When NOT to use /goal in Codex
This transcript explains when not to use the /goal command in Codex, outlining three scenarios where it is inappropriate. It also defines the three key properties that make a goal well-suited for the tool: a durable objective, an evidence-based finish line, and a multi-turn path.
Summary
The speaker opens by cautioning that /goal is not a universal tool and should not be applied to simple, one-line edits. The reasoning is that /goal is designed for outcomes rather than outputs — it is too heavy a mechanism for trivial tasks.
The speaker then identifies vague or unmeasurable objectives as another poor fit for /goal. Examples given include 'make my customers happy' and 'refactor this code,' both of which lack a clear, definitive completion condition that can be reliably measured. Without measurability, there is no way to confirm the goal has been achieved.
Finally, the speaker outlines the three properties that make /goal most effective: (1) a durable objective that remains steady over time, (2) an evidence-based finish line that can be measured, and (3) a path that requires multiple turns of investigation to complete. When all three properties are present, /goal is the right tool for the job.
Key Insights
- The speaker argues that /goal is too large a tool for simple, one-line edits, emphasizing that it is designed for outcomes rather than outputs.
- The speaker claims that vague goals like 'make my customers happy' are poor candidates for /goal because they lack a reliable, definitive completion condition.
- The speaker identifies 'refactor this code' as a bad example for /goal, implying that goals without measurable endpoints are not appropriate for the tool.
- The speaker states that /goal works best when the objective is durable — meaning it stays steady over time rather than shifting during execution.
- The speaker argues that /goal is strongest when reaching the objective requires multiple turns of investigation, distinguishing it from tasks solvable in a single step.
Topics
Transcript
[0:00] Goals are not the right tool for every job. Do not use goal for something that is a very simple one-line edit. It is just too big of a tool for the job. Your goal wouldn't be like, "Make sure this line of code is removed." You really want an outcome, not an output, almost, for it to be a good goal. Also, don't use a goal when the finish line is vague. If you're like, "{slash}goal make my customers happy." I think that is just a very vague goal. It's very hard to measure, and there's no reliable, [0:30] definitive completion condition. And so, that's not very good. The other example they give is like, "Refactor this code."…
Full transcript available for MurmurCast members
Sign Up to AccessMore from How I AI
GPT-5.6's video editing via Codex is genuinely one of my favorite new workflows
A product manager describes using GPT-5.6 with Codex to automate video editing for social media clips. By simply dragging a file and providing natural language instructions, the AI generated five polished, fast-paced hype videos from a long conference talk recording, dramatically reducing the time-intensive manual clipping process.
Theoretically Intelligent vs. Practically Effective: Why GPT-5.6 Sol Beats Fable for Product Work
An executive contrasts two AI models (Fable and Soul/GPT-5.6 Sol), arguing that Soul is superior for product work because it prioritizes practical effectiveness over theoretical intelligence. The speaker values the ability to ship products to customers and understand end-user goals over theoretical sophistication.
Build a harness when the same workflow needs the same setup and the same outcomes, every time
Building a harness for repetitive workflows allows you to be more prescriptive about job execution, resulting in greater efficiency, consistency, and better outcomes. Rather than explaining requirements to an AI agent each time, a harness lets you use a simpler interface like pasting a link while the agent already understands the intended task.
GPT 5.6-Sol vs. Claude Fable: Why OpenAI’s new model crushes my benchmark
Claire Vo compares OpenAI's new GPT 5.6 models (Soul, Terra, Luna) against Claude's Fable using her custom "How I AI" benchmark, finding that GPT 5.6 Soul excels at practical product work, prototyping, and natural communication, while Fable is theoretically intelligent but pedantic and difficult to collaborate with.
Context offloading is an underrated AI use case
The speaker highlights context offloading as an underrated AI use case, where AI serves as a safety net for routine cognitive tasks like email management and personal finances. Rather than adding new capabilities, AI reduces anxiety about missing important information or making mistakes by handling monitoring tasks, thereby freeing up mental bandwidth.