Hello {{first_name|Motivated and Miffed Community}},

OpenAI launched something this week called ChatGPT Work, and the name is doing a lot of lifting. This is OpenAI's bet that the chat window was a phase — that the real product is an agent that takes a goal, disappears for a while, and comes back with finished files. Today's issue unpacks what actually shipped, what's underneath it, and the part nobody's saying out loud.

→ Here's the catch with agents that work for hours: they still need you to pick the work. The Six is my free tool for deciding your six tasks before the tools start running. Pick your six →

The Move: Run a five-minute pre-mortem before you delegate anything — to a person or an agent. Ask: "It's two weeks from now and this went sideways. Why?"

The Evidence: The technique comes from decision researcher Gary Klein, building on research by Mitchell, Russo, and Pennington showing that "prospective hindsight" — imagining an outcome has already happened — increases your ability to identify reasons for it by about 30%.

Start Here: Before your next handoff, write down the top two failure reasons you imagine, then add one sentence to your instructions addressing each. Under five minutes.

From elections and inflation to sports, tech, and more, Kalshi lets you trade on the real-world events you already follow. Buy “Yes” or “No” contracts based on what you think will happen, then earn returns if you’re right.

Pick a market, make your prediction, and put your knowledge to work.

Trade $25, Get Up to $500

_{Bonus credit varies from $15 to $500. Terms apply.}

OpenAI introduced ChatGPT Work, an agent inside ChatGPT that gathers information across your apps and workflows to create finished materials — sheets, slides, docs, and web apps — and stays with complex projects for hours by breaking them into smaller steps and completing them independently. It connects via plugins to tools like Slack, Microsoft Teams, Google Drive, and SharePoint, and Scheduled Tasks let it keep projects moving when you're away — turning new team messages into updated docs or slides, for example.

The structural shift is the deliverable. You've been the one carrying context between your tools and pasting outputs into their final homes (like a very tired courier). This moves the finish line inside the product — the agent does the gathering, the drafting, and the formatting, and you review the artifact. The tell is where it came from: Codex started as a coding agent, and OpenAI says more than a million people were already using it for non-coding work — the users productized this before OpenAI did.

Why it matters: The AI companies are no longer competing for your questions — they're competing for your job's output.

Under the hood is GPT-5.6, which is its own story. OpenAI released three versions: Sol, the most powerful; Luna, optimized for speed; and Terra, balanced for everyday use. Sam Altman's pitch was efficiency — he says the model is 54% more token efficient on agentic coding, noting that "every enterprise now is thinking about spend and the value they're getting in exchange for AI."

The quieter story: the rollout was reportedly staggered after a request from the Trump administration, and Altman described making "many changes" following a "collaborative back and forth" with the government — the kind of pre-release engagement policymakers have also had with other frontier labs. Frontier model launches now have a review step that looks less like a product calendar and more like a regulatory one.

Why it matters: The release schedule for the world's most capable models now runs partly through Washington.

The launch came with housekeeping that signals strategy. OpenAI is merging the Codex app into a new ChatGPT desktop app, with the old one renamed ChatGPT Classic, and the standalone Atlas browser is being sunset, its agentic-browsing learnings absorbed into ChatGPT itself. Pricing follows the Codex pattern: usage-metered, so longer tasks consume more of your plan — a seat license is becoming more like a utility bill.

Read the map, not just the launch: this lands months after Anthropic's Claude Cowork, and Microsoft has rolled out Copilot Cowork — every major lab is now converging on the same shape: one surface, agents underneath, your files and apps as the raw material. The experiments (separate browser, separate coding app) are being folded back into the mothership.

Why it matters: When every lab consolidates into one agentic surface, the moat becomes whichever one already lives inside your workflow.

OpenAI also introduced GPT-Live, a voice model meant to replace ChatGPT's existing voice mode. It can listen and speak at the same time — and it shows it's paying attention with acknowledgment phrases like "mhmm." We have officially taught the machines active listening (some of your coworkers still haven't unlocked this feature).

Why it matters: The interface goal has shifted from "answers questions correctly" to "feels like a person" — and that changes what we'll tolerate, and expect, from software.

→ The agents can run for hours now. Deciding what they run on is still your job — The Six makes that a two-minute morning habit.

If this issue clicked for you, these might too:

This week rhymed: coding agent → work agent → your job description. The tools stopped waiting for your next message.

Stay MOTIVATED,

Gio