Lakers say the Mark Williams/Dalton Knecht trade has been rescinded.

🍿Its here: the official Angular documentary. 🍿 Check it out on our YouTube channel. You can find the link in our bio or in the comments.

BREAKING: Sacramento is finalizing a trade to send De'Aaron Fox to the San Antonio Spurs in a multi-team trade that moves Chicago Bulls' Zach LaVine to the Kings, sources tell ESPN.

🐹 bridging the Go AI gap: introducing Agent-API 🐹 There's a big problem with AI tooling today: it's very Python-centric and can quickly lock you into a sole LLM provider. This can make adding AI capabilities to your existing Go backends difficult. A few weeks ago, I decided to work on this problem and make unlocking AI for Go simple, efficient, and safe. Introducing Agent-API: A fast, agnostic, and powerful Go AI framework for one-shot workflows, building autonomous agents, working with LLM providers, and integrating into your existing Go backends. There are currently 3 components to this toolkit: 1. The core APIs - This is where the interfaces for LLM providers are defined. It gives a common shape around messages, AI tools, and agents. 2. gsv - Go Schema Validator - gsv was born out of a desire to have a more dynamic workflow for defining tool parameters, compiling JSON schemas, validating arguments from AI tool calls, storing values, and overall, make working with AI tool json much easier in Go. 3. The providers - currently, the one I'm focused on is the ollama-provider - Each provider implements the core APIs making it easy to drop in and replace various providers (or mix and match abunch of them all at once). --- Let's look at a quick example: First, we can define an ollama provider that uses qwen2.5: ``` provider := ollama.NewOllamaProvider(log, " http://localhost", 11434, "qwen2.5") ``` Then, we can make an agent that uses that provider: ``` agent := agent.NewAgent(provider) ``` Let's give the agent some tools in its toolbelt! Here's a very simple "calculator" function that gets added to the agent: ``` err = agent.AddTool(tool.Tool{ Name: "calculator", Description: "Performs basic arithmetic operations: supported operations are 'add' and 'multiply'", WrappedFunction: wrappedCalc, JSONSchema: schema, }) ``` The "wrappedCalc" is a calculator function that's been wrapped to work with agent-api and is ready to accept arguments from an LLM (via tool calling). "schema" is a compiled json schema from gsv: this enables agent-api to validate and verify the incoming json from the LLM tool call. Now, we can send a message! ``` response, err := agent.SendMessage(ctx, "What is 5 + 3?") ``` Qwen on Ollama will utilize the tool to make the calculation and get an accurate result back! Check out some of the other examples in the examples repo: -- ⭐️Very early days and these APIs are still in their infancy so they may change unexpectedly but I'm very excited for where this is headed and I hope to make Go a first-class citizen in AI! Please star, contribute, drop an issue/idea/feature, or ask questions!!!