.Ansem 🐂🀄️ wanted a bot to automatically short ETH whenever the timeline starts posting bullish ETHBTC charts, so we built him an AI agent using Syntax V2 (launching soon). Here’s how you can build one too, even if you're new to this! A 🧵~

☝️First, let's head to Syntax V2 and click on “Create a new Agent”. Then lets copy @blknoiz06's tweet, add some context to our prompt, and clarify it's from the Twitter timeline. Now, let's start the Q&A session with the agent builder.

📝 The agent builder will propose a few things like a description and high-level instructions. It will also ask for a name. Keep it simple for now. It will ask more questions to refine the details, like the specific amount of ETH you want to trade.

🧐 You can then specify what the agent should monitor on Twitter. Let’s say all relevant tweets. It's almost ready to generate the agent instructions. While we wait, let's explain what's happening behind the scenes.

🔍 We're using our Agent Creation Flow, which starts with you interacting with the agent builder. In this video example, we use Ansem's tweet as a prompt. The agent builder collects all the info needed to create your agent.

🧩 Next, the segmentation process brings our L.E.A Model in action, and organizes the info into three categories: Listen for ETHBTC-related tweets, Evaluate if they're bullish, and then Act to short ETH. The code for each step starts generating automatically.

🛠️ The code critic process kicks in, evaluating the generated code with static analysis, unit tests, and suggesting improvements. Once everything checks out, the final code is presented to you.

📈 Once the agent is deployed you can chat with it and set the amount of ETH you’d like to short. You can create multiple positions (jobs) like these, shorting a different amount in each job, and the agent will automatically transact on-chain for you. That's how easy it is to create an agent on Syntax!

@blknoiz06 It’s infinitely harder to make a profitable trading bot than this, I’ve made tons of them and screwed around a lot with historical data, every indicator, trained custom deep networks, etc Everybody that uses these agents will lose money but the people creating the service

@blknoiz06 OMG THE TECH IS REAL

@blknoiz06 $SPEC