Rohan Varma

Introducing ChatGPT for Dishwashing

GPT-6 achieves state-of-the-art results on dishwashing benchmarks for long-running jobs and difficult stains.

The Navier–Stokes Millennium Prize Problem is a $1 million challenge about whether the equations describing fluid motion can break down. In this video, I accidentally found a counter-example while trying to optimize my dishwasher. But in reality, OpenAI used 10,000 AI agents to find a proof in 88 hours: openai.com/index/navier-s…

Parody of the original GPT-6 Astra ad

@OpenAI #ChatGPT_Partner #chatgpt

Credits:

Written and Directed by Joma
Co-Writer - Henry Connor Coan IV @coan_henry
Production by Dauntlus Studios @dauntlus.studios
Producer/Co-Director - Marcus Liew @dauntlus
Camera Operator - Amos Lee @amoselijahlee
Art Director - Alethea Soo @aletheasoo
Art Assistant - Alicia Lim @aliciatkl_
Assistant Editor - GPT-6 Astra

Music: Tensions Run High by Soundridemusic
youtube.com/watch?v=Ly9H63…

Everyone is so focused on 5x, 10x, and 20x with these AI subscriptions, but no one ever asks what “x” is

Hmm lol @DoorDash @andyfang please upgrade the customer support bot or use humans 🙏🏾

One week since we launched dots!

We’ve been working through your feedback and shipped a bunch of improvements:
- Faster cloud browsing
- Fixed extra phone notifications while you’re already using ChatGPT
- Cleaner code blocks and attachments on web
- Fewer pauses while waiting for a reply
- Faster sidebar loading, smoother animations and less loading jank
- Fixed character rendering in Safari and Firefox
- Fixed Outlook setup getting stuck
- Clearer plan guidance + a way to retry failed activity loads
- More stable safety banners on web
- Fixed clipped approval buttons and oversized Gmail previews hiding approvals
- Fixed enterprise access issues on web and when starting voice
- Better dot organization of when to use new versus existing codex threads

Up next:
- Voice reliability (being improved everyday, but we’ll shout when it’s up to par with our expectations)
- Increasing dot’s control of your Codex app: manage project folders, have full Codex thread context, rename threads, and more
- Better organization to know what your dot is working on at any given moment
- Connect your dot to multiple computers
- Better Windows support for local computer control
- what else?

Thanks for trying our product and giving us lots of feedback. Keep it coming!

⌘.

Send the prompt below to your dot - it has been super useful for helping me manage my time!

"Monitor my calendar and let me know at 8a and 5p of all changes to the calendar: any new events added, any events deleted, and any events with new times. DM me on Slack at 8a everyday with a brief on all important meetings and list of meetings you think I could cancel."

Moving over to @OpenAI Dot from Muse. I'm using Chatgpt/Codex anyway daily for my personal stuff so simpler that way.

(I like keeping my work ai use with @linear since it has the context I need).

Also introducing sleepy Doomhammer orc peon:
Work, work. Okay dokey.

cool dot interaction

i asked a teammate for a slide yesterday on a random slack thread. dot saw the message, decided to keep tabs on the figma, the placeholder was still there this morning, so it gave me a heads up

all on its own

I'd love to chat with you all about dots, hear your ideas, and get feedback on some of the improvements we are planning!

I'm blocking off time everyday for the next couple weeks for this.

Down to chat with anyone willing to spare some time, but especially so if:
- you are daily driving a personal assistant tool (dots, bots, muse, etc) to do complex work, especially engineering (or have tried)
- you have been a power user of Codex, ChatGPT, or other agentic products but aren't finding much value from personal assistants
- you have dreams for how all these tools can work and want to steer us towards that direction

DM me on here and my dot will send you a scheduling link!

dot has become my primary interface to AI over the last few weeks.

my dot's name is boo. he's cute!

in a year, i bet i'll still be interacting in this way with @ChatGPT—e.g. as an always-on persistent agent—but boo will have disappeared. it reminds me a lot of early OpenClaw days—we loved the little personality quirks of our Claws, but we dropped them as soon as there was something functionally more powerful available

Improvements to Dot’s voice interface in progress!

If there is anything we’re missing on @athyuttamre’s list that you’d like to see, please share!

Proactivity is one of the defining features of Dots.

I think Grok Bot is super cool and being built by a legendary team of builders. I’m glad they were able to add proactivity to their product as well!

@poteto we can work off of a shared roadmap, delegate to dots and bots, go grab some beers, and see what happens :)

Very cool dots use cases!

🎯

My dot is now managing my Claude.

Before:
You manage your Chat/Codex threads

Now:
You delegate to your dot, that manages your Chat/Codex threads

Delegation isn’t for everyone or everything (I mean this earnestly) - but when you want to work this way, it’s a magical experience.

I’ll share some demos later on how this is a fundamental leap in usefulness!

You can move AI products forward in two ways:
- increase capabilities
- increase accessibility

When we went from chat -> agents, that was a capability leap.

When we went from CLI -> ADE (Codex App), that was an accessibility leap.

Im over generalizing, but the Codex App fundamentally makes it easier to do things which are reasonably possible to do with the CLI tool. However, most people weren’t doing them until we built the App: multi agent orchestration, agent to agent hand offs, review focused development, backgrounding tasks, simple connectors, and more.

Many enterprises and individuals became a lot more productive when switching from the CLI to the App, even with no feature or model jumps.

In a lot of ways, Dots also represent an accessibility leap.

You can, in theory, set up your Codex to:
- spin itself back up to keep working
- prompt itself to carry on another task
- spin up sidecar agents to find work to do on your behalf
- embody itself in all platforms you work in
- summarize the work it’s doing into digestible chunks
- surface opportunities for work proactively
- use automations creatively
- decide to operate out of a single agent control plane
- and more

And to be honest, I know dozens of people working this way. Many of you on X are doing this today!

But the reason we built dots, and what gets us excited every day, is how to bring this frontier way of working to our 1.2B+ users and 2.5 million businesses.

I’ve had conversations with a handful of enterprises leaders about dots since we launched the product, and it’s incredible how game changing this is for them. AI’s capability overhang today is largely an understanding, training, and enablement problem. With dots, they don’t have to train any of their employees on how to use it. Everyone already knows how to work with another person. It just works.

We’re so early in this product category and there is so much left to build. I’m excited to see where it goes, and I’m excited to see how much more we can do on behalf of all of our users with this new modality.

It sounds like folks aren’t that impressed with the flight booking product we built.

I currently have my dot navigating the US immigration system for a friend moving here - is that sufficiently complex or do most people typically want to do that by hand too 😬

Now that I mostly talk to dots, going back to chat / agent interfaces feels oddly inorganic.

Talking to dots feels natural, like talking to a person.

Watching Codex call tools and stream intermediate outputs kinda breaks the illusion for me.

Pretty crazy how quickly that feeling developed and how fine, excited even, I was with the previous way until we built this new way.

We're iterating on the UI to make it more natural for you and your Dot to work on Codex threads together, but this is a cool way to manage it today!