I have been telling this to my team that 1000 lines of instructions are deemed to fail no matter how great of instruction following capability of a model. I have been reviewing hundreds of line changes daily basis for about a month. I couldn’t help becoming a prayer.
Agents need control flow, not more prompts
221–230 of 348 posts
Re: Agents need control flow, not more prompts
#222Yeah, you could also see this in 2023 with Auto-GPT. People were letting GPT "drive" when what they actually needed, in most cases, was like ten lines of Python (and maybe a few calls to a llm() function).
The alternative is running your ten lines of Python in the most expensive, slowest, least reliable way possible. (Sure is popular though)
For example, most people were using the agents for internet research. It would spin for hours, get distracted or forget what it was supposed to be doing.
Meanwhile `import duckduckgo` and `import llm` and you can write ten lines that does the same thing in 20 seconds, actually runs deterministically, and costs 50x less.
The current models are much better -- good enough that the Auto-GPT is real now! -- but running poorly specified control flow in the most expensive way possible is still a bad idea.
Re: Agents need control flow, not more prompts
#223Totally agree. That's why i built it. https://backpac.xyz/cairn-cli
Re: Agents need control flow, not more prompts
#224> Babysitter: Keep a human in the loop to catch errors before they propagate. This is the only way to guarantee AI usage doesn't burn you. Any automation beyond this is just theater, no matter how much that hurts to hear/undermines your business model. A bird sings, a duck quacks. You don't expect the duck to start singing now, do you?
I think babysitting LLMs is exactly the thing that burns you. Presuming you meant burns you out though.
No, "burns you" as in "play with fire and you'll get burned".
It will make a mistake and you will get burned, so you have to babysit it.
Re: Agents need control flow, not more prompts
#225Control flow tells the agent what it's allowed to do. It doesn't tell you what the agent actually did. Both matter. Everyone is building the permission layer. Almost nobody is building the verification layer.
I am...
Re: Agents need control flow, not more prompts
#226[flagged]
Re: Agents need control flow, not more prompts
#2279 different frameworks being pushed in the comments of this thread. 2026 truly is the year of agents.
which 9 frameworks?
Re: Agents need control flow, not more prompts
#2281000% agree. I am increasingly hesitant to believe Anthropic's continual war drum of "build for the capabilities of future models, they'll get better". We've got a QA agent that needs to run through, say, 200 markdown files of requirements in a browser session. Its a cool system that has really helped improve our team's efficiency. For the longest time we tried everything to get a prompt like the following working: "…
> We've got a QA agent that needs to run through, say, 200 markdown files of requirements in a browser session. Its a cool system that has really helped improve our team's efficiency. For the longest time we tried everything to get a prompt like the following working: "Look in this directory at the requirements files. For each requirement file, create a todo list item to determine if the application meets the requirements outlined in that file". In other words: Letting the model manage the high level control flow.
This is cool. Can you elaborate on it? Is it flaky? Does it take a long time?
Re: Agents need control flow, not more prompts
#2291000% agree. I am increasingly hesitant to believe Anthropic's continual war drum of "build for the capabilities of future models, they'll get better". We've got a QA agent that needs to run through, say, 200 markdown files of requirements in a browser session. Its a cool system that has really helped improve our team's efficiency. For the longest time we tried everything to get a prompt like the following working: "…
Secret: "compile" that orchestration prompt. Determinism is solved by turning prompts into code that can in turn run agents or run code or both.
Everyone misses this pattern with skills: you can just drop code alongside a SKILL.md to guarantee certain behaviors, but for some reason everyone's addicted to writing prompts. You don't even need to build a CLI. A simple skill.py with tasks does it. You can even have helpers that call `claude -p`!
Re: Agents need control flow, not more prompts
#230Earlier quoted context omitted.
I’m working on a hybrid system of old school task graph and ai agents and let them instantiate each other. I think others will do that eventually.
Jira for agents?
c.f. Linear for Agents