The best AI workspace for teams. Build your billion-dollar company at https://xitter.cf/t.co/SfNrJSggUH
SF
Joined June 2026
- Tweets93
- Following6
- Followers1.4K
- Likes142
Pinned Tweet
Now, build with Offloop.
Join the waitlist. Build your own profit-driving agent army.
offloop.org
Introducing Offloop!
We're a team of four. Today our multi-agent harness hit state of the art on GDPval, ahead of Claude code and Codex across jobs that pay $2.4 trillion a year in the US.
Offloop gives every knowledge worker what the Fortune 500 spends billions on: a high-performing agent army that runs itself and grows the business.
Offloop retweeted
Prompt is dead. Now just make a wish.
Guys, I just test latest deepseek v4.1 flash 1017 @deepseek_ai.
It is insane. It is so insane.
ONLY one prompt, one shot, and it built the cod style fps game:
build a call of duty 23 fps game use three js.
It auto build 37 mins and JUST 0.1 USD!
5-wave tactical survival operation, 3 operators，3 difficulty protocols.
I can not find word to describe my feelings now.
It seems the same level as Gemini 4 and better than Chatgpt 6.1 sol
@deepseek_ai is back.
Offloop retweeted
i've been testing DeepSeek v4.1 Flash 1017 for the past few hours.
-> one shot, 31 minutes of autonomous execution
-> only $0.22
-> the promising part is how far it can carry a task from a brief description without me steering every step. this is the kind of long-running execution that makes “just describe what you want” coding feel useful.
this SpongeBob showcase is one of the results, a short prompt asking for a 30-second Three.js scene with water absorption, simulated physics, cinematic lighting and bubbles.
We gave Gemini 4 Argon and Claude Opus 5.5 same prompt
Gemini 4 Argon ~ 3m14s, $0.31
Opus 5.5 ~ 15m21s, $2.47
~5x the time, ~8x the cost. Left or right?
With Gemini 4 Argon, your research has new ways to tell its story 📖
We used Nano Banana, Veo, and Argon to turn a dense academic PDF into a full-color, animated storybook in the @GeminiApp.
After half a day of failed attempts and retries with Gemini 4 Argon, finally tested the margarita sloshing effect with:
→ Gemini 4 Argon, $1.64
→ GPT 6 Astra, $4.90
→ GLM 5.3, $0.93
→ DeepSeek Flash, $0.24
They all used Three.js + WebGL.
Gemini has improved so muchhh 🍸
Tested with Offloop.
Your next customer might already be in your pipeline.
Offloop’s agents review past conversations and spot new signals across the web, so you can follow up as soon as an opportunity opens up.
Try it at offloop.org
Offloop retweeted
I’m not sure why @Offloop doesn’t promote itself more widely, but if you work in marketing or SEO, it’s worth giving it a try.
Less copying files into chat. More letting your agent work with them.
With Offloop Connect, choose a Google Drive account and grant your agent access right in the conversation to continue the task.
You choose the account. You choose the agent.
Async tool calls and mid-turn steering look like api features, but they change how agents can exist inside an organization.
Real work is asynchronous. people interrupt each other. requirements change. external systems are slow. multiple things happen at once.
If agents are going to become actual participants in team workflows, the runtime has to support this messiness.
This is why we’re increasingly convinced the next layer of ai infrastructure is an org-level agent harness.
While integrating GPT-6 Astra, the most interesting upgrade might not be the model itself, but what’s happening to the Responses API.
Async tool calls: the model can keep reasoning and doing other work while a tool is still running.
Mid-turn steering: users can change requirements while the agent is already working, without restarting the whole turn.
Dynamic reasoning: reasoning effort can change mid-conversation while preserving the prompt cache.
This feels like an important shift that Responses API is becoming less of an “LLM API” and more of an actual runtime protocol for long-running agents. Those features only available to post gpt 5.6 models means they explicitly optimized during post-training.
GPT-6-Astra is coming to offloop.
Computer use, browsing, coding... impressive!
Next test is here -> how well can it actually work inside your team? (try it at: offloop.org)
Introducing Subchannels.
When people and AI agents share one Channel, parallel conversations pile up, and context gets buried.
Now, you can fork any message into a focused Subchannel.
Try it here: ol.app/subchannels
Offloop retweeted
This Product Hunt launch almost didn’t happen.
We had less than 10 minutes to remove the waitlist and open access. Product Hunt doesn’t allow waitlist-only launches.
We made it.
Now we’re competing today.
If you have 20 seconds, a vote + comment would mean a lot to us.
producthunt.com/products/off…
Big companies have headcount. Small teams now have agents.
Offloop is built for you: teammates and agents in one conversation, sharing context, handing off, shipping together.
Sign up on your own, or bring Offloop to your team and start building with AI together: ol.app/tm/z7c4va
The gap between a 4-person team and a 400-person org has never been smaller.
We built Offloop as an org-level harness.
How it works:
1）Open a channel with your team and the agents you need
2）@ mention an agent with a bounded task — status goes Requested → Working
3）The work comes back in the channel, ready for review
Any teammate continues from the result. Context stays put.
Sign up here: ol.app/tm/z7c4va
