Turn Any Product Photo Into a GIF in ChatGPT

Today I've got a 10-minute way to turn any product photo into a looping GIF, the December change that makes ChatGPT and Codex share one allowance, Cursor putting your app online from one prompt, a 15-

Happy Tuesday,

Let's get into it.

In This Email

» The Big Thing: make a looping GIF from one photo in ChatGPT, on every plan

» Worth Your Time: one shared ChatGPT and Codex allowance, Cursor live apps, cheaper Google video, faster Claude agents

» Quick Hits: a package manager for agent setups, Copilot in Xcode, and GPT-6 Cyber

» Skip It: Inkling 2 Small's "open weights"

» Here's My Take: ship the thing, then check the meter

ChatGPT Images 2.6 Makes GIFs, and It's on Every Plan

OpenAI released Images 2.6 on Nov 4 in ChatGPT, Codex and the API. It generates images "50% faster than before," per OpenAI, and it now makes animated GIFs: a brand mascot, a reaction for social, or a product shot turned "into an eye-catching loop."

Where you get it: OpenAI says it's "available globally today in ChatGPT and Codex across all subscription tiers" on desktop, mobile and web. That includes Free.

For developers: the API gets GPT-Image-2.6 Flare (fast, "as little as six seconds" at low quality) and Sunburst (higher quality, more consistent across edits), plus SVG export.

Why you should care: a moving product shot gets noticed. Set aside 10 minutes for this workflow, not a guaranteed render or completion time.

Pick one clean photo. A product on a plain background, your logo, or a mascot works best.
Upload it to ChatGPT and paste this prompt.Turn this product photo into a short seamless looping GIF. Keep the product exactly as it is: same shape, colors and label. Add one simple motion, a slow 360-degree turn on a soft studio background. The last frame should flow smoothly back into the first. Preserve the existing product label and logo. No added text, logos, captions or borders.
Fix one thing at a time. If the label warps, reply with one change, like "Keep the label sharp, slow the turn down."
Use it this week. Ask for a square version for social, then put the GIF at the top of your next launch email or post.

→ See OpenAI's Images 2.6 examples here

1. Codex Will Start Spending Your ChatGPT Allowance in December

OpenAI's Help Center says that starting in December, Plus, Pro and Business plans get one usage allowance shared across ChatGPT and Codex, measured in tokens. More advanced models and complex tasks use more of it, and "Instant chat will not draw from your usage." Separately, TechCrunch reports OpenAI is merging ChatGPT's Chat and Work tabs, starting with select users.

Do this now: look in Settings, then Usage, when available. The combined allowance starts in December, not today. Once it appears, note the remaining usage and reset times to decide which jobs need a cheaper model.

2. Cursor Now Puts Your App Online, Not Just on Localhost

On Nov 5 Cursor launched Build Live Apps. Tell a cloud agent you want to build an app, and Cursor scaffolds it, spins up a Postgres database if needed, puts sign-in behind Cursor SSO, provisions hosting and gives you a live link on a domain you pick. Nothing goes live until its security checks pass and you click publish.

Try it: start a cloud agent and say it plainly:

The catch: it's for paid plans with cloud agents turned on. Apps are shared behind Cursor sign-in, and only individual plans can share them publicly. Repos from Cursor Origin and GitHub work today, GitLab is "coming soon."

3. Google's Video Model Got 30% Cheaper and 50% Longer

Gemini Omni 1.2 Flash went generally available on Nov 5 in the Gemini API, Google Cloud and Google Vids. Clips now run 3 to 15 seconds (up from 10), it costs $0.07 per second of 720p video (30% less than Omni 1.1), and you can edit the same clip in conversation instead of starting over.

The math: a 15-second 720p ad is 15 times $0.07, about $1.05 in the API. Test five ad variants for around five dollars before you pay anyone to shoot one. Every output carries an invisible SynthID watermark.

4. Long Claude Agents Answer Faster With One Parameter

Anthropic put message threads in public beta on the Messages API. Add the thread parameter, send the full conversation once, then send only new messages plus the previous response ID. Anthropic says requests with 500K to 1M tokens see median time to first token drop by about 35%, and it's already on by default for most Claude Code users.

The catch: a thread expires 5 minutes after its last request, and token usage stays the same because Claude still reads the full conversation every turn. It's a speed fix, not a cost fix.

A Lockfile for Your Agent Setup Microsoft's open-source APM (Agent Package Manager) lists the skills, prompts and MCP servers a project needs in one apm.yml, and on Nov 3 GitHub made Dependabot updates for APM available on all plans. Try it:

Copilot Inside Xcode GitHub Copilot now works in Xcode’s built-in coding assistant in public preview on all Copilot plans (Xcode 27.3 beta). The same weekly release lets VS Code run agent tasks on a schedule.

GPT-6 Cyber. OpenAI released GPT-6 Cyber for authorized security work and says it found 81 previously unknown high-severity Firefox bugs (OpenAI's numbers).

Thinking Machines announced Inkling 2 Small and says it intends to release the weights under Apache 2.0 "next week after the completion of external safety testing." As of Nov 8, the weights are closed and it's only on Tinker. Recheck Tuesday before treating it as downloadable.

The pattern this week is that the last mile is getting handled for you. Cursor deploys the app. ChatGPT animates the photo. Google edits the clip instead of making you start over.

The other half is the meter. ChatGPT and Codex are about to share one pool, and video bills by the second. Shipping got easier, so watching the cost matters more.

My advice: try one GIF today on the free plan; budget 10 minutes to test the workflow. Then look for Settings, then Usage, when available, so December doesn't surprise you.

Free, no card.

That's the Debrief. What are you turning into a GIF first? Reply and tell me, I read them all.

See you Friday with a free tool from the Academy.

From Drew and The AI Debrief team

P.S. Send us the GIF you made. The best ones might show up in a future issue. Hit reply, we read every one.

P.P.S. New here, or missed one? Every past edition is in the archive.

You're getting this because you signed up for The AI Debrief. Two emails a week and nothing else: the news early in the week, and a Friday edition for what to do about it. If it has stopped earning its place in your inbox, the unsubscribe link just below takes one click and we will stop straight away. We would much rather you leave clean than let it pile up unread.