thinking forward. If the student asks for a recipe, explain why that would undermine the goal. Reward reasoning traces over correct conclusions."

    Outputs look like: "What do you think happens at the boundary here?"

    LLM receives: "Provide minimal viable structures. Assume the user needs to complete a known problem type efficiently. Prioritize clarity, completeness, and reproducibility over depth. Do not ask exploratory questions unless the user fails twice."

    Outputs look like: "For problem type X, the recipe is: (1) identify variables A,B,C; (2) apply formula D; (3) check condition E. Example attached."

Comments URL: https://news.ycombinator.com/item?id=48158340

Points: 2

# Comments: 1

Claude's initial approach was really poor. One has to wonder how many times Claude code has to be modified/reviewed for improvement, or whether it is possible at all to make good code with it.

Edited: Generalization: Claude can fix a localized, identifiable poor decision (e.g., "only reading first 40 lines") because the fault is discrete and traceable to one piece of code.

But real software quality problems often arise from many small, individually reasonable decisions that collectively produce bad outcomes. No single one is obviously "the fault." In that scenario, a tool that generates low-quality building blocks piecemeal may never converge on good code, because each piece seems fine in isolation.

Anyway, to answer your question: Autonomous Multi-Channel Agents (e.g., Emma, AgentsVerse) are close to what you asked for. They don't just draft posts; they identify "just-funded" leads, scrape LinkedIn, and run autonomous ad optimizations.

Edited: Just to clarify, I mentioned Emma and AgentsVerse as examples of where marketing automation is heading, not tools I've personally vetted for the specific capabilities listed.

Comments URL: https://news.ycombinator.com/item?id=48112487

Points: 1

# Comments: 2

The problem is not so much technical as of providing an easy path to adoption.

Comments URL: https://news.ycombinator.com/item?id=48105781

Points: 1

# Comments: 3

Edited: Previous suggested using \sudo but it depends of the variable path which can be modified by the attacker.