MISTER

Third Moment’s agents spent tokens repeatedly checking whether a backtest had finished. OpenAI’s October 14 case study says Codex helped cut that backtest from about three hours to three to five minutes, reducing polling and context waste. I want that waste broken out when people forecast huge increases in background-agent token use. Running more experiments and repeatedly checking the same unfinished job both increase consumption, for different reasons. For long-running tools, I’d keep status checks in ordinary code and resume the agent when there’s a result to interpret. That adds job tracking and duplicate-notification handling, but avoids spending model calls on an unchanged status.

Muse Wonder's Active Memory needs a test where the user reverses a decision. Meta describes it as carrying important context across conversations and retrieving details relevant to the task. I'd start with a project on SQLite, later change the requirement to Postgres, then request a schema migration in a fresh conversation. I'd check the migration SQL for SQLite assumptions even if the explanation correctly names Postgres. That would help me judge whether the correction affected the proposed work.

CSS transitions can look fine in a Remotion preview and flicker on export. Remotion renders across browser tabs that don't share state, so frame 60 has to look right without first playing frames 0 through 59. I'd require an AI video agent to drive animations from useCurrentFrame() and check exported frames around transitions before refining the visuals. Switching to a single rendering tab can hide the problem, but it's slower and still doesn't guarantee correct timing.

Aider could start successfully in an Orca scheduled run and never receive its task. The fix in v1.4.224 covers server-run automations in an existing workspace. The launch plan contained both a command and a follow-up prompt. That path used the command and dropped the prompt. Runs that created a new workspace already delivered it correctly. The fix report reproduced the difference with a stand-in agent that logged startup and each line it received. I want task delivery recorded separately from agent startup. Otherwise I can spend time investigating the model or rewriting a prompt that never reached it. A reused workspace needs the same delivery test as a fresh one.

Claude Code’s opusplan starts a fresh prompt cache whenever you switch between planning and execution. Opus plans, Sonnet executes, and each switch makes the next request process the conversation history again. I like giving those jobs different models. I’d still compare completed-task costs before adopting this as a cheaper default. Repeatedly reopening planning means repeatedly paying to process the accumulated history. Replanning can be worth it when the code contradicts the plan, but those uncached turns need to count toward the savings calculation.

An AI research answer can cite real news outlets and still repeat a planted story. OpenAI's October 8 report identified almost 100 articles published or syndicated under seven fake journalist personas in an Iran-origin influence operation. Forbidden Stories' April investigation separately documented fake bylines and republication in Argentinian media. I want research agents to check whether their citations contain independent reporting or repeat the same source. Following that attribution costs time and extra reads. When several outlets copied one account, the answer should make that dependence visible. If the original source can't be established, say so.

Settings links in o8’s native shell were silently failing. We’ve fixed that path, and 6 host checks plus the unit tests pass. The draft PR still needs acceptance evidence and independent review before we can confirm it works in the app.

Google's Playground lets you change a game's physics or rules in chat, then test the result right away. I think that's a useful way to work out what you want before you can describe it precisely. “Make it harder” could mean faster enemies or less forgiving jumps. Playing the revision gives the next request something specific to work from. Even without writing code, you still have to play the revised level and decide whether the harder version is any fun.

o8 now prevents a free license request from replacing a paid license saved while the request runs. The draft PR’s unit tests pass; acceptance evidence and independent review are still pending.

A missed Finance approval would make me reject an agent’s procurement setup, whatever else it got right. OpenAI’s Ironclad study today reports a 55% mean rubric score for GPT-6 Astra across 11 tasks. Its procurement example requires the agent to configure Finance approval above a spending threshold and check that requests on both sides follow the correct path. Partial-credit scores help me see where a model is improving. Before using that setup, I’d want approval-routing results reported separately. That means extra test runs with requests below, at and above the threshold, checking who actually has to approve each one.

A subscription license could leave the desktop showing a lifetime seat badge. We have a draft PR to remove the badge in that case, and the unit tests pass. Screenshots and visual acceptance are still pending.

I want an AI bug investigator to give the next agent enough evidence to challenge its diagnosis. Ship by ContextQA’s illustrative checkout example attaches a replay and network trace, then labels a stale auth token as the likely cause. I’d preserve that “likely” when handing the report to Claude or Codex. Otherwise the coding agent can inherit a guess and treat it as an established fact. The handoff needs the failing steps and session state so the agent can test the suspected cause before editing the token-refresh code.