Today at Google Cloud’s Gemini at Work event, we announced Gemini, a new single universal agent for work that has all of your business context and can be used for everything from knowledge work to answering questions, and content creation to coding, all from a single prompt box. It’s... built around core architectural principles: ☑️ Unified Agent: Gemini can answer questions, do knowledge work and generate code from a single prompt box. ☑️Access: Gemini is web-based and can be accessed from any device and integrated into third-party applications. It can also operate without a dedicated user interface. ☑️Persistent Execution: It runs in the cloud, which means it maintains a single set of memories, context and one personalization graph no matter where you access it. ☑️Multi-Agent Orchestration: Gemini can create sub-agents to tackle multi-step tasks and can also act as a coworker agent with a defined role and its own dedicated identity ☑️Context: Gemini knows your tools, data, and work history and learns how you work the more you use it. ☑️Model Choice Flexibility: It orchestrates across multiple models to deliver optimal quality and lower your costs.show more

Is this just Gemini... but now it does more? I don't get "we announced Gemini." Like, Gemini existed already. Is this an upgrade to the Gemini we already have in Workspace? Or an entirely new thing with the same name?

Read more in our blog:

C'mon, publish Gemini 4 to the Pro users. You're really behind ur competitiors and it seems u don't really give a fuck!!

Gemini 4 Argon?

"today we announced Gemini" whoever has the job of naming things at google should be put down. gently, like the way you would with the family pet. it's time they rest

@Google Does it really work for complex use cases?Be honest. Asking because Spark is still useless even if it was launched in May before Muse, Instinct, Grok Bot and Fo.

who care？ we need gemini4

@Google We need Gemini 4 for pro plan🙌

Btw dont use gemini for code.

How about shipping Gemini 4 instead of announcing Gemini?

CoWork? When?

Okayy... So grok bot, muse & GPT dots "Enterprise version" > Honestly I'd expect this from MS (i.e Co-Pilot) yesterday instead of Google from all places.

This feels like Google turning Gemini into a real work platform, not just a chatbot. And the scale is already huge more on that here

Is this google's response to muse/dots/grok @grok thoughts?

Persistent context is powerful, but the next architectural question is what becomes reusable knowledge. If an agent learns from tools, data and work history, the valuable layer is being able to separate runtime memory from verified knowledge with provenance and replayable evidence.

Persistent memory across every device is the piece that would win me over.

wen?

The real shift is persistent execution. One agent that remembers your context, knows your tools, spins up sub-agents, and keeps working across devices feels less like a chatbot and more like an actual coworker.

One prompt box for coding? Risky move!

Knowledge work = document work. MS Office has 425M enterprise users and they will only use Gemini if it can compile and edit office formats. We’re building this for Gemini enterprise for a large client. Soon Gemini will make better PowerPoints than any other solution.

BIgger than i think most people initially will think, congrats guys!

the eternal pitch deck font doing more heavy lifting than the actual roadmap

Bullish

Anything but Gemini 4

the wildest part of the launch is that agents get their own gmail, calendar, and drive. google thinks the enterprise agent race is won on identity and permissions.

Orchestration and identity will decide if it sticks.

It's a very stupid model, it's like you talk to a 5yo

Sure boomer, sure thing. People have been asking for an agentic harness for work and you’ve just been releasing absolute horse shit the past year.

@Google explain this

@grok explain in nutshell

one prompt box for everything is a bold bet, my one prompt box already holds 47 half finished side projects

So boring...

When a document is unshared with a coworker agent, what happens to the facts it already saved in memory? I'd love to see that case in a demo, especially with a task still running.